gps.py

#!usr/bin/env python
# -*- coding: utf-8 -*-
# kristine larson, june 2017
# toolbox for GPS/GNSS data analysis
import datetime
from datetime import date
import math
import os
import pickle
import re
import scipy.signal as spectral
from scipy.interpolate import interp1d
import subprocess
import sys

import matplotlib.pyplot as plt
import numpy as np
import wget

import read_snr_files as snr

# various numbers you need in the GNSS world
# mostly frequencies and wavelengths
class constants:
    c= 299792458 # m/sec
#   GPS frequencies and wavelengths
    fL1 = 1575.42 # MegaHz 154*10.23
    fL2 = 1227.60 # 120*10.23
    fL5 = 115*10.23
#  GPS wavelengths
    wL1 = c/(fL1*1e6) # meters wavelength
    wL2 = c/(fL2*1e6)
    wL5 = c/(fL5*1e6)
#   galileo frequency values
    gal_L1 = 1575.420
    gal_L5 = 1176.450
    gal_L6 = 1278.70
    gal_L7 = 1207.140
    gal_L8 = 1191.795
#  galileo wavelengths, meters
    wgL1 = c/(gal_L1*1e6)
    wgL5 = c/(gal_L5*1e6)
    wgL6 = c/(gal_L6*1e6)
    wgL7 = c/(gal_L7*1e6)
    wgL8 = c/(gal_L8*1e6)

#   beidou frequencies and wavelengths
    bei_L2 = 1561.098
    bei_L7 = 1207.14
    bei_L6 = 1268.52
    wbL2 = c/(bei_L2*1e6)
    wbL7 = c/(bei_L7*1e6)
    wbL6 = c/(bei_L6*1e6)

#   Earth rotation rate used in GPS Nav message
    omegaEarth = 7.2921151467E-5 #	%rad/sec
    mu = 3.986005e14 # Earth GM value


class wgs84:
    """
    wgs84 parameters
    """
    a = 6378137. # meters
    f  =  1./298.257223563 # flattening factor
    e = np.sqrt(2*f-f**2) # 

def define_filename(station,year,doy,snr):
    """
    given station name, year, doy, snr type
    year doy and snr are integers
    returns snr filenames (both uncompressed and compressed)
    author: Kristine Larson
    19mar25: return compressed filename too
    20apr12: fixed typo in xz name!
    """
    xdir = os.environ['REFL_CODE']
    cdoy = '{:03d}'.format(doy)
    cyy = '{:02d}'.format(year-2000)
    f= station + str(cdoy) + '0.' + cyy + '.snr' + str(snr)
    fname = xdir + '/' + str(year) + '/snr/' + station + '/' + f 
    fname2 = xdir + '/' + str(year) + '/snr/' + station + '/' + f  + '.xz'
    return fname, fname2

def define_and_xz_snr(station,year,doy,snr):
    """
    given station name, year, doy, snr type
    returns snr filename
    author: Kristine Larson
    19mar25: return compressed filename too
    20apr12: fixed typo in xz name! now try to compress here
    """
    xdir = os.environ['REFL_CODE']
    cdoy = '{:03d}'.format(doy)
    cyy = '{:02d}'.format(year-2000)
    f= station + str(cdoy) + '0.' + cyy + '.snr' + str(snr)
    fname = xdir + '/' + str(year) + '/snr/' + station + '/' + f
    fname2 = xdir + '/' + str(year) + '/snr/' + station + '/' + f  + '.xz'
    snre = False
    if os.path.isfile(fname):
        print('snr file exists')
        snre = True
    else:
        if os.path.isfile(fname2):
            print('found xz compressed snr file - try to unxz it')
            subprocess.call(['unxz', fname2])
        # check that it was success
            if os.path.isfile(fname):
                snre = True
#   return fname2 but mostly for backwards compatibility
    return fname, fname2, snre 

def define_filename_prevday(station,year,doy,snr):
    """
    given station name, year, doy, snr type
    returns snr filename for the PREVIOUS day
    author: Kristine Larson
    """
    xdir = os.environ['REFL_CODE']
    year = int(year)
    doy = int(doy)
    if (doy == 1):
        pyear = year -1
        print('found january 1, so previous day is december 31')
        doyx,cdoyx,cyyyy,cyy = ymd2doy(pyear,12,31)
        pdoy = doyx 
    else:
#       doy is decremented by one and year stays the same
        pdoy = doy - 1
        pyear = year

    cdoy = '{:03d}'.format(pdoy)
    cyy = '{:02d}'.format(pyear-2000)
    f= station + str(cdoy) + '0.' + cyy + '.snr' + str(snr)
    fname = xdir + '/' + str(year) + '/snr/' + station + '/' + f 
    fname2 = xdir + '/' + str(year) + '/snr/' + station + '/' + f + '.xv'
    print('snr filename for the previous day is ', fname) 
    return fname, fname2

def read_inputs(station):
    """
    given station name, read LSP parameters for strip_snrfile.py
    author: Kristine M Larson
    19sep22 change name of longitude variable, add error messages
    20may20 add peak to noise as an returned variable

    """
#   directory name is currently defined using REFL_CODE
    peak2noise = 0 # assume it is not set here but in the main code,
    # which is how it used to be
    xdir = os.environ['REFL_CODE']
    fname = xdir + '/input/' + station
    print('default inputs should be in this file: ', fname)
#   default location values - not used now
    lat = 0; lon = 0; h = 0;
#   counter variables
    k = 0
    lc = 0
#   elevation angle list, degrees
    elang = []
#   azimuth angle list, degrees
    azval = []
    freqs = [] ; reqAmp = []
#   this limits the reflector height values you compute LSP for
    Hlimits = []
#   default polyfit, elevation angle limits for DC removal
    polyFit = 3; pele = [5, 30]
#   desired reflector height precision, in meters
    desiredP = 0.01
    ediff = 2
    noiseRegion = [0.25, 6]
    try:
        f = open(fname, 'r')
#       read all the lines
        l=f.readlines()
#       figure out what is on each
        for line in l:
#           print(line)
            if line[0] == '#':
                lc = lc + 1
            else:
                k=k+1
                nfo=line.split()
                if k==1:
# read in the latitude, longitude, and height. currently this information is not used
                    lat = float(nfo[0])
                    lon = float(nfo[1])
                    h = float(nfo[2])
                    print("latitude/longitude/height:", lat,lon,h)
# read in the elevation angle ranges
                if k==2:
                    elang.append(float(nfo[0]))
                    elang.append(float(nfo[1]))
                    print('elevation angle limits', elang[0], elang[1])
                    # these are elevation angle limits for polyfit
                    if (len(nfo)==4):
                        pele = [float(nfo[2]), float(nfo[3]) ]
# read in the azimuth ranges
                if k==3:
                    naz = len(nfo)
                    for j in range(naz):
                        azval.append(float(nfo[j]))
# read in the frequencies and the required spectral amplitudes
                if k==4:
                    nnp = int(len(nfo)/2)
                    for j in range(nnp):
                        ni = 2*j; nj = 2*j+1
                        freqs.append(int(nfo[ni]))
                        reqAmp.append(float(nfo[nj]))
                        print('Frequency ', int(nfo[ni]), ' Amplitude ', float(nfo[nj]))
# read in reflector height restrictions etc
                if k==5:
                    np = len(nfo)
                    polyFit = int(nfo[0])
# this is desired precision of the LSP
                    desiredP = float(nfo[1])
                    H1 = Hlimits.append(float(nfo[2]))
                    H2 = Hlimits.append(float(nfo[3]))
                    ediff = float(nfo[4])
                    print('Polynomial Fit ', polyFit, ' Precision: ', desiredP,' ediff ', ediff)
                    if (np > 5):
# range of reflector heights used for noise calculation
                        ns1 = float(nfo[5])
                        ns2 = float(nfo[6])
                        noiseRegion = [ns1,ns2]
                        # this means peak to noise is on the line too
                        if np > 7:
                            peak2noise = float(nfo[7])
                        
        f.close
    except:
        print('Some kind of problem reading the required input file: ' + fname)
        print('Please use make_input_file.py if it does not exist. It will need at a minimum the ')
        print('latitude, longitude, and height of the station.  This location does not need to be precise')
        print('If you do not have your site location handy, just enter 0,0,0.')
        print('The rest of the inputs can be defaults, but if ')
        print('your site requires RH > 6 meters, please adjust using -h1 and -h2. Exiting now.')
        sys.exit()

    return lat,lon,h,elang, azval, freqs, reqAmp,polyFit, desiredP, Hlimits, ediff, pele,noiseRegion, peak2noise

def satclock(week, epoch, prn, closest_ephem):
    """
    inputs: gps week, second of week, satellite number (PRN)
    and ephemeris. returns clock clorrection in  meters
    note: although second order correction exists, it is not used  
    """
    # what is sent should be the appropriate ephemeris for given
    # satellite and time
    prn, week, Toc, Af0, Af1, Af2, IODE, Crs, delta_n, M0, Cuc,\
    ecc, Cus, sqrta, Toe, Cic, Loa, Cis, incl, Crc, perigee, radot, idot,\
    l2c, week, l2f, sigma, health, Tgd, IODC, Tob, interval = closest_ephem

    correction = (Af0+Af1*(epoch-Toc))*constants.c
    return correction[0]


def ionofree(L1, L2):
    """
    input are L1 and L2 observables (either phase or pseudorange, in meters)
    output is L3 (meters)
    author: kristine larson
    """
    f1 = constants.fL1
    f2 = constants.fL2
    
    P3 = f1**2/(f1**2-f2**2)*L1-f2**2/(f1**2-f2**2)*L2
    return P3

def azimuth_angle(RecSat, East, North):
    """
    kristine larson
    inputs are receiver satellite vector (meters)
    east and north unit vectors, computed with the up vector
    returns azimuth angle in degrees
    """
    staSatE = East[0]*RecSat[0] + East[1]*RecSat[1] + East[2]*RecSat[2]
    staSatN = North[0]*RecSat[0] + North[1]*RecSat[1] + North[2]*RecSat[2]
#    azangle = 0
    azangle = np.arctan2(staSatE, staSatN)*180/np.pi
    if azangle < 0:
        azangle = 360 + azangle
# 
    return azangle

def rot3(vector, angle):
    """
    input a vector (3) and output the same vector rotated by angle
    in radians apparently.
    from ryan hardy
    """
    rotmat = np.matrix([[ np.cos(angle), np.sin(angle), 0],
                        [-np.sin(angle), np.cos(angle), 0],
                        [             0,             0, 1]])
    vector2 = np.array((rotmat*np.matrix(vector).T).T)[0]
    return vector2

def xyz2llh(xyz, tol):
    """
    inputs are station coordinate vector xyz (x,y,z in meters), tolerance for convergence
    outputs are lat, lon in radians and wgs84 ellipsoidal height in meters
    kristine larson
    """
    x=xyz[0]
    y=xyz[1]
    z=xyz[2]
    lon = np.arctan2(y, x)
    p = np.sqrt(x**2+y**2)
    lat0 = np.arctan((z/p)/(1-wgs84.e**2))
    b = wgs84.a*(1-wgs84.f)
    error = 1
    a2=wgs84.a**2
    i=0 # make sure it doesn't go forever
    while error > tol and i < 10:
        n = a2/np.sqrt(a2*np.cos(lat0)**2+b**2*np.sin(lat0)**2)
        h = p/np.cos(lat0)-n
        lat = np.arctan((z/p)/(1-wgs84.e**2*n/(n+h)))
        error = np.abs(lat-lat0)
        lat0 = lat
        i+=1
    return lat, lon, h

def xyz2llhd(xyz):
    """
    inputs are station vector xyz (x,y,z in meters), tolerance for convergence is hardwired
    outputs are lat, lon in degrees and wgs84 ellipsoidal height in meters
    author : kristine larson
    """
    x=xyz[0]
    y=xyz[1]
    z=xyz[2]
    lon = np.arctan2(y, x)
    p = np.sqrt(x**2+y**2)
    lat0 = np.arctan((z/p)/(1-wgs84.e**2))
    b = wgs84.a*(1-wgs84.f)
    error = 1
    a2=wgs84.a**2
    i=0 # make sure it doesn't go forever
    tol = 1e-10
    while error > tol and i < 10:
        n = a2/np.sqrt(a2*np.cos(lat0)**2+b**2*np.sin(lat0)**2)
        h = p/np.cos(lat0)-n
        lat = np.arctan((z/p)/(1-wgs84.e**2*n/(n+h)))
        error = np.abs(lat-lat0)
        lat0 = lat
        i+=1
    return lat*180/np.pi, lon*180/np.pi, h



def zenithdelay(h):
    """
    author: kristine larson
    input the station ellipsoidal (height) in meters
    the output is a very simple zenith troposphere delay in meters
    """

    zd = 0.1 + 2.31*np.exp(-h/7000.0)
    return zd

def up(lat,lon):
    """
    author: kristine larson
    inputs rae latitude and longitude in radians
    returns the up unit vector, and local east and north used for azimuth calc.
    """
    xo = np.cos(lat)*np.cos(lon)
    yo = np.cos(lat)*np.sin(lon)
    zo = np.sin(lat)
    u= np.array([xo,yo,zo])    
#    c ... also define local east/north for station: took these from fortran
    North = np.zeros(3)
    East = np.zeros(3)
    North[0] = -np.sin(lat)*np.cos(lon)
    North[1] = -np.sin(lat)*np.sin(lon)
    North[2] = np.cos(lat)
    East[0] = -np.sin(lon)
    East[1] = np.cos(lon)
    East[2] = 0
    return u, East, North

def norm(vect):
    """
    given a three vector - return its norm
    """  
    nv = np.sqrt(np.dot(vect,vect))
    return nv

def elev_angle(up, RecSat):
    """
    inputs:
    up - unit vector in up direction
    RecSat is the Cartesian vector that points from receiver 
    to the satellite in meters
    the output is elevation angle in radians
    author: kristine larson
    """
    ang = np.arccos(np.dot(RecSat,up) / (norm(RecSat)))
    angle = np.pi/2.0 - ang
    return angle

def sp3_interpolator(t, tow, x0, y0, z0, clock0):
    """
    got this from ryan hardy - who coded it for my class?
    inputs are??? tow is GPS seconds
    xyz are the precise satellite coordinates (in meters)
    clocks are likely satellite clock corrections (microseconds)
    i believe n is the order fit, based on what i recall.
    these values do not agree with my test cases in matlab or fortran
    presumably there is an issue with the estimation of the coefficients.
    they are good enough for calculating an elevation angle used in reflectometry
    """
    # ryan set it to 7 - which is not recommended by the paper
    # ryan confirmed that he doesn't know why this doesn't work ...
    n = 7 # don't know why this was being sent before
    coeffs = np.zeros((len(t), 3, n))
    # whatever ryan was doing was not allowed here.  had to make
    # sure these are treated as integers
    s1 = int(-(n-1)/2)
    s2 = int((n-1)/2+1)
#    print(s1,s2)
    
    omega = 2*2*np.pi/(86164.090530833)
    x = np.zeros(len(t))
    y = np.zeros(len(t))
    z = np.zeros(len(t))
    clockf = interp1d(tow, clock0, bounds_error=False, fill_value=clock0[-1])
    clock = clockf(t)
    # looks like it computes it for a number of t values?
    for i in range(len(t)):
        # sets up a matrix with zeros in it - 7 by 7
        independent = np.matrix(np.zeros((n, n)))
        # no idea what this does ...
        m = np.sort(np.argsort(np.abs(tow-t[i]))[:n])
        tinterp = tow[m]-np.median(tow[m])
        # set x, y, and z to zeros
        xr = np.zeros(n)
        yr = np.zeros(n)
        zr = np.zeros(n)
        # coefficients are before and after the time
        for j in range(s1, s2):
            independent[j] = np.cos(np.abs(j)*omega*tinterp-(j > 0)*np.pi/2)
        for j in range(n):
            xr[j], yr[j], zr[j] = rot3(np.array([x0[m], y0[m], z0[m]]).T[j], omega/2*tinterp[j])
 #           print(j, xr[j], yr[j], zr[j])
			
        independent = independent.T
        eig =  np.linalg.eig(independent)
        iinv  = (eig[1]*1/eig[0]*np.eye(n)*np.linalg.inv(eig[1]))
# set up the coefficients
        coeffs[i, 0] = np.array(iinv*np.matrix(xr).T).T[0]
        coeffs[i, 1] = np.array(iinv*np.matrix(yr).T).T[0]
        coeffs[i, 2] = np.array(iinv*np.matrix(zr).T).T[0]
        
        j = np.arange(s1, s2)
        # time since median of the values?
        tx = (t[i]-np.median(tow[m]))
        r_inertial =  np.sum(coeffs[i][:, j]*np.cos(np.abs(j)*omega*tx-(j > 0)*np.pi/2), -1)

        x[i], y[i], z[i] = rot3(r_inertial, -omega/2*tx)
        # returns xyz, in meters ? and satellite clock in microseconds
        return x*1e3, y*1e3, z*1e3, clock
 
    
def readPreciseClock(filename):     
    """
    author: kristine larson
    filename of precise clocks
    returns prn, time (gps seconds of the week), and clock corrections (in meters)
    """          
    StationNFO=open(filename).readlines()
    c= 299792458 # m/sec
 
    nsat = 32
    k=0
# this is for 5 second clocks
    nepochs = 17280
    prn = np.zeros(nepochs*nsat)
    t = np.zeros(nepochs*nsat)
    clockc = np.zeros(nepochs*nsat)
# reads in the high-rate clock file 
    for line in StationNFO:
        if line[0:4] == 'AS G':
            lines= line[7:60]
            sat = int(line[4:6])
            year = int(lines.split()[0])
            month = int(lines.split()[1])
            day = int(lines.split()[2])
            hour = int(lines.split()[3])
            minutes = int(lines.split()[4])
            second = float(lines.split()[5])
            [gw, gpss] = kgpsweek(year, month, day, hour, minutes, second)
            clock = float(lines.split()[7])
            prn[k] = sat
            t[k]=int(gpss)
            clockc[k]=c*clock
            k += 1
    return prn, t, clockc


def ymd2doy(year,month,day):
    """
    takes in year, month, day and returns day of year (doy)
    and the character version of that day of year
    19mar20: and now it returns character versions of 4 and 2 character years
    """
    today=datetime.datetime(year,month,day)
    doy = (today - datetime.datetime(today.year, 1, 1)).days + 1
    cdoy = '{:03d}'.format(doy)
    cyyyy = '{:04d}'.format(year)
    cyy = '{:02d}'.format(year-2000)
    return doy, cdoy, cyyyy, cyy

def rinex_unavco(station, year, month, day):
    """
    author: kristine larson
    picks up a RINEX file from unavco.  it tries to pick up an o file,
    but if it does not work, it tries the "d" version, which must be
    decompressed.  the location of this executable is defined in the crnxpath
    variable. This is from the main unavco directory - not the highrate directory.

    WARNING: only rinex version 2 in this world
    """
    exedir = os.environ['EXE']
    crnxpath = hatanaka_version() 
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    rinexfile,rinexfiled = rinex_name(station, year, month, day)
    unavco= 'ftp://data-out.unavco.org'
    filename1 = rinexfile + '.Z'
    filename2 = rinexfiled + '.Z'
    # URL path for the o file and the d file
    url1 = unavco+ '/pub/rinex/obs/' + cyyyy + '/' + cdoy + '/' + filename1
    url2 = unavco+ '/pub/rinex/obs/' + cyyyy + '/' + cdoy + '/' + filename2
    try:
        print('try to get observation file')
        wget.download(url1,filename1)
        status = subprocess.call(['uncompress', filename1])
        print('found the rinex file')
    except:
        print('did not find observation file, try hatanaka version')
        try:
            wget.download(url2,filename2)
            status = subprocess.call(['uncompress', filename2])
            status = subprocess.call([crnxpath, rinexfiled])
            status = subprocess.call(['rm', '-f', rinexfiled])
            print('found d file and converted to observation file')
        except:
            print('failed to find either RINEX file at unavco')


def rinex_sopac(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a hatanaka RINEX file from SOPAC - converts to o
    hatanaka exe hardwired  for my machine
    """
    exedir = os.environ['EXE']
    crnxpath = hatanaka_version()
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    sopac = 'ftp://garner.ucsd.edu'
    oname,fname = rinex_name(station, year, month, day) 
    # compressed version??
    file1 = fname + '.Z'
    path1 = '/pub/rinex/' + cyyyy + '/' + cdoy + '/' 
    url1 = sopac + path1 + file1 

    #file2 = oname   + '.Z'
    #path2 = '/pub/rinex/' + cyyyy + '/' + cdoy + '/' 
    #url2 = sopac + path2 + file2 
    try:
        wget.download(url1,file1)
        subprocess.call(['uncompress', file1])
        subprocess.call([crnxpath, fname])
        subprocess.call(['rm', '-f',fname])
        print('successful Hatanaka download from SOPAC ')
    except:
        print('not able to download from SOPAC',file1)
        subprocess.call(['rm', '-f',file1])
        subprocess.call(['rm', '-f',fname])

def rinex_sonel(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a hatanaka RINEX file from SOPAC - converts to o
    hatanaka exe hardwired  for my machine
    """
    exedir = os.environ['EXE']
    crnxpath = hatanaka_version()
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    sonel = 'ftp://ftp.sonel.org'
    oname,fname = rinex_name(station, year, month, day) 
    file1 = fname + '.Z'
    path1 = '/gps/data/' + cyyyy + '/' + cdoy + '/' 
    url = sonel + path1 + file1 
    print(url)

    # I don't think they have normal RINEX, so only hatanaka
    try:
        wget.download(url,file1)
        subprocess.call(['uncompress', file1])
        subprocess.call([crnxpath, fname])
        subprocess.call(['rm', '-f',fname])
        print('successful Hatanaka download from SONEL ')
    except:
        print('some kind of problem with Hatanaka download from SONEL',file1)
        subprocess.call(['rm', '-f',file1])

def rinex_cddis(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a hatanaka RINEX file from CDDIS - converts to o

    June 2020, changed  to use secure ftp
    if day is zero, then month is assumed to be doy
    
    """
    exedir = os.environ['EXE'] # do not need this?
    crnxpath = hatanaka_version()
    if (day == 0):
        doy = month
        year, month, day, cyyyy,cdoy, YMD = ydoy2useful(year,doy)
        cyy = cyyyy[2:4]
    else:
        doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)

    cddis = 'ftp://ftp.cddis.eosdis.nasa.gov'
    oname,fname = rinex_name(station, year, month, day)
    file1 = oname + '.Z'
    url = cddis + '/pub/gnss/data/daily/' + cyyyy + '/' + cdoy + '/' + cyy + 'o/' + file1

    # try new way using secure ftp
    dir_secure = '/pub/gnss/data/daily/' + cyyyy + '/' + cdoy + '/' + cyy + 'o/'
    file_secure = file1

    try:
        #wget.download(url,file1)
        cddis_download(file_secure,dir_secure)
        subprocess.call(['uncompress', file1])
    except:
        print('some issue at CDDIS',file1)
        subprocess.call(['rm', '-f',file1])
    if os.path.exists(oname):
        print('successful RINEX 2.11 download from CDDIS')
    else:
        print('unsuccessful RINEX 2.11 download from CDDIS')

def rinex_nz(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a RINEX file from GNS New zealand
    you can input day =0 and it will assume month is day of year
    20jul10 - changed access point and ending
    """
    # if doy is input 
    if day == 0:
        doy=month
        d = doy2ymd(year,doy);
        month = d.month; day = d.day
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    gns = 'ftp://ftp.geonet.org.nz/gnss/rinex/'
    oname,fname = rinex_name(station, year, month, day)
    file1 = oname + '.gz'
    url = gns +  cyyyy + '/' + cdoy +  '/' + file1
    print(url)

    try:
        wget.download(url,file1)
        subprocess.call(['gunzip', file1])
        print('successful download from GeoNet New Zealand')
    except:
        print('some kind of problem with download',file1)
        subprocess.call(['rm', '-f',file1])

def rinex_bkg(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a lowrate RINEX file BKG
    you can input day =0 and it will assume month is day of year
    """
    crnxpath = hatanaka_version()
    # if doy is input 
    if day == 0:
        doy=month
        d = doy2ymd(year,doy);
        month = d.month; day = d.day
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    gns = 'ftp://igs.bkg.bund.de/EUREF/obs/'
    oname,fname = rinex_name(station, year, month, day)
    # they store hatanaka and normal unix compression
    file1 = fname + '.Z'
    url = gns +  cyyyy + '/' + cdoy +  '/' + file1
    print(url)

    try:
        wget.download(url,file1)
        subprocess.call(['uncompress', file1])
        print('successful download from BKG')
        subprocess.call([crnxpath, fname])
        subprocess.call(['rm', '-f',fname])
        print('successful Hatanaka download and translation from BKG')
    except:
        print('some kind of problem with download',file1)
        subprocess.call(['rm', '-f',file1])

def rinex_nrcan(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a RINEX file from GNS New zealand
    you can input day =0 and it will assume month is day of year

    Note: this code does not work - let me know if you have a way
    to access NRCAN anonymously
    """
    exedir = os.environ['EXE']
    crnxpath = exedir + '/CRX2RNX'
    # if doy is input 
    if day == 0:
        doy=month
        d = doy2ymd(year,doy);
        month = d.month; day = d.day
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    # was using this ...
    gns = 'ftp://gauss.geod.nrcan.gc.ca/data/ftp/naref/pub/rinex/'
    gns = 'ftp://gauss.geod.nrcan.gc.ca/data/ftp/naref/pub/data/rinex/'
    # user narefftp
    # password 4NAREF2use
    oname,fname = rinex_name(station, year, month, day)
    # only hatanaka in canada and normal compression
    file1 = fname + '.Z'
    url = gns +  cyy + cdoy + '/' +  file1
    print(url)

    try:
        wget.download(url,file1)
        subprocess.call(['uncompress', file1])
        print('successful download from NRCAN ')
    except:
        print('some kind of problem with download',file1)
        subprocess.call(['rm', '-f',file1])

def getnavfile(year, month, day):
    """
    author: kristine larson
    given year, month, day it picks up a GPS nav file from SOPAC
    and stores it in the ORBITS directory
    returns the name of the file,  its directory, and a boolean
    19may7 now checks for compressed and uncompressed nav file
    19may20 now allows day of year input if day is set to zero
    20apr15 check for xz compression

    if the day is zero it assumes month is doy

    """
    foundit = False
    ann = make_nav_dirs(year)
    if (day == 0):
        doy = month
        year, month, day, cyyyy,cdoy, YMD = ydoy2useful(year,doy)
        cyy = cyyyy[2:4]
    else:
        doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    navname,navdir = nav_name(year, month, day)
    nfile = navdir + '/' + navname
    if os.path.exists(navdir):
        print('navdir already exists')
    else:
        subprocess.call(['mkdir',navdir])

    if os.path.exists(nfile):
        foundit = True
    elif os.path.exists(nfile + '.xz' ):
        print('xz compressed navfile exists online')
        subprocess.call(['unxz',nfile + '.xz'])
        foundit = True
    else:
        print('go pick up the navfile')
        navstatus = navfile_retrieve(navname, cyyyy,cyy,cdoy) 
        if navstatus:
            print('mv the navfile')
            subprocess.call(['mv',navname, navdir])
            foundit = True
        else:
            print('no navfile')

    return navname,navdir,foundit

def getsp3file(year,month,day):
    """
    author: kristine larson
    retrieves IGS sp3 precise orbit file from CDDIS
    inputs are year, month, and day 
    modified in 2019 to use wget 
    returns the name of the file and its directory

    20jun14 - add CDDIS secure ftp
    """
    name, fdir = sp3_name(year,month,day,'igs') 
    print(name,fdir)
    cddis = 'ftp://cddis.nasa.gov'


    # try new way using secure ftp
    #dir_secure = '/pub/gnss/data/daily/' + cyyyy + '/' + cdoy + '/' + cyy + 'o/'
    #file_secure = file1

    if (os.path.isfile(fdir + '/' + name ) == True):
        print('sp3file already exists')
    else:
        gps_week = name[3:7]
        file1 = name + '.Z'
        filename1 = '/gnss/products/' + str(gps_week) + '/' + file1
        url = cddis + filename1 
        # new secure ftp way
        sec_dir = '/gnss/products/' + str(gps_week) + '/'
        sec_file = file1
        try:
            #wget.download(url,file1)
            cddis_download(sec_file, sec_dir) 
            subprocess.call(['uncompress',file1])
            store_orbitfile(name,year,'sp3') 
        except:
            print('some kind of problem-remove empty file')
            subprocess.call(['rm',file1])

#   return the name of the file so that if you want to store it
    return name, fdir

def getsp3file_flex(year,month,day,pCtr):
    """
    author: kristine larson
    retrieves sp3 orbit files from CDDIS
    inputs are year, month, and day  (integers), and 
    pCtr, the processing center  (3 characters)
    returns the name of the file and its directory
    20apr15 check for xz compression

    20jun14 add CDDIS secure ftp

    unfortunately this won't work with the long sp3 file names. use mgex instead
    """
    # returns name and the directory
    name, fdir = sp3_name(year,month,day,pCtr) 
    print(name, fdir)
    gps_week = name[3:7]
    file1 = pCtr + name[3:8] + '.sp3.Z'
    name = pCtr + name[3:8] + '.sp3'
    foundit = False
    ofile = fdir + '/' + name
    # new CDDIS way
    sec_dir = '/gnss/products/' + str(gps_week) + '/'
    sec_file = file1

    if (os.path.isfile(ofile ) == True):
        print('sp3file already exists online')
        foundit = True
    elif (os.path.isfile(ofile + '.xz') == True):
        print('xz compressed sp3file already exists online')
        subprocess.call(['unxz', ofile + '.xz'])
        foundit = True
    else:
        filename1 = '/gnss/products/' + str(gps_week) + '/' + file1
        cddis = 'ftp://cddis.nasa.gov'
        url = cddis + filename1 
        try:
            cddis_download(sec_file, sec_dir) 
            #wget.download(url,file1)
            subprocess.call(['uncompress',file1])
            store_orbitfile(name,year,'sp3') 
            foundit = True
        except:
            print('some kind of problem-remove empty file, if it exists')
            subprocess.call(['rm','-f',file1])
#   return the name of the file so that if you want to store it
    return name, fdir, foundit

def getsp3file_mgex(year,month,day,pCtr):
    """
    author: kristine larson
    retrieves MGEX sp3 orbit files 
    inputs are year, month, and day  (integers), and 
    pCtr, the processing center  (3 characters)
    right now it checks for the "new" name, but in reality, it 
    assumes you are going to use the GFZ product
    20apr15  check for xz compression

    20jun14 add CDDIS secure ftp. what a nightmare
    20jun25 add Shanghai GNSS orbits - as they are available sooner than GFZ
    20jun25 added French and JAXA orbits
    20jul01 allow year, doy as input instead of year, month, day
    20jul10 allow Wuhan, but only one of them.
    """
    foundit = False
    # this returns sp3 orbit product name
    if day == 0:
        print('assume you were given doy in the month input')
        doy = month
        year,month,day= ydoy2ymd(year,doy)
    name, fdir = sp3_name(year,month,day,pCtr) 
    gps_week = name[3:7]
    igps_week = int(gps_week)
    print('GPS week', gps_week,igps_week)
    file1 = name + '.Z'

    # get the sp3 filename for the new format
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    if pCtr == 'gbm': # GFZ
        file2 = 'GFZ0MGXRAP_' + cyyyy + cdoy + '0000_01D_05M_ORB.SP3.gz'
    if pCtr == 'wum': # Wuhan, but only the one at midnite(they are updated more frequently)
        print('use previous day for Wuhan ultra products since gnssSNR does not allow two day files')
        nyear, ndoy = nextdoy(year,doy)
        cndoy = '{:03d}'.format(ndoy); cnyear = '{:03d}'.format(nyear)
        file2 = 'WUM0MGXULA_' + cnyear + cndoy + '0000_01D_05M_ORB.SP3.gz'
    if pCtr == 'grg': # french group
        file2 = 'GRG0MGXFIN_' + cyyyy + cdoy + '0000_01D_15M_ORB.SP3.gz'
    if pCtr == 'sha': # shanghai observatory
        file2 = 'SHA0MGXRAP_' + cyyyy + cdoy + '0000_01D_05M_ORB.SP3.gz'
    # try out JAXA - should have GPS and glonass
    if pCtr == 'jax':
        file2 = 'JAX0MGXFIN_' + cyyyy + cdoy + '0000_01D_05M_ORB.SP3.gz'
    name2 = file2[:-3] 

    # where the files used to live at CDDIS
    #cddis = 'ftp://cddis.nasa.gov'
    dirlocation = '/gps/products/mgex/' + str(gps_week) + '/'
    #url = cddis + dirlocation  + file1; 
    #url2 = cddis + dirlocation + file2; 
    # this is the default setting - no file exists
    mgex = 0
    n1 = os.path.isfile(fdir + '/' + name)
    n1c = os.path.isfile(fdir + '/' + name + '.xz')
    if (n1 == True):
        print('first kind of MGEX sp3file already exists online')
        mgex = 1
        foundit = True
    elif (n1c  == True): 
        print('xz first kind of MGEX sp3file already exists online-unxz it')
        fx =  fdir + '/' + name + '.xz'
        subprocess.call(['unxz', fx])
        mgex = 1
        foundit = True

    n2 = os.path.isfile(fdir + '/' + name2)
    n2c = os.path.isfile(fdir + '/' + name2 + '.xz')
    if (n2 == True):
        print('second kind of MGEX sp3file already exists online')
        mgex = 2 ; foundit = True
    elif (n2c == True):
        print('xz second kind of MGEX sp3file already exists online')
        mgex = 2 ; foundit = True
        fx =  fdir + '/' + name2 + '.xz'
        subprocess.call(['unxz', fx])

    if (mgex == 2):
        name = name2
    if (mgex == 1):
        name = file1[:-2]
# there has to be a better way ... but for now  this works
# only try to download if neither exists

# new using secure ftp
# this is the directory
    secure_dir = '/gps/products/mgex/' + str(gps_week) + '/'
    if (mgex == 0):
        secure_file = file1
        name = file1[:-2]
        print('first file input',secure_file)
        
#https://cddis.nasa.gov/Data_and_Derived_Products/GNSS/gnss_mgex_products.html
# CDDIS claims the old way stopped at GPS week 1945

        if (igps_week < 1944):
            try:
            # secure filename # 1??
                cddis_download(secure_file, secure_dir)
                if os.path.isfile(file1):
                    subprocess.call(['uncompress', file1])
                    store_orbitfile(name,year,'sp3') ; foundit = True
            except:
                print('did not find', file1)
        else:
            print('will only use the long orbit name for weeks after 1944')
        if not foundit:
            name = file2[:-3]
            # new secure filename 
            secure_file = file2
            print('next secure file to check',secure_file)
            try:
                cddis_download(secure_file, secure_dir)
                if os.path.isfile(secure_file):
                    subprocess.call(['gunzip', file2])
                    store_orbitfile(name,year,'sp3') 
                    foundit = True
            except:
                print('did not find',file2)


    return name, fdir, foundit


def kgpsweek(year, month, day, hour, minute, second):
    """
    inputs are year (4 char), month, day, hour, minute, second
    outputs: gps week and second of the week
    author: kristine larson
    modified from a matlab code
    """

    year = np.int(year)
    M = np.int(month)
    D = np.int(day)
    H = np.int(hour)
    minute = np.int(minute)
    
    UT=H+minute/60.0 + second/3600. 
    if M > 2:
        y=year
        m=M
    else:
        y=year-1
        m=M+12
        
    JD=np.floor(365.25*y) + np.floor(30.6001*(m+1)) + D + (UT/24.0) + 1720981.5
    GPS_wk=np.floor((JD-2444244.5)/7.0);
    GPS_wk = np.int(GPS_wk)
    GPS_sec_wk=np.rint( ( ((JD-2444244.5)/7)-GPS_wk)*7*24*3600)            
     
    return GPS_wk, GPS_sec_wk
def kgpsweekC(z):
    """
    takes in time tag from a RINEX file and converts to gps week/sec
    so the input is a character string of length 26.  in this 
    kind of string, the year is only two characters
    author: kristine larson
    """
    y= np.int(z[1:3])
    m = np.int(z[4:6])
    d=np.int(z[7:9])
    hr=np.int(z[10:12])
    mi=np.int(z[13:15])
    sec=np.float(z[16:26])
    gpsw,gpss = kgpsweek(y+2000,m,d,hr,mi,sec)
    return gpsw, gpss

def igsname(year,month,day):
    """
    take in year, month, day
    returns IGS sp3 filename and COD clockname (5 sec)
    author: kristine larson
    """
    [wk,sec]=kgpsweek(year,month,day,0,0,0)
    x=int(sec/86400)
    dd = str(wk) + str(x) 
    name = 'igs' + str(wk) + str(x) + '.sp3'
    # i think at some point they changed to lower case?
   # clockname = 'COD' + dd + '.CLK_05S.txt'
    clockname = 'cod' + dd + '.clk_05s'

    return name, clockname
def read_sp3(file):
    """
    borrowed from Ryan Hardy, who got it from David Wiese
    """
    try:      
        f = open(file)
        raw = f.read()
        f.close()
        lines  = raw.splitlines()
        nprn = np.int(lines[2].split()[1])
        lines  = raw.splitlines()[22:-1]
        epochs = lines[::(nprn+1)]
        nepoch =  len(lines[::(nprn+1)])
        week, tow, x, y, z, clock, prn = np.zeros((nepoch*nprn, 7)).T
        for i in range(nepoch):
            year, month, day, hour, minute, second = np.array(epochs[i].split()[1:], dtype=float)
            week[i*nprn:(i+1)*nprn], tow[i*nprn:(i+1)*nprn] = \
				kgpsweek(year, month, day, hour, minute, second)
            for j in range(nprn):
                prn[i*nprn+j] =  int(lines[i*(nprn+1)+j+1][2:4])
                x[i*nprn+j] = np.float(lines[i*(nprn+1)+j+1][4:18])
                y[i*nprn+j] = np.float(lines[i*(nprn+1)+j+1][18:32])
                z[i*nprn+j] = np.float(lines[i*(nprn+1)+j+1][32:46])
                clock[i*nprn+j] = np.float(lines[(i)*(nprn+1)+j+1][46:60])
    except:
        print('sorry - the sp3file does not exist')
        week,tow,x,y,z,prn,clock=[0,0,0,0,0,0,0]
		
    return week, tow, prn, x, y, z, clock

def myreadnav(file):
    """
    input is navfile name
    output is complicated - broadcast ephemeris blocks
    author: Kristine Larson, April 2017
    """
# input is the nav file
    try:
        f = open(file, 'r')
        nav = f.read()
        f.close()
        nephem = (len(nav.split('END OF HEADER')[1].splitlines())-1)/8
        nephem = int(nephem) #    print(nephem)         
        lines = nav.split('END OF HEADER')[1].splitlines()[1:]
        table = np.zeros((nephem, 32))
        print('Total number of ephemeris messages',nephem)
        for i in range(nephem):
            for j in range(8):
                if j == 0:
                    prn = int(lines[i*8+j][:2])
                    year = int(lines[i*8+j].split()[1])
                    if year > 76:
                        year += 1900
                    else:
                        year += 2000
                    month = int(lines[i*8+j].split()[2])
                    day = int(lines[i*8+j].split()[3])
                    hour = int(lines[i*8+j].split()[4])
                    minute = int(lines[i*8+j].split()[5])
                    second = float(lines[i*8+j][17:22])
                    table[i, 0] = prn
#                    print('Ephem for: ', prn, year, month, day, hour, minute)
                    week, Toc = kgpsweek(year, month, day, hour, minute, second)
                    table[i, 1] =  week
                    table[i, 2] = Toc
                    Af0 = np.float(lines[i*8][-3*19:-2*19].replace('D', 'E'))
                    Af1 = np.float(lines[i*8][-2*19:-1*19].replace('D', 'E'))
                    Af2 = np.float(lines[i*8][-19:].replace('D', 'E'))
                    table[i,3:6] = Af0, Af1, Af2
                elif j != 7:
                    for k in range(4):
                        value = np.float(lines[i*8+j][19*k+3:19*(k+1)+3].replace('D', 'E'))
                        table[i,2+4*j+k] = value
                elif j== 7:
                    table[i,-2]= np.float(lines[i*8+j][3:19+3].replace('D', 'E'))
                    if not lines[i*8+7][22:].replace('D', 'E').isalpha():
                        table[i,-1]= 0
                    else:
                        table[i, -1] = np.float(lines[i*8+7][22:41].replace('D', 'E'))
# output is stored as:
#
# 0-10   prn, week, Toc, Af0, Af1, Af2, IODE, Crs, delta_n, M0, Cuc,\
# 11-22    ecc, Cus, sqrta, Toe, Cic, Loa, Cis, incl, Crc, perigee, radot, idot,\
# 23-24?                   l2c, week, l2f, sigma, health, Tgd, IODC, Tob, interval 
# week would be 24 by this scheme?
# Toe would be 14 
#	
        ephem = table
    except:
        print('This ephemeris file does not exist',file)
        ephem = []
    return ephem
def myfindephem(week, sweek, ephem, prn):
    """
# inputs are gps week, seconds of week
# ephemerides and PRN number
# returns the closest ephemeris block after the epoch
# if one does not exist, returns the first one    
    author: kristine larson
"""
    t = week*86400*7+sweek
# defines the TOE in all the ephemeris 
# he is taking the week and adding ToE (14?)
# poorly coded is all i'm gonna say

    teph = ephem[:, 24]*86400*7+ephem[:, 14]     
    prnmask = np.where(ephem[:, 0]== prn)    

    [nr,nc]=np.shape(prnmask)
#    print(nr,nc)
    if nc == 0:
        print('no ephemeris for that PRN number')
        closest_ephem = []
    else:
        try:          
            signmask = np.where(t >= teph[prnmask])
            proxmask =  np.argmin(t-teph[prnmask][signmask])
            closest_ephem = ephem[prnmask][signmask][proxmask]
        except:
#           print('using first ephemeris - but not after epoch')
            closest_ephem = ephem[prnmask][0]
        
  
    return closest_ephem

def findConstell(cc):
    """
    input is one character (from rinex satellite line)
    output is integer added to the satellite number
    0 for GPS, 100 for Glonass, 200 for Galileo, 300 for everything else?
    author: kristine larson, GFZ, April 2017
    """
    if (cc == 'G' or cc == ' '):
        out = 0
    elif (cc == 'R'): # glonass
        out = 100
    elif (cc == 'E'): # galileo
        out = 200
    else:
        out = 300
        
    return out
def myscan(rinexfile):
    """
    stripping the header code came from pyrinex.  
    data are stored into a variable called table
    columns 0,1,2 are PRN, GPS week, GPS seconds, and observables
    rows are the different observations. these should be stored 
    properly - this is a kluge
    """
    f=open(rinexfile,'r')
    lines = f.read().splitlines(True)
    lines.append('')
    # setting up a set or directionary
    # sets must be unique - so that is hwy he checks to see if it already exists
    header={}        
    eoh=0
# looks like it reads all the header lines, so you can extract them as you like
    for i,line in enumerate(lines):
        if "END OF HEADER" in line:
            eoh=i
            break
#        print(line[60:].strip())
        if line[60:].strip() not in header:
            header[line[60:].strip()] = line[:60].strip()
        else:
            header[line[60:].strip()] += " "+line[:60].strip()
    
    header['APPROX POSITION XYZ'] = [float(i) for i in header['APPROX POSITION XYZ'].split()]
    w = header['APPROX POSITION XYZ']
#    print(w)
#    approxpos = [float(i) for i in header['APPROX POSITION XYZ'].split()]
    header['# / TYPES OF OBSERV'] = header['# / TYPES OF OBSERV'].split()
#    typesObs = header['# / TYPES OF OBSERV'].split()
    aa=header['# / TYPES OF OBSERV']
    types = aa[1:] # this means from element 1 to the end
    # these are from the pyrinex verison of hte code
    header['# / TYPES OF OBSERV'][0] = int(header['# / TYPES OF OBSERV'][0])
    header['INTERVAL'] = float(header['INTERVAL'])
    # need to get approx position of the receiver
    x,y,z = header['APPROX POSITION XYZ']
    numobs = int(np.ceil(header['# / TYPES OF OBSERV'][0]))

# initial three columns in the newheader variable
    newheader = 'PRN\tWEEK\tTOW'
# add in the observation types from this file
    for j in range(numobs):
        newheader += '\t'+types[j]
# header using the Ryan Hardy style
#    print(newheader)

#    # set the line reader to after the end of the header
    # this tells it where to start
    i=eoh+1
    # so try to implement ryan hardy's storing procedure, where 0 is prn,
    # 1 is week, 2 is seconds of week, 3-N are the observables
    table = np.zeros((0, numobs+3))
    print('number of observables ', numobs)
    if numobs > 10:
        print('Tooooo many observables. I cannot deal with this')
        return
    print('line number ' , eoh)
    l = 0 # start counter for blocks
    while True:
        if not lines[i]: break
        if not int(lines[i][28]):
#            print(lines[i])
#            z=lines
            [gw,gs]=kgpsweekC(lines[i])
#            print('week and sec', gw,gs)    
            numsvs = int(lines[i][30:32])  # Number of visible satellites at epoch
#            print('number of satellites',numsvs)
            # strictly speaking i don't understand this line.
            table = np.append(table, np.zeros((numsvs, numobs+3)), axis=0)
            table[l:l+numsvs, 1] = gw
            table[l:l+numsvs, 2] = gs
            #headlength.append(1 + numsvs//12)
            sp = []
                
            if(numsvs>12):
                for s in range(numsvs):
                    xv = findConstell(lines[i][32+(s%12)*3:33+(s%12)*3]) 
                    sp.append(xv + int(lines[i][33+(s%12)*3:35+(s%12)*3]))
                    if s>0 and s%12 == 0:
                        i+= 1  # For every 12th satellite there will be a new row with satellite names                sats.append(sp) # attach satellites here
            else:
                for s in range(numsvs):
                    xv = findConstell(lines[i][32+(s%12)*3:33+(s%12)*3])
                    sp.append(xv + int(lines[i][33+(s%12)*3:35+(s%12)*3]))

#            print(len(sp), 'satellites in this block', sp)
            for k in range(numsvs): 
                table[l+k,0]=sp[k]
                if (numobs > 5):
                    for d in range(5):
                        gg = d*16
                        f=lines[i+1+2*k][gg:gg+14]
                        if not(f == '' or f.isspace()):
                            val = np.float(lines[i+1+2*k][gg:gg+14])
                            table[l+k, 3+d] = val
                    for d in range(numobs-5):
                        gg = d*16
                        f=lines[i+2+2*k][gg:gg+14]
                        if not (f == '' or f.isspace()):
                            val = np.float(lines[i+2+2*k][gg:gg+14])
                            table[l+k, 3+5+d] = val
                else:
                    for d in range(numobs):
                        gg = d*16
                        f = lines[i+1+2*k][gg:gg+14]
                        if (f == '' or f.isspace()):
                            val = np.float(lines[i+2+2*k][gg:gg+14])
                            table[l+k, 3+d] = val

            i+=numsvs*int(np.ceil(header['# / TYPES OF OBSERV'][0]/5))+1
            l+=numsvs
        else:
            print('there was a comment or some header info in the rinex')
            flag=int(lines[i][28])
            if(flag!=4):
                print('this is a flag that is not 4', flag)
            skip=int(lines[i][30:32])
            print('skip this many lines',skip)
            i+=skip+1
            print('You are now on line number',i)
    [nr,nc]=np.shape(table)
    print('size of the table variable is ',nr, ' by ', nc)    
    # code provided by ryan hardy - but not needed???
#    format = tuple(np.concatenate((('%02i', '%4i', '%11.7f'), 
#							((0, numobs+3)[-1]-3)*['%14.3f'])))
# I think this takes the newheader and combines it with the information in
# the table variable
    obs =  dict(zip(tuple(newheader.split('\t')), table.T))
    return obs,x,y,z

def geometric_rangePO(secweek, prn, rrec0, sweek, ssec, sprn, sx, sy, sz, sclock):
    """
    Calculates and returns geometric range (in metres) given
    time (week and sec of week), prn, Cartesisan 
    receiver coordinates rrec0(meters)
    using the precise ephemeris instead of the broadcast
    returns clock correction (clockC) in meters
    Author: Kristine Larson, May 2017
    June 21, 2017 returns transmit time (in seconds) so I can calculate
    relatistic correction
    """
    error = 1
    # find the correct sp3 data for prn
    m = [sprn==prn]
    nx,ny,nz,nc = sp3_interpolator(secweek, ssec[m], sx[m], sy[m], sz[m], sclock[m]) 
    SatOrb = np.array([nx[0],ny[0],nz[0]])
    geo=norm(SatOrb-rrec0)
    c=constants.c
    clockC = nc*1e-6*c
    oE = constants.omegaEarth
    deltaT = norm(SatOrb - rrec0)/constants.c
    ij=0
    while error > 1e-16:  
        nx,ny,nz,nc = sp3_interpolator(secweek-deltaT, ssec[m], sx[m], sy[m], sz[m], sclock[m]) 
        SatOrb = np.array([nx[0],ny[0],nz[0]])
#        SatOrb, relcorr = propagate(week, secweek-deltaT, closest_ephem)
        Th = -oE * deltaT
        xs = SatOrb[0]*np.cos(Th)-SatOrb[1]*np.sin(Th)
        ys = SatOrb[0]*np.sin(Th)+SatOrb[1]*np.cos(Th)
        SatOrbn = [xs, ys, SatOrb[2]]
        geo=norm(SatOrbn-rrec0)
        deltaT_new = norm(SatOrbn-rrec0)/constants.c               
        error = np.abs(deltaT - deltaT_new)
        deltaT = deltaT_new
        ij+=1
    #    print(ij)
    return geo,SatOrbn,clockC, deltaT
def read_files(year,month,day,station):
    """
    """   

    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    # i have a function for this ....
    rinexfile = station + cdoy + '0.' + cyy + 'o'
    navfilename = 'auto'  + cdoy + '0.' + cyy +  'n'
    if os.path.isfile(rinexfile):
        print('rinexfile exists')
    else:
        print(rinexfile)
        print('get the rinex file')
        rinex_unavco(station, year, month, day)
    # organize the file names
    print('get the sp3 and clock file names')
    sp3file, cname = igsname(year,month,day)

    # define some names of files
    if os.path.isfile(navfilename):
        print('nav exists')
    else:
        print('get nav')
        navname,navdir,foundit = getnavfile(year,month,day)
    print('read in the broadcast ephemeris')
    ephemdata = myreadnav(navfilename)
    if os.path.isfile(cname):
        print('file exists')
    else:        
        print('get the CODE clock file')
        codclock(year,month,day)
    pname = cname[0:9] + 'pckl'
    print('pickle', pname)
    # if file exists already     
    if os.path.isfile(pname):
        print('read existing pickle file')
        f = open(pname, 'rb')
        [prns,ts,clks] = pickle.load(f)
        f.close()
    else:
        print('read and save as pickle')
        prns, ts, clks = readPreciseClock(cname)
        # and then save them
        f = open(pname, 'wb')
        pickle.dump([prns,ts,clks], f)
        f.close()
    if os.path.isfile(sp3file):
        print('sp3 exsts')
    else:
        print('get sp3')
        getsp3file(year,month,day)
    print('read in the sp3 file', sp3file)
    
    sweek, ssec, sprn, sx, sy, sz, sclock = read_sp3(sp3file)
    
#20    print('len returned data', len(ephemdata), navfilename
    rinexpickle = rinexfile[0:11] + 'pclk'
    if os.path.isfile(rinexpickle):
        print('rinex pickle exists')
        f=open(rinexpickle,'rb')
        [obs,x,y,z]=pickle.load(f)
        f.close()
    else:     
        print('read the RINEX file ', rinexfile)
        obs,x,y,z = myscan(rinexfile)
        print('save as pickle file')
        f=open(rinexpickle,'wb')
        pickle.dump([obs,x,y,z], f)
        f.close()
        
    return ephemdata, prns, ts, clks, sweek, ssec, sprn, sx, sy,sz,sclock,obs,x,y,z
def precise_clock(prn,prns,ts,clks,gps_seconds):
    """
    input prn and gps_seconds, and contents of precise clock file,
    i think.
    units of returned variable are?
    """
    m = [prns == prn]
    preciset = ts[m]
    precisec = clks[m]
#           # trying to interpolate
# could do this much faster ...
    clockf = interp1d(preciset, precisec, bounds_error=False)
    scprecise = clockf(gps_seconds)
    return scprecise
def precise_clock_test(ttags, prns,ts,clks):
    """
    input timetags (gps seconds of the week)
    and precise clock values.returns interpolated values
    units of returned variable are?
    """
    NP = len(ttags)
    print('number of time tags', NP)
    newclks = np.zeros((NP, 32))
    for i in range(32): 
        prn = i+1
 #        print('checking', prn)
        m = [prns == prn]
        [nr,nc]=np.shape(m)
 #       print(nr,nc)
        preciset = ts[m]
        precisec = clks[m]
        clockf = interp1d(preciset, precisec, bounds_error=False)
        k=0
        for t in ttags:
            newclks[k,i]= clockf(t)
            k +=1
#        print(prn, numrows, numcols)
    return newclks

def propagate(week, sec_of_week, ephem):
    """
    inputs are GPS week, seconds of the week, and the appropriate 
    ephemeris block from the navigation message
    returns the x,y,z, coordinates of the satellite 
    and relativity correction (also in meters), so you add,
    not subtract
    Kristine Larson, April 2017

    """

# redefine the ephem variable
    prn, week, Toc, Af0, Af1, Af2, IODE, Crs, delta_n, M0, Cuc,\
    ecc, Cus, sqrta, Toe, Cic, Loa, Cis, incl, Crc, perigee, radot, idot,\
    l2c, week, l2f, sigma, health, Tgd, IODC, Tob, interval = ephem
    sweek = sec_of_week
    # semi-major axis
    a = sqrta**2
    t = week*7*86400+sweek
    tk = t-Toe
    # no idea if Ryan Hardy is doing this correctly - it should be in a function
    tk  =  (tk - 302400) % (302400*2) - 302400
    n0 = np.sqrt(constants.mu/a**3)
    n = n0+ delta_n
    Mk = M0 + n*tk
    i = 0
    Ek = Mk
    E0 = Mk + ecc*np.sin(Mk)
    # solve kepler's equation
    while(i < 15 or np.abs(Ek-E0) > 1e-12):
        i +=1
        Ek = Mk + ecc*np.sin(E0) 
        E0 = Mk + ecc*np.sin(Ek)
    nuk = np.arctan2(np.sqrt(1-ecc**2)*np.sin(Ek),np.cos(Ek)-ecc)
    Phik = nuk + perigee
    duk = Cus*np.sin(2*Phik)+Cuc*np.cos(2*Phik)
    drk = Crs*np.sin(2*Phik)+Crc*np.cos(2*Phik)
    dik = Cis*np.sin(2*Phik)+Cic*np.cos(2*Phik)
    uk = Phik + duk
    rk = a*(1-ecc*np.cos(Ek))+drk
       
    ik = incl+dik+idot*tk
    xkp = rk*np.cos(uk)
    ykp = rk*np.sin(uk)
    Omegak = Loa + (radot-constants.omegaEarth)*tk -constants.omegaEarth*Toe
    xk = xkp*np.cos(Omegak)-ykp*np.cos(ik)*np.sin(Omegak)
    yk = xkp*np.sin(Omegak)+ykp*np.cos(ik)*np.cos(Omegak)
    zk = ykp*np.sin(ik) 
    # using class
    F = -2*np.sqrt(constants.mu)/constants.c
    relcorr = F*ecc*sqrta*np.sin(Ek)
#    return [xk, yk, zk], relcorr
    return [xk[0], yk[0], zk[0]], relcorr

def mygeometric_range(week, secweek, prn, rrec0, closest_ephem):
    """
    Calculates and returns geometric range (in metres) given
    time (week and sec of week), prn, receiver coordinates (cartesian, meters)
    this assumes someone was nice enough to send you the closest ephemeris
    returns the satellite coordinates as well, so you can use htem
    in the A matrix
    Kristine Larson, April 2017
    """
    error = 1

    SatOrb, relcorr = propagate(week, secweek, closest_ephem)
    # first estimate of the geometric range
    geo=norm(SatOrb-rrec0)

    deltaT = norm(SatOrb - rrec0)/constants.c
    while error > 1e-16:     
        SatOrb, relcorr = propagate(week, secweek-deltaT, closest_ephem)
        Th = -constants.omegaEarth * deltaT
        xs = SatOrb[0]*np.cos(Th)-SatOrb[1]*np.sin(Th)
        ys = SatOrb[0]*np.sin(Th)+SatOrb[1]*np.cos(Th)
        SatOrbn = [xs, ys, SatOrb[2]]
        geo=norm(SatOrbn-rrec0)
        deltaT_new = norm(SatOrbn-rrec0)/constants.c               
        error = np.abs(deltaT - deltaT_new)
        deltaT = deltaT_new
    return geo,SatOrbn


def readobs(file,nepochX):
    """
    inputs: filename and number of epochs from the RINEX file you want to unpack
    returns: receiver Cartesian coordinates (meters) and observation blocks
    Kristine Larson, April 2017
    18aug20: updated so that 0,0,0 is returned if there are no receiver coordinates
    """
    f = open(file, 'r')
    obs = f.read()
    f.close()
    testV = obs.split('APPROX POSITION XYZ')[0].split('\n')[-1].split()
#   set default receiver location values,              
    x0=0
    y0=0
    z0=0
    if len(testV) == 3:
        x0, y0, z0 = np.array(obs.split('APPROX POSITION XYZ')[0].split('\n')[-1].split(), 
    
							dtype=float)
    types = obs.split('# / TYPES OF OBSERV')[0].split('\n')[-1].split()[1:]
    print(types)
    tfirst =  obs.split('TIME OF FIRST OBS')[0].split('\n')[-1].split()
    print(tfirst)
    ntypes = len(types)
	#Identify unique epochs
    countstr = '\n '+tfirst[0][2:]+' '+'%2i' % np.int(tfirst[1])+' '+'%2i' % np.int(tfirst[2])
    print(countstr) # so this is 07 10 13, e.g.
    # figures out how many times this string appears between END OF HEADER and end of file
    nepoch = obs.split('END OF HEADER')[1].count(countstr)
    
    table = np.zeros((0, ntypes+3))
	#Identify kinds of observations in the file
    header = 'PRN\tWEEK\tTOW'
    t0 = 0
    for i in range(ntypes):
        header += '\t'+types[i]
    l = 0
#    print('countstr',countstr[0])
    print('header', header)
    epochstr_master = re.split(countstr, obs)
    print('number of data epochs in the file', nepoch)
    # only unpack a limited number of epochs    #nepoch = 5
    print('restricted number of epochs to be read ', nepochX)
    for i in range(nepochX):
        epochstr = epochstr_master[i+1]        
        # this will only work with new rinex files that use G as descriptor
# this only finds GPS!
        prnstr = re.findall(r'G\d\d|G \d', epochstr.split('\n')[0])
        print(prnstr)
#        print('prnstr', prnstr)
        # number of satellites for an observation block
        # this code won't work if there are more than 12 satellites (i think)
        nprn = len(prnstr)
        # append
        table = np.append(table, np.zeros((nprn, ntypes+3)), axis=0)
        # decode year, month, day hour, minute ,second
        # convert 2 char to 4 char year
        # but it appears that the year nad month and day 
        # are being hardwired from the header, which is both  very very odd
        # and stupid
        year = int(tfirst[0])
        month = int(tfirst[1])
        day = int(tfirst[2])
        hour = int(epochstr.split()[0])
        minute = int(epochstr.split()[1])
        second = float(epochstr.split()[2])
        print(year, month, day, hour, minute, second)
        week, tow = kgpsweek(year, month, day, hour, minute, second)
        print('reading week number ', week, 'sec of week', tow)
        table[l:l+nprn, 1] = week
        table[l:l+nprn,2] = tow
        for j in range(nprn):
            table[l+j, 0] = np.int(prnstr[j][1:])
            # split up by end of line markers, which makes sense
#            print('something',2*j+1)
            line0 = epochstr.split('\n')[2*j+1]#.split()
            line = re.findall('.{%s}' % 16, line0+' '*(80-len(line0)))
# this seems to get strings of a certain width, which he will later change 
#through float# also where he made his mistake on reading last two characters 
#            print(j,len(line), line)

            for k in range(len(line)):
                if line[k].isspace():
                    table[l+j, 3+k] = np.nan
                    continue
                table[l+j, 3+k] = np.float(line[k][:-2])
#                print(table[l+j,3+k])
                # if more than 5 observaitons, he has to read the next line
            if ntypes > 5:
                line2 = epochstr.split('\n')[2*j+2].split()
#                print(line2)
                for k in range(len(line2)):
                    table[l+j, 3+len(line)+k] = np.float(line2[k][:-2])
        l += nprn	
    format = tuple(np.concatenate((('%02i', '%4i', '%11.7f'), 
							((0, ntypes+3)[-1]-3)*['%14.3f'])))
    # kinda strange - but ok - format statemenst for PRN, week, sec of week, etc
#    print(format)
# I guess it is making a super observable so that it can later be parsed.
# very strange 
    obs =  dict(zip(tuple(header.split('\t')), table.T))

    return np.array([x0,y0,z0]),obs

def tmpsoln(navfilename,obsfilename):
    """
    kristine larson
    inputs are navfile and obsfile names
    should compute a pseudorange solution
    """
    r2d = 180.0/np.pi
#  elevation mask
    emask = 10 
 
    ephemdata = myreadnav(navfilename)
    if len(ephemdata) == 0:
        print("empty ephmeris or does not exist")
        return
    #cartesian coordinates - from the header
    # number of epochs you want to read
    nep = 2
    recv,obs=readobs2(obsfilename,nep)
    print('A priori receiver coordinates', recv)
    if recv[0] == 0.0:
        print('using new a priori')
        recv[0]=-2715532
        recv[1] = -881995
        recv[2] = 5684286
        print('now using', recv)
    lat, lon, h = xyz2llh(recv,1e-8)
    print("%15.7f %15.7f"% (lat*r2d,  lon*r2d) )
    # zenith delay - meters
    zd = zenithdelay(h)
    u,east,north = up(lat,lon)
    epoch = np.unique(obs['TOW'])
    NN=len(epoch)
    NN = 2 # only do two positions
    for j in range(NN):
        print('Epoch', j+1)
        sats = obs['PRN'][np.where(epoch[j]==obs['TOW'])]
        gps_seconds = np.unique(obs['TOW'][np.where(epoch[j]==obs['TOW'])])
        gps_weeks = np.unique(obs['WEEK'][np.where(epoch[j]==obs['TOW'])])
        # not sure this does anything
        gps_weeks = gps_weeks.tolist()
        gps_seconds = gps_seconds.tolist()

        P1 = obs['C1'][np.where(epoch[j]==obs['TOW'])]
        P2 = obs['P2'][np.where(epoch[j]==obs['TOW'])]
#       do not know why this is here
#       S1 = obs['S1'][np.where(epoch[j]==obs['TOW'])]
        print('WEEK', gps_weeks, 'SECONDS', gps_seconds)
        k=0
        # set up the A matrix with empty stuff
        M=len(sats)
        print('Number of Satellites: ' , M)
        A=np.zeros((M,4))
        Y=np.zeros(M)
        elmask = np.zeros(M, dtype=bool)
        for prn in sats:
            closest = myfindephem(gps_weeks, gps_seconds, ephemdata, prn) #           
            p1=P1[k]
            p2=P2[k]
            p3 = ionofree(p1,p2)
            N=len(closest)
            if N > 0:    
                satv, relcorr = propagate(gps_weeks, gps_seconds, closest)
#               print("sat coor",gps_weeks,gps_seconds,satv)
                r=np.subtract(satv,recv)
                elea = elev_angle(u, r) # 
                tropocorr = zd/np.sin(elea)
                R,satv = mygeometric_range(gps_weeks, gps_seconds, prn, recv, closest)
                A[k]=-np.array([satv[0]-recv[0],satv[1]-recv[1],satv[2]-recv[2],R])/R
                elea = elev_angle(u,np.subtract(satv,recv))
                elmask[k] = elea*180/np.pi > emask
#               satellite clock correction
                satCorr = satclock(gps_weeks, gps_seconds, prn, closest)
                # prefit residual, ionosphere free pseudorange - geometric rnage
                # plus SatelliteClock - relativity and troposphere corrections
                Y[k] = p3-R+satCorr  -relcorr-tropocorr
#               print(int(prn), k,p1,R, p1-R)
                print(" {0:3.0f} {1:15.4f} {2:15.4f} {3:15.4f} {4:10.5f}".format(prn, p3, R, Y[k], 180*elea/np.pi))
                k +=1
# only vaguest notion of what is going on here - code from Ryan Hardy                
#       applying an elevation mask
        Y=np.matrix(Y[elmask]).T
        A=np.matrix(A[elmask])
        soln = np.array(np.linalg.inv(A.T*A)*A.T*Y).T[0]
#       update Cartesian coordinates
        newPos = recv+soln[:3]
        print('New Cartesian solution', newPos)
#       receiver clock solution
#        rec_clock = soln[-1]
        lat, lon, h = xyz2llhd(newPos)
        print("%15.7f %15.7f %12.4f "% (lat,  lon,h) )
# print("%15.5f"% xyz[0])
  
def readobs2(file,nepochX):
    """
    inputs: filename and number of epochs from the RINEX file you want to unpack
    returns: receiver Cartesian coordinates (meters) and observation blocks
    Kristine Larson, April 2017
    18aug20: updated so that 0,0,0 is returned if there are no receiver coordinates
    18aug21: version to include Glonass satellites etc
    """
    f = open(file, 'r')
    obs = f.read()
    f.close()
    testV = obs.split('APPROX POSITION XYZ')[0].split('\n')[-1].split()
#   set default receiver location values,              
    x0=0
    y0=0
    z0=0
    if len(testV) == 3:
        x0, y0, z0 = np.array(obs.split('APPROX POSITION XYZ')[0].split('\n')[-1].split(), 
							dtype=float)
    types = obs.split('# / TYPES OF OBSERV')[0].split('\n')[-1].split()[1:]
    print(types)
    tfirst =  obs.split('TIME OF FIRST OBS')[0].split('\n')[-1].split()
    print(tfirst)
    ntypes = len(types)
	#Identify unique epochs
    countstr = '\n '+tfirst[0][2:]+' '+'%2i' % np.int(tfirst[1])+' '+'%2i' % np.int(tfirst[2])
    print(countstr) # so this is 07 10 13, e.g.
    # figures out how many times this string appears between END OF HEADER and end of file
    nepoch = obs.split('END OF HEADER')[1].count(countstr)
    
    table = np.zeros((0, ntypes+3))
	#Identify kinds of observations in the file
    header = 'PRN\tWEEK\tTOW'
    t0 = 0
    for i in range(ntypes):
        header += '\t'+types[i]
    l = 0
#    print('countstr',countstr[0])
    print('header', header)
    epochstr_master = re.split(countstr, obs)
    print('number of data epochs in the file', nepoch)
    # only unpack a limited number of epochs    #nepoch = 5
    print('restricted number of epochs to be read ', nepochX)
    for i in range(nepochX):
#       clear satarray
        satarray = []
        epochstr = epochstr_master[i+1]        
        print(epochstr[21:23])
#       this is now the number of satellites properly read from the epochstr variable
        nprn = int(epochstr[21:23])
        print('Epochstr number of satellites',nprn)
        # this will only work with new rinex files that use G as descriptor
        for jj in range(nprn):
            i2 = 26+jj*3
            i1 = i2-3 
            satName = epochstr[i1:i2] 
            print(satName)
            if satName[0] == 'R':
#               found glonass
                list.append(satarray, 100+int(satName[1:3]))
            else:
#               assume rest are GPS for now
                list.append(satarray, int(satName[1:3]))
            print(satarray)
        prnstr = re.findall(r'G\d\d|G \d', epochstr.split('\n')[0])
        
        print(nprn, prnstr)
        if nprn > 12:
           # add the commant to read next line
           print('read next line of satellite names')
        # append
        table = np.append(table, np.zeros((nprn, ntypes+3)), axis=0)
        # decode year, month, day hour, minute ,second
        # convert 2 char to 4 char year
        # but it appears that the year and month and day 
        # are being hardwired from the header, which is both  very very odd
        # and stupid
        year = int(tfirst[0])
        month = int(tfirst[1])
        day = int(tfirst[2])
        hour = int(epochstr.split()[0])
        minute = int(epochstr.split()[1])
        second = float(epochstr.split()[2])
        print(year, month, day, hour, minute, second)
        week, tow = kgpsweek(year, month, day, hour, minute, second)
        print('reading week number ', week, 'sec of week', tow)
#       now you are saving the inforamtion to the table variable
        table[l:l+nprn, 1] = week
        table[l:l+nprn,2] = tow
        for j in range(nprn):
#            store this satellite
            print(j, satarray[j])
#            table[l+j, 0] = np.int(prnstr[j][1:])
# try this - using new definition of observed satellite
            table[l+j, 0] = satarray[j]
            # split up by end of line markers, which makes sense
#            print('something',2*j+1)
            line0 = epochstr.split('\n')[2*j+1]#.split()
            line = re.findall('.{%s}' % 16, line0+' '*(80-len(line0)))
# this seems to get strings of a certain width, which he will later change 
#through float# also where he made his mistake on reading last two characters 
#            print(j,len(line), line)

            for k in range(len(line)):
                if line[k].isspace():
                    table[l+j, 3+k] = np.nan
                    continue
                table[l+j, 3+k] = np.float(line[k][:-2])
#                print(table[l+j,3+k])
                # if more than 5 observaitons, he has to read the next line
            if ntypes > 5:
                line2 = epochstr.split('\n')[2*j+2].split()
#                print(line2)
                for k in range(len(line2)):
                    table[l+j, 3+len(line)+k] = np.float(line2[k][:-2])
        l += nprn	
    format = tuple(np.concatenate((('%02i', '%4i', '%11.7f'), 
							((0, ntypes+3)[-1]-3)*['%14.3f'])))
    # kinda strange - but ok - format statemenst for PRN, week, sec of week, etc
#    print(format)
# I guess it is making a super observable so that it can later be parsed.
# very strange 
    obs =  dict(zip(tuple(header.split('\t')), table.T))

    return np.array([x0,y0,z0]),obs

def get_ofac_hifac(elevAngles, cf, maxH, desiredPrec):
    """
    computes two factors - ofac and hifac - that are inputs to the
    Lomb-Scargle Periodogram code.
    We follow the terminology and discussion from Press et al. (1992)
    in their LSP algorithm description.

    INPUT
    elevAngles:  vector of satellite elevation angles in degrees 
    cf:(L-band wavelength/2 ) in meters    
    maxH:maximum LSP grid frequency in meters
    desiredPrec:  the LSP frequency grid spacing in meters
    i.e. how precise you want he LSP reflector height to be estimated
    OUTPUT
    ofac: oversampling factor
    hifac: high-frequency factor
    """
# in units of inverse meters
    X= np.sin(elevAngles*np.pi/180)/cf     

# number of observations
    N = len(X) 
# observing Window length (or span)
# units of inverse meters
    W = np.max(X) - np.min(X)         

# characteristic peak width, meters
    cpw= 1/W

# oversampling factor
    ofac = cpw/desiredPrec 

# Nyquist frequency if the N observed data samples were evenly spaced
# over the observing window span W, in meters
    fc = N/(2*W)

# Finally, the high-frequency factor is defined relative to fc
    hifac = maxH/fc  

    return ofac, hifac

def strip_compute(x,y,cf,maxH,desiredP,pfitV,minH):
    """
    strips snr data
    inputs; max reflector height, desiredP is desired precision in meters
    pfitV is polynomial fit order
    minH - do not allow LSP below this value
    returns 
    max reflector height and its amplitude
    min and max observed elevation angle
    riseSet is 1 for rise and -1 for set
    author: Kristine Larson
    """
    ofac,hifac = get_ofac_hifac(x,cf,maxH,desiredP)
#   min and max observed elevation angles
    eminObs = min(x); emaxObs = max(x)
    if x[0] > x[1]:
        riseSet = -1
    else:
        riseSet = 1

#   change so everything is rising, i.e. elevation angle is increasing
    ij = np.argsort(x)
    x = x[ij]
    y = y[ij]

    x = np.sin(x*np.pi/180)
#   polynomial fit done before

#   scale by wavelength
    x=x/cf
#    y=newy
#   get frequency spacing
    px = freq_out(x,ofac,hifac) 
#   compute spectrum using scipy
    scipy_LSP = spectral.lombscargle(x, y, 2*np.pi*px)

#   find biggest peak
#   scaling required to get amplitude spectrum
    pz = 2*np.sqrt(scipy_LSP/len(x))
#   now window
#    ij = np.argmax(px > minH)
#    new_px = px[ij]
    new_pz = pz[(px > minH)]
    new_px = px[(px > minH)]
    px = new_px
    pz = new_pz
#   find the max
    ij = np.argmax(pz)
    maxF = px[ij]
    maxAmp = np.max(pz)
    return maxF, maxAmp, eminObs, emaxObs,riseSet, px,pz

def window_data(s1,s2,s5,s6,s7,s8, sat,ele,azi,seconds,edot,f,az1,az2,e1,e2,satNu,pfitV,pele):
    """
    author kristine m. larson
    also calculates the scale factor for various GNNS frequencies.  currently
    returns meanTime in UTC hours and mean azimuth in degrees
    cf, which is the wavelength/2
    currently works for GPS, GLONASS, GALILEO, and Beidou
    new: pele are the elevation angle limits for the polynomial fit. these are appplied
    before you start windowing the data
    """
    cunit = 1
    dat = []; x=[]; y=[]
#   get scale factor
#   added glonass, 101 and 102
    if (f == 1) or (f==101) or (f==201):
        dat = s1
    if (f == 2) or (f == 20) or (f == 102) or (f==302):
        dat = s2
    if (f == 5) or (f==205):
        dat = s5
#   these are galileo frequencies (via RINEX definition)
    if (f == 206) or (f == 306):
        dat = s6
    if (f == 207) or (f == 307):
        dat = s7
    if (f == 208):
        dat = s8
#   get the scaling factor for this frequency and satellite number
#   print(f,satNu)
    cf = arc_scaleF(f,satNu)

#   if not, frequency does not exist, will be tripped by Nv
#   remove the direct signal component
    if (cf > 0):
        x,y,sat,azi,seconds,edot  = removeDC(dat, satNu, sat,ele, pele, azi,az1,az2,edot,seconds) 

#
    Nv = len(y); Nvv = 0 ; 
#   some defaults in case there are no data in this region
    meanTime = 0.0; avgAzim = 0.0; avgEdot = 1; Nvv = 0
    avgEdot_fit =1; delT = 0.0
#   no longer have to look for specific satellites. some minimum number of points required 
    if Nv > 30:
        model = np.polyfit(x,y,pfitV)
        fit = np.polyval(model,x)
#       redefine x and y as old variables
        ele = x
        dat = y - fit
#       ok - now figure out what is within the more restricted elevation angles
        x =   ele[(ele > e1) & (ele < e2) & (azi > az1) & (azi < az2)]
        y =   dat[(ele > e1) & (ele < e2) & (azi > az1) & (azi < az2)]
        ed = edot[(ele > e1) & (ele < e2) & (azi > az1) & (azi < az2)]
        a =   azi[(ele > e1) & (ele < e2) & (azi > az1) & (azi < az2)]
        t = seconds[(ele > e1) & (ele < e2) & (azi > az1) & (azi < az2)]
        sumval = np.sum(y)
        if sumval == 0:
            x = []; y=[] ; Nv = 0 ; Nvv = 0
#   since units were changed to volts/volts, the zeros got changed to 1 values
        if sumval == Nv:
            x = []; y=[] ; Nv = 0 ; Nvv = 0
        Nvv = len(y)
#       calculate average time in UTC (actually it is GPS time) in hours and average azimuth
#       this is fairly arbitrary, but can't be so small you can't fit a polymial to it
        if (Nvv > 10):
            dd = np.diff(t)
#           edot, in radians/sec
            model = np.polyfit(t,x*np.pi/180,1)
#  edot in radians/second
            avgEdot_fit = model[0]
            avgAzim = np.mean(a)
            meanTime = np.mean(t)/3600
            avgEdot = np.mean(ed) 
#  delta Time in minutes
            delT = (np.max(t) - np.min(t))/60 
# average tan(elev)
            cunit =np.mean(np.tan(np.pi*x/180))
#           return tan(e)/edot, in units of radians/hour now. used for RHdot correction
    if avgEdot == 0:
        outFact1 = 0
    else:
        outFact1 = cunit/(avgEdot*3600) 
    outFact2 = cunit/(avgEdot_fit*3600) 
    # was debugging
    #if (satNu == 25) and (f == 20):
    #    print('Edot,cunit,az', avgEdot_fit, cunit,  avgAzim, outFact2)
    return x,y,Nvv,cf,meanTime,avgAzim,outFact1, outFact2, delT

def arc_scaleF(f,satNu):
    """
    input a frequency and put out a scale factor cf which is wavelength*0.5 
    """ 
#   default value for w so that if someone inputs an illegal frequency, it does not crash
    w = 0
    if f == 1:
        w = constants.wL1
    if (f == 2) or (f == 20):
        w = constants.wL2
    if f == 5:
        w = constants.wL5
#   galileo satellites
#   must be a smarter way to do this
    if (f > 200) and (f < 210):
        if (f == 201):
            w = constants.wgL1
        if (f == 205):
            w = constants.wgL5
        if (f == 206):
            w = constants.wgL6
        if (f == 207):
            w = constants.wgL7
        if (f == 208):
            w = constants.wgL8
#
#   add beidou 18oct15
    if (f > 300) and (f < 310):
        if (f == 302):
            w = constants.wbL2
        if (f == 307):
            w = constants.wbL7
        if (f == 306):
            w = constants.wbL6

#   glonass satellite frequencies
    if (f == 101) or (f == 102):
        w = glonass_channels(f,satNu) 
    cf = w/2
    return cf 

def freq_out(x,ofac,hifac):
    """
    inputs: x 
    ofac: oversamping factor
    hifac
    outputs: two sets of frequencies arrays
    """
#
# number of points in input array
    n=len(x)
#
# number of frequencies that will be used
    nout=np.int(0.5*ofac*hifac*n)
	 
    xmax = np.max(x) 
    xmin = np.min(x) 
    xdif=xmax-xmin 
# starting frequency 
    pnow=1.0/(xdif*ofac) 
    pstart = pnow
    pstop = hifac*n/(2*xdif)
# 
# output arrays
#    px = np.zeros(nout)
#    for i in range(0,nout):
#        px[i]=pnow
#        pnow=pnow+1.0/(ofac*xdif)
# simpler way
    pd = np.linspace(pstart, pstop, nout)
    return pd

def find_satlist(f,snrExist):
    """
    inputs: frequency and boolean numpy array that tells you 
    if a signal is (potentially) legal
    outputs: list of satellites to use 
    author: kristine m. larson
    """
# set list of GPS satellites for now
# 
# Block III will be 4, 18, 23
#   these are the only L2C satellites as of 18oct10
    l2c_sat = [1, 3, 5, 6, 7, 8, 9, 10, 12, 15, 17, 24, 25, 26, 27, 29, 30, 31, 32]
    # updated on 20 jul 15 - really should make this time dependent ....
    l2c_sat = [1, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15, 17, 23, 24, 25, 26, 27, 29, 30, 31, 32]

#   only L5 satellites thus far
    l5_sat = [1, 3,  6,  8, 9, 10, 24, 25, 26, 27, 30,  32]
    # l5_sat = [1, 3,  4, 6,  8, 9, 10, 24, 25, 26, 27, 30,  32]
#   assume l1 and l2 can be up to 32
    l1_sat = np.arange(1,33,1)
    satlist = []
    if f == 1:
        satlist = l1_sat
    if f == 20:
        satlist = l2c_sat
    if f == 2:
        satlist = l1_sat
    if f == 5:
        satlist = l5_sat
#   i do not think they have 26 - but ....
#   glonass L1
    if (f == 101) or (f==102):
# only have 24 frequencies defined
        satlist = np.arange(101,125,1)
#   galileo - 40 max?
#   use this to check for existence, mostly driven by whether there are 
#   extra columns (or if they are non-zero)
    gfs = int(f-200)

    if (f >  200) and (f < 210) and (snrExist[gfs]):
        satlist = np.arange(201,241,1)
#   galileo has no L2 frequency, so set that always to zero
    if f == 202:
        satlist = []
#   pretend there are 32 Beidou satellites for now
    if (f > 300):
        satlist = np.arange(301,333,1)

    if len(satlist) == 0:
        print('     illegal frequency: no sat list being returned')
    return satlist

def glonass_channels(f,prn):
    """
    inputs frequency and prn number
    returns wavelength for glonass satellite in meters
    logic from Simon Williams, matlab
    converted to function by KL, 2017 November
    converted to python by KL 2018 September
    """
#   we define glonass by adding 100.  remove this for definition of the wavelength
    if (prn > 100):
        prn = prn - 100
    lightSpeed = 299792458
    slot = [14,15,10,20,19,13,12,1,6,5,22,23,24,16,4,8,3,7,2,18,21,9,17,11]
    channel = [-7,0,-7,2,3,-2,-1,1,-4,1,-3,3,2,-1,6,6,5,5,-4,-3,4,-2,4,0]
    slot = np.matrix(slot)
    channel = np.matrix(channel)
#   main frequencies
    L1 = 1602e6
    L2 = 1246e6
#   deltas
    dL1 = 0.5625e6
    dL2 = 0.4375e6

    ch = channel[(slot == prn)]
    ch = int(ch)
#   wavelengths in meters
#    print(prn,ch,f)
    l = 0.0
    if (f == 101):
        l = lightSpeed/(L1 + ch*dL1)
    if (f == 102):
        l = lightSpeed/(L2 + ch*dL2)
    return l
def open_outputfile(station,year,doy,extension):
    """
    inputs: station name, year, doy, and station name
    opens output file in REFL_CODE/year/results/station directory
    return fileID
    author kristine m. Larson
    if the results directory does not exist, it tries to make it. i think
    june 2019, added snrending to output name
    """
    if os.path.isdir('logs'):
        print('log directory exists')
    else:
        print('making log directory ')
        subprocess.call(['mkdir', 'logs'])
    fout = 0
#   primary reflector height output goes to this directory
    xdir = os.environ['REFL_CODE']
    cdoy = '{:03d}'.format(doy)
#   extra file with rejected arcs
    w = 'logs/reject.' + str(year) + '_' + cdoy  + station + '.txt'
    print('open output file for rejected arcs',w)
    frej=open(w,'w+')
#    frej=open('reject.txt','w+')
#   put a header in the file
    frej.write("%year, doy, maxF,sat,UTCtime, Azim, Amp,  eminO, emaxO,  Nv,freq,rise,Edot, PkNoise,  DelT,   MJD \n")
    filedir = xdir + '/' + str(year)  + '/results/' + station 
#    filepath1 =  filedir + '/' + cdoy  + '.txt'
#   changed to a function
    filepath1,fexit = LSPresult_name(station,year,doy,extension)
    print('output will go to:', filepath1)
    try:
        fout=open(filepath1,'w+')
#       put a header in the output file
        fout.write("% gnssIR_python, https://github.com/kristinemlarson \n")
        fout.write("% Phase Center corrections have NOT been applied \n")
        fout.write("% year, doy, maxF,sat,UTCtime, Azim, Amp,  eminO, emaxO,  Nv,freq,rise,EdotF, PkNoise  DelT     MJD   refr-appl\n")
        fout.write("% (1)  (2)   (3) (4)  (5)     (6)   (7)    (8)    (9)   (10) (11) (12) (13)  (14)     (15)     (16)   (17)\n")
        fout.write("%            m         hrs    deg   v/v    deg    deg                  hrs            min             1 is yes  \n")
    except:
        print('problem on first attempt - so try making results directory')
        f1 = xdir + '/' + str(year) + '/results/'
        subprocess.call(['mkdir',f1])
        # os.system(cm)
        f2 = xdir + '/' + str(year) + '/results/' + station
        subprocess.call(['mkdir',f2])
        # os.system(cm)
        try:
            fout=open(filepath1,'w+')
            print('successful open')
        except:
            print('problems opening the file')
            sys.exit()
    return fout, frej

def removeDC(dat,satNu, sat,ele, pele, azi,az1,az2,edot,seconds):
    """
#   remove direct signal using given elevation angle (pele) and azimuth 
    (az1,az2) constraints, return x,y as primary used data and windowed
    azimuth, time, and edot
#   removed zero points, which 10^0 have value 1.  used 5 to be sure?
    """
    p1 = pele[0]; p2 = pele[1]
#   look for data within these azimuth and elevation angle constraints
    x = ele[(sat == satNu) & (ele > p1) & (ele < p2) & (azi > az1) & (azi < az2) & (dat > 5)]
    y = dat[(sat == satNu) & (ele > p1) & (ele < p2) & (azi > az1) & (azi < az2) & (dat > 5)]
    edot = edot[(sat == satNu) & (ele > p1) & (ele < p2) & (azi > az1) & (azi < az2) & (dat > 5)]
    seconds = seconds[(sat == satNu) & (ele > p1) & (ele < p2) & (azi > az1) & (azi < az2) & (dat > 5)]
    azi = azi[(sat == satNu) & (ele > p1) & (ele < p2) & (azi > az1) & (azi < az2) & (dat > 5)]

    return x,y,sat,azi,seconds,edot

def quick_plot(plt_screen, gj,station,pltname,f):
    """
    inputs plt_screen variable (1 means go ahead) and integer variable gj
    which if > 0 there is something to plot
    also station name for the title
    pltname is png filename, if requested
    author: kristine m. larson
    """
    if (plt_screen == 1  and gj > 0):
        plt.subplot(212)
        plt.xlabel('Reflector height(m)')
        plt.ylabel('SNR Spectral Amplitude')
        plt.subplot(211)
        plt.title('Station:' + station + '/freq:' + str(f))
        plt.ylabel('SNR (volts/volts)')
        plt.xlabel('elevation angle(degrees)')
        if pltname != 'None':
            plt.savefig(pltname)
        else:
            print('plot file not saved ')
        plt.show()

def print_file_stats(ele,sat,s1,s2,s5,s6,s7,s8,e1,e2):
    """
    inputs 
    """
    gps = ele[(sat > 0) & (sat < 33) & (ele < e2) ]
    glonass = ele[(sat > 100) & (sat < 125) & (ele < e2) ]
    beidou = ele[(sat > 300) & (sat < 340) & (ele < e2) ]
    galileo = ele[(sat > 200) & (sat < 240) & (ele < e2) ]
    print('GPS     obs ', len(gps) )
    print('Glonass obs ', len(glonass))
    print('Galileo obs ', len(galileo))
    print('Beidou  obs ', len(beidou))

    return



def diffraction_correction(el_deg, temp=20.0, press=1013.25):
    """ Computes and return the elevation correction for refraction in the atmosphere.

    Computes and return the elevation correction for refraction in the atmosphere such that the elevation of the
    satellite plus the correction is the observed angle of incidence.

    Based on an empirical model by G.G. Bennet.
    This code was provided by Chalmers Group, Joakim Strandberg and Thomas Hobiger

    Parameters
    ----------
    el_deg : array_like
        A vector of true satellite elevations in degrees for which the correction is calculated.

    temp : float, optional
        Air temperature at ground level in degrees celsius, default 20 C.

    press : float, optional
        Air pressure at ground level in hPa, default 1013.25 hPa.

    Returns
    -------
    corr_el_deg : 1d-array
        The elevation correction in degrees.

    References:
    ----------
        Bennett, G. G. 'The calculation of astronomical refraction in marine navigation.'
        Journal of Navigation 35.02 (1982): 255-259.
    """
    el_deg = np.array(el_deg)

    corr_el_arc_min = 510/(9/5*temp + 492) * press/1010.16 * 1/np.tan(np.deg2rad(el_deg + 7.31/(el_deg + 4.4)))

    corr_el_deg = corr_el_arc_min/60

    return corr_el_deg
def mjd(y,m,d,hour,minute,second):
    """
    inputs: year, month, day, hour, minute,second
    output: modified julian day
    using information from http://infohost.nmt.edu/~shipman/soft/sidereal/ims/web/MJD-fromDatetime.html
    coded by kristine m. larson
    """
    if  (m <= 2):
        y, m = y-1, m+12
    if ((y, m, d) >= (1582, 10, 15)):
        A = int(y / 100)
        B = 2 - A + int(A / 4)
    else:
        B = 0
    C = int(365.25 * y)
    D = int(30.6001 *(m + 1))
    mjd = B + C + D + d - 679006
#   calculate seconds
    s = hour*3600 + minute*60 + second
    fracDay = s/86400
    return mjd, fracDay


def doy2ymd(year, doy):
    """
    inputs: year and day of year (doy)
    returns: some kind of datetime construct which can be used to get MM and DD
    """

    d = datetime.datetime(year, 1, 1) + datetime.timedelta(days=(doy-1))
    print('ymd',d)
    return d 

def getMJD(year,month,day,fract_hour):
    """
    inputs are year, month, day and fractional hour
    return is modified julian day (real8)
    """
#   convert fract_hour to HH MM SS
#   ignore fractional seconds for now
    hours = math.floor(fract_hour) 
    leftover = fract_hour - hours
    minutes = math.floor(leftover*60)
    seconds = math.floor(leftover*3600 - minutes*60)
#    print(fract_hour, hours, minutes, seconds)
    MJD, fracS = mjd(year,month,day,hours,minutes,seconds)
    MJD = MJD + fracS
    return MJD
def update_plot(plt_screen,x,y,px,pz):
    """
    input plt_screen integer value from gnssIR_lomb.
    (value of one means update the SNR and LSP plot)
    and values of the SNR data (x,y) and LSP (px,pz)
    """
    if (plt_screen == 1):
        plt.subplot(211)  
        plt.plot(x,y)
        #plt.title(station)
        plt.subplot(212)  
        plt.plot(px,pz)
def open_plot(plt_screen):
    """
    simple code to open a figure, called by gnssIR_lomb
    """
    if (plt_screen == 1):
        plt.figure()

def quick_rinex_snr(year, doy, station, option, orbtype,receiverrate,dec_rate):
    """
    inputs: year and day of year (integers) and station name
    option is for the snr creation
    orbtype can be nav or sp3.  if the former, then gpsSNR is used.
    if the later, then gnssSNR
    this assumes you follow my definitions for where things go,
    i.e. REFL_CODE and ORBITS
    it currently checks Unavco, SOPAC, and SONEL. I should add CDDIS
    author: kristine m. larson
    19may20, added decimation
    19sep12, I got tired of code crashing for files > 20 observables.  I am thus using teqc
    20apr15, xz compression added
    """
    # define directory for the conversion executables
    exedir = os.environ['EXE']
    # FIRST, check to see if the SNR file already exists
#    snrname_full,snrname_compressed = define_filename(station,year,doy,option)
    snrname_full, snrname_compressed, snre = define_and_xz_snr(station,year,doy,option)
    if (snre == True):
        print('snrfile already exists:', snrname_full)
    else:
        print('snr does not exist so pick up orbits and rinex')
        d = doy2ymd(year,doy); 
        month = d.month; day = d.day
        # new function to do the whole orbit thing
        foundit, f, orbdir, snrexe = get_orbits_setexe(year,month,day,orbtype) 
        # if you have the orbit file, you can get the rinex file
        if foundit:
            # now you can look for a rinex file
            rinexfile,rinexfiled = rinex_name(station, year, month, day)
            print('rinexfile should be ', rinexfile)
            if (os.path.isfile(rinexfile) == True):
                print('rinex file exists')
            else:
                print('go get the rinex file')
                if receiverrate == 'low':
                    print('seeking low rate at unavco')
                    rinex_unavco(station, year, month, day) 
                else:
                    print('seeking high rate at unavco')
                    rinex_unavco_highrate(station, year, month, day) 
                if os.path.isfile(rinexfile):
                    print('you have the rinex file')
                else:
                    # only keep looking if you are seeking lowrate data
                    if receiverrate == 'low':
                        try:
                            print('try to get it from SOPAC')
                            rinex_sopac(station, year, month, day)
                        except:
                            print('no success at SOPAC.')
                    if not os.path.isfile(rinexfile) and receiverrate == 'low':
                        try:
                            print('try SONEL')
                            rinex_sonel(station, year, month, day)
                        except:
                            print('got nothing - I give up')

    # check to see if you found the rinex file
    # should check that the orbit really exists too
            oexist = os.path.isfile(orbdir + '/' + f) == True
            rexist = os.path.isfile(rinexfile) == True
            # if rinex exists
            if rexist:
                exc = teqc_version()
                # get rid of all the observables i do not need
                if os.path.isfile(exc):
                    print('teqc executable exists, so let us use it')
                    foutname = 'tmp.' + rinexfile
                    fout = open(foutname,'w')
                # not sure this will work
                    print('try to fix glonass bug using teqc input')
                    subprocess.call([exc, '-O.obs','S1+S2+S5+S6+S7+S8', '-n_GLONASS', '27', rinexfile],stdout=fout)
                    fout.close()
                # store it in the original rinex filename
                    subprocess.call(['rm','-f',rinexfile])
                    subprocess.call(['mv','-f',foutname, rinexfile])
            if (rexist and dec_rate > 0):
                try:
                    print('decimate using teqc ', dec_rate, ' seconds')
                    print('testing subprocess while decimating')
                    exc = exedir + '/teqc' 
                    rinexout = rinexfile + '.tmp'; cdec = str(dec_rate)
                    fout = open(rinexout,'w')
                    subprocess.call([exc, '-O.dec', cdec, rinexfile],stdout=fout)
                    # not sure this is needed
                    fout.close()
                    # try using subprocess instead of os.system
                    status = subprocess.call(['mv','-f', rinexout, rinexfile])
                except:
                    print('decimation failed')
            if (oexist and rexist):
            #convert to SNR file
                snrname = snr_name(station, year,month,day,option)
                orbfile = orbdir + '/' + f
                try:
                    subprocess.call([snrexe, rinexfile, snrname, orbfile, str(option)])
                    status = subprocess.call(['rm','-f', rinexfile ])
                except:
                    print('no success making SNR file')
                if os.path.isfile(snrname): 
#                make sure it exists and is non-zero size before moving it
                    if (os.stat(snrname).st_size == 0):
                        print('you created a zero file size which could mean a lot of things')
                        print('bad exe, bad snr option, do not really have the orbit file')
                        status = subprocess.call(['rm','-f', snrname ])
                    else:
                        print('a SNR file was created and it is non-zero in length')
                        print(snrname_full)
                        store_snrfile(snrname,year,station) 
            else:
                print('rinex file or orbit file does not exist, so there is nothing to convert')

def store_orbitfile(filename,year,orbtype):
    """
    simple code to move an orbit file to the right place 
    inputs are the filename, the year, and the kind of orbit
    (sp3 or nav)
    """
    xdir = os.environ['ORBITS'] + '/' + str(year)
    # check that directories exist
    if not os.path.isdir(xdir): #if year folder doesn't exist, make it
        os.makedirs(xdir)
    xdir = os.environ['ORBITS'] + '/' + str(year) + '/' + orbtype
    if not os.path.isdir(xdir): #if year folder doesn't exist, make it
        os.makedirs(xdir)
    if (os.path.isfile(filename) == True):
        print('moving ', filename, ' to ', xdir)
        status = subprocess.call(['mv','-f', filename, xdir])
    else:
        print('The orbit file did not exist, so it was not stored')
    return xdir


def store_snrfile(filename,year,station):
    """
    simple code to move an snr file to the right place 
    inputs are the filename, the year, and the station name
    """
    xdir = os.environ['REFL_CODE'] + '/' + str(year)
    # check that directories exist
    if not os.path.isdir(xdir): #if year folder doesn't exist, make it
        os.makedirs(xdir)
    xdir = xdir + '/snr'
    if not os.path.isdir(xdir): #if year folder doesn't exist, make it
        os.makedirs(xdir)
    xdir = xdir + '/' + station 
    if not os.path.isdir(xdir): #if year folder doesn't exist, make it
        os.makedirs(xdir)
    if (os.path.isfile(filename) == True):
        status = subprocess.call(['mv','-f', filename, xdir])
    else:
        print('the SNR file does not exist, so nothing was moved')

def rinex_name(station, year, month, day):
    """
    author: kristine larson
    given station (4 char), year, month, day, return rinexfile name
    and the hatanaka equivalent
    """
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)

    fnameo = station + cdoy + '0.' + cyy + 'o'
    fnamed = station + cdoy + '0.' + cyy + 'd'
    return fnameo, fnamed

def snr_name(station, year, month, day,option):
    """
    author: kristine larson
    given station (4 char), year, month, day, and snr option,
    return snr filename (and directory) using my system
    """
    doy,cdoy,cyyy,cyy = ymd2doy(year,month,day)

    fname = station + cdoy + '0.' + cyy + '.snr' + str(option)
    return fname

def nav_name(year, month, day):
    """
    kristine m. larson
    inputs are year month and day
    returns nav file name and directory
    """
    if (day == 0):
#        print('using doy instead of month and day')
        cdoy = '{:03d}'.format(month)
        cyyyy = str(year)
        cyy = '{:02d}'.format(year-2000)
    else:
        doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    navfilename = 'auto'  + cdoy + '0.' + cyy  +  'n'
    navfiledir = os.environ['ORBITS'] + '/' + cyyyy + '/nav'
    return navfilename,navfiledir

def sp3_name(year,month,day,pCtr):
    """
    kristine m. larson
    inputs are year month and day and processing center
    returns sp3 file name and directory
    """
    name,clkn=igsname(year,month,day)
    gps_week = name[3:7]
    sp3name = pCtr + name[3:8] + '.sp3'
    sp3dir = os.environ['ORBITS'] + '/' + str(year) + '/sp3'
    return sp3name, sp3dir

def rinex_unavco_obs(station, year, month, day):
    """
    author: kristine larson
    picks up a RINEX file from unavco.  
    normal observation file - not Hatanaka
    new version i was testing out
    """
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    rinexfile,rinexfiled = rinex_name(station, year, month, day) 
    unavco= 'ftp://data-out.unavco.org' 
    filename = rinexfile + '.Z'
    url = unavco+ '/pub/rinex/obs/' + cyyyy + '/' + cdoy + '/' + filename
    print(url)
    try:
        wget.download(url,filename)
        subprocess.call(['uncompress',filename])
    except:
        print('some kind of problem with download',rinexfile)


def rinex_unavco_highrate(station, year, month, day):
    """
    author: kristine larson
    picks up a RINEX file from unavco.  it tries to pick up an o file,
    but if it does not work, it tries the "d" version, which must be
    decompressed.  the location of this executable is defined in the crnxpath
    variable. This is from the main unavco directory - not the highrate directory.

    WARNING: only rinex version 2 in this world
    """
    exedir = os.environ['EXE']
    crnxpath = exedir + '/CRX2RNX'
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    rinexfile,rinexfiled = rinex_name(station, year, month, day)
    unavco= 'ftp://data-out.unavco.org'
    filename1 = rinexfile + '.Z'
    filename2 = rinexfiled + '.Z'
    # URL path for the o file and the d file
    url1 = unavco+ '/pub/highrate/1-Hz/rinex/' + cyyyy + '/' + cdoy + '/' + station + '/' + filename1
    url2 = unavco+ '/pub/highrate/1-Hz/rinex/' + cyyyy + '/' + cdoy + '/' + station + '/' + filename2
    print(url1)
    print(url2)
    try:
        print('try to get o file')
        wget.download(url1,filename1)
        subprocess.call(['uncompress',filename1])
    except:
        print('did not find o file - will look for a d file')
        try:
            wget.download(url2,filename2)
            subprocess.call(['uncompress',filename2])
            #convert from hatanaka
            subprocess.call([crnxpath, rinexfiled])
            #remove the hatanaka file
            subprocess.call(['rm','-f',rinexfiled])
            print('found d file and converted to o file')
        except:
            print('failed to find either RINEX file at unavco')

def big_Disk_in_DC(station, year, month, day):
    """
    author: kristine larson
    picks up a RINEX file from CORS.  
    """
    exedir = os.environ['EXE']
    crnxpath = exedir + '/CRX2RNX'
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    rinexfile,rinexfiled = rinex_name(station, year, month, day)
    comp_rinexfiled = rinexfiled + '.Z'
    mainadd = 'ftp://www.ngs.noaa.gov/cors/rinex/'
    url = mainadd + str(year) + '/' + cdoy+ '/' + station + '/' + comp_rinexfiled 
    try:
        wget.download(url, out=comp_rinexfiled)
        status = subprocess.call(['uncompress', comp_rinexfiled])
        status = subprocess.call([crnxpath, rinexfiled])
    # rm the hatanaka file
        status = subprocess.call(['rm','-f',rinexfiled])
    except:
        print('some problem in download - maybe the site does not exist on this archive')

def ydoy2ymd(year, doy):
    """
    inputs: year and day of year (doy)
    returns: useful stuff
    """

    d = datetime.datetime(year, 1, 1) + datetime.timedelta(days=(doy-1))
    month = int(d.month)
    day = int(d.day)
    return year, month, day

def rewrite_UNR_highrate(fname,station,year,doy):
    """
    takes a filename from was already retrieved? from 
    UNReno, reads it, rewrites as all numbers for other uses.
    no header, but year, month, day, day of year, seconds vertical, east, north
    the latter three are in meters
    stores in $REFL_CODE/yyyy/pos/station
    """
# make sure the various output directories  are there
    xdir = os.environ['REFL_CODE'] 
    dir1 = xdir + '/' + str(year)
    if not os.path.isdir(dir1):
        status = subprocess.call(['mkdir', dir1])

    dir1 = xdir + '/' + str(year) + '/' + 'pos'
    if not os.path.isdir(dir1):
        status = subprocess.call(['mkdir', dir1])

    dir1 = xdir + '/' + str(year) + '/' + 'pos' + '/' + station
#   make filename for the output
    yy,mm,dd, cyyyy, cdoy, YMD = ydoy2useful(year,doy)
    outputfile = dir1 + '/' + cdoy + '_hr.txt'
    print('file will go to: ' , outputfile)
    if not os.path.isdir(dir1):
        print('use subprocess to make directory')
        status = subprocess.call(['mkdir', dir1])
    try:
        x=np.genfromtxt(fname, skip_header=1, usecols = (3, 4, 5, 6, 7, 8, 9, 10))
        N = len(x)
        print('open outputfile',outputfile)
        f=open(outputfile,'w+')
        for i in range(0,N):
            f.write(" {0:4.0f} {1:2.0f} {2:2.0f} {3:3.0f} {4:7.0f} {5:9.4f} {6:9.4f} {7:9.4f} \n".format(x[i,0], x[i,1],x[i,2],x[i,3], x[i,4],x[i,5],x[i,6],x[i,7]))
        print('delete the original Blewitt file')
        subprocess.call(['rm','-f', fname])
        f.close()
    except:
        print('problem with accessing the file')

def month_converter(month):
    """
    brendan gave this to me - give it a 3 char month, returns integer
    """
    months = ['JAN', 'FEB', 'MAR', 'APR', 'MAY', 'JUN', 'JUL', 'AUG', 'SEP', 'OCT', 'NOV', 'DEC']
    return months.index(month) + 1

def char_month_converter(month):
    """
    integer month to 3 character month
    """
    months = ['JAN', 'FEB', 'MAR', 'APR', 'MAY', 'JUN', 'JUL', 'AUG', 'SEP', 'OCT', 'NOV', 'DEC']
    return months[(month-1)]

def UNR_highrate(station,year,doy):
    """
    input station name and it picks up the 5 minute time series from UNR website
    returns name of the file and an iostat

    """
    yy,mm,dd, cyyyy, cdoy, YMD = ydoy2useful(year,doy)
    stationUP = station.upper()
    dirdir = 'ftp://gneiss.nbmg.unr.edu/rapids_5min/kenv/' + cyyyy + '/' + cdoy + '/'
    filename = YMD + station.upper() + '_fix.kenv'
    url = dirdir + filename
    if (os.path.isfile(filename) == True):
        print('file already exists')
        goodDownload = True
    else:
        try:
            wget.download(url,filename)
            goodDownload = True
        except:
            print(url)
            print('could not get the highrate file from UNR')
            goodDownload = False
    return filename, goodDownload

def mjd_to_date(jd):
    """
# KL converted from this reference
# https://gist.github.com/jiffyclub/1294443
    Converts Modified Julian Day to y,m,d
    
    Algorithm from 'Practical Astronomy with your Calculator or Spreadsheet', 
        4th ed., Duffet-Smith and Zwart, 2011.
    
    Parameters
    ----------
    jd : float
        Julian Day
        
    Returns
    -------
    year : int
        Year as integer. Years preceding 1 A.D. should be 0 or negative.
        The year before 1 A.D. is 0, 10 B.C. is year -9.
        
    month : int
        Month as integer, Jan = 1, Feb. = 2, etc.
    
    day : float
        Day, may contain fractional part.
        
    
    """
   # first step is to change MJD to jd, since original code expected it 
    jd = jd + 2400000.5
    # start of the original code
    jd = jd + 0.5
    
    F, I = math.modf(jd)
    I = int(I)
    
    A = math.trunc((I - 1867216.25)/36524.25)
    
    if I > 2299160:
        B = I + 1 + A - math.trunc(A / 4.)
    else:
        B = I
        
    C = B + 1524
    
    D = math.trunc((C - 122.1) / 365.25)
    
    E = math.trunc(365.25 * D)
    
    G = math.trunc((C - E) / 30.6001)
    
    day = C - E + F - math.trunc(30.6001 * G)
    
    if G < 13.5:
        month = G - 1
    else:
        month = G - 13
        
    if month > 2.5:
        year = D - 4716
    else:
        year = D - 4715
        
    day = int(day)
    return year, month, day

def getseries(site):
    """
    originally from brendan crowell.
    picks up two UNR time series - stores in subdirectory called tseries 
    input is station name (four character, lower case)
    """
    # i changed this to download ENV instead of XYZ (or both?)
    #if tseries folder doesn't exist, make it
    if not os.path.exists('tseries'): 
        os.makedirs('tseries')
    # change station to uppercase
    siteid = site.upper()
    # why still ITRF 2008?
    fname = 'tseries/' + siteid + '.IGS08.tenv3'
    # NA12 env time series
    fname2 = 'tseries/' + siteid + '.NA12.tenv3'
    # NA12 env rapid time series
    fname3 = 'tseries/' + siteid + '.NA12.rapid.tenv3'

    if (os.path.isfile(fname) == True):
        print ('Timeseries file ' + fname + ' already exists')
    else:
        url = 'http://geodesy.unr.edu/gps_timeseries/tenv3/IGS08/' + siteid + '.IGS08.tenv3'
        wget.download(url, out='tseries/')
#
    if (os.path.isfile(fname2) == True):
        print ('Timeseries file ' + fname2 + ' already exists')
    else:
        url = 'http://geodesy.unr.edu/gps_timeseries/tenv3/NA12/' + siteid + '.NA12.tenv3'
        wget.download(url, out='tseries/')

    if (os.path.isfile(fname3) == True):
        print ('Timeseries file ' + fname3 + ' already exists')
    else:
        url = 'http://geodesy.unr.edu/gps_timeseries/rapids/tenv3/NA12/' + siteid + '.NA12.tenv3'
        wget.download(url, out=fname3)

def rewrite_tseries(station):
    """
    given a station name, look at a daily blewitt position (ENV) 
    file and write a new file that is less insane to understand
    """
    siteid = station.upper()
    # NA12 env time series
    fname = 'tseries/' + siteid + '.NA12.tenv3'
    fname_rapid = 'tseries/' + siteid + '.NA12.rapid.tenv3'
    outputfile = 'tseries/' + station+ '_na12.env'
    print(fname,outputfile)
    try:
        x=np.genfromtxt(fname, skip_header=1, usecols = (3, 7, 8, 9, 10, 11, 12,13))
        N = len(x)
        print(N,'open outputfile',outputfile)
        f=open(outputfile,'w+')
        for i in range(0,N):
            mjd = x[i,0]
            yy,mm,dd = mjd_to_date(mjd) 
            doy, cdoy, cyyyy, cyy = ymd2doy(yy,mm,dd)
            east = x[i,1] + x[i,2]
            north= x[i,3] + x[i,4]
            # adding in the antenna
            vert = x[i,5] + x[i,6] +  x[i,7]
            f.write(" {0:4.0f} {1:2.0f} {2:2.0f} {3:3.0f} {4:13.4f} {5:13.4f} {6:13.4f} \n".format(yy,mm,dd,doy,east,north,vert))
        f.close()
    except:
        print('some problem writing the output')

def rewrite_tseries_igs(station):
    """
    given a station name, look at a daily blewitt position (ENV)
    file and write a new file that is less insane to understand
    """
    siteid = station.upper()
    # NA12 env time series
    fname = 'tseries/' + siteid + '.IGS08.tenv3'
    outputfile = 'tseries/' + station + '_igs08.env'
    print(fname,outputfile)
    try:
        x=np.genfromtxt(fname, skip_header=1, usecols = (3, 7, 8, 9, 10, 11, 12,13))
        N = len(x)
        print(N)
        print(N,'open outputfile',outputfile)
        f=open(outputfile,'w+')
        for i in range(0,N):
            mjd = x[i,0]
            yy,mm,dd = mjd_to_date(mjd)
            doy, cdoy, cyyyy, cyy = ymd2doy(yy,mm,dd)
            east = x[i,1] + x[i,2]
            north= x[i,3] + x[i,4]
            # adding in the antenna
            vert = x[i,5] + x[i,6] +  x[i,7]
            f.write(" {0:4.0f} {1:2.0f} {2:2.0f} {3:3.0f} {4:13.4f} {5:13.4f} {6:13.4f} \n".format(yy,mm,dd,doy,east,north,vert))
        f.close()
    except:
        print('some problem writing the output')

def codclock(year,month,day):
    """
    author: kristine lasron
    pick up 5 second clocks from the bernese group...

    not using this - so have not added CDDIS secure ftp
    """
    n,nn=igsname(year,month,day)
    gps_week = n[3:7]
    file1 = nn + '.Z'
    print(nn, gps_week, file1)
    h = 'ftp://cddis.gsfc.nasa.gov/gnss/products/' 
    url = h + str(gps_week) + '/' + file1
    print(url)
    try:
        wget.download(url, out=file1)
        subprocess.call(['gunzip','-f',file1])
    except:
        print('some kind of problem downloading clock files')

def llh2xyz(lat,lon,height):
    """
    inputs lat,lon (in degrees) and ellipsoidal height (in meters)
    returns Cartesian values in meters.

    Ref: Decker, B. L., World Geodetic System 1984,
    Defense Mapping Agency Aerospace Center.
    modified from matlab version kindly provided by CCAR
    """
    A_EARTH = 6378137;
    f= 1/298.257223563;
    NAV_E2 = (2-f)*f; # also e^2
    deg2rad = math.pi/180.0

    slat = math.sin(lat*deg2rad);
    clat = math.cos(lat*deg2rad);
    r_n = A_EARTH/(math.sqrt(1 - NAV_E2*slat*slat))
    x= (r_n + height)*clat*math.cos(lon*deg2rad)
    y= (r_n + height)*clat*math.sin(lon*deg2rad)
    z= (r_n*(1 - NAV_E2) + height)*slat
    return x, y, z

def LSPresult_name(station,year,doy,extension):
    """
    given station name, year, doy, and extension
    returns the location of the LSP result
    also returns boolean if it already exists (so that
    information can be used)
    extension is now used
    """
    # for testing
    xdir = os.environ['REFL_CODE']
    cyear = str(year)
    cdoy = '{:03d}'.format(doy)
    # this is default location where the results will go
    filedir = xdir + '/' + cyear  + '/results/' + station
    # this is now also done in the result_directories function,
    # but I guess no harm is done
    if not os.path.isdir(filedir):
        print('making new results bdirectory ')
        subprocess.call(['mkdir', filedir])
    filedirx = filedir + '/' + extension
    # this is what you do if there is an extension
    if not os.path.isdir(filedirx):
        print('making new results subdirectory ')
        subprocess.call(['mkdir', filedirx])

    filepath1 =  filedirx + '/' + cdoy  + '.txt'

    print('output for this date will go to:', filepath1)
    if os.path.isfile(filepath1):
        print('A result file already exists')
        fileexists = True
    else:
        print('A result file does not exist')
        fileexists = False
    return filepath1, fileexists

def result_directories(station,year,extension):
    """
    inputs station, year, and extension
    makes output directories for results
    jun30, 2019
    kristine larson
    """
    xdir = os.environ['REFL_CODE']
    cyear = str(year)

    f1 = xdir + '/' + cyear
    if not os.path.isdir(f1):
        subprocess.call(['mkdir',f1])

    f1 = f1 + '/results'
    if not os.path.isdir(f1):
        subprocess.call(['mkdir',f1])


    f1 = f1 + '/' + station
    if not os.path.isdir(f1):
        subprocess.call(['mkdir',f1])

    if (extension != ''):
        f1 = f1 + '/' + extension
        if not os.path.isdir(f1):
            subprocess.call(['mkdir',f1])
    else:
        print('no extension')

    f1 = xdir + '/' + cyear + '/phase'
    if not os.path.isdir(f1):
        subprocess.call(['mkdir',f1])

    f1 = f1 + '/' + station
    if not os.path.isdir(f1):
        subprocess.call(['mkdir',f1])

def write_QC_fails(delT,delTmax,eminObs,emaxObs,e1,e2,ediff,maxAmp, Noise,PkNoise,reqamp):
    """
    prints out various QC fails to the screen

    """
    if delT > delTmax:
        print('       delT {0:.1f} minutes '.format(delT ))
    if eminObs  > (e1 + ediff):
        print('       emin {0:.1f} '.format(eminObs ))
    if emaxObs  < (e2 - ediff):
        print('       emax {0:.1f} '.format(emaxObs ))
    if maxAmp < reqamp:
        print('       Ampl {0:.1f} '.format(maxAmp  ))
    if maxAmp/Noise < PkNoise:
        print('       PkN  {0:.1f} '.format(maxAmp/Noise ))
        
def define_quick_filename(station,year,doy,snr):
    """
    given station name, year, doy, snr type
    returns snr filename but without default directory structure
    author: Kristine Larson
    19mar25: return compressed filename too
    """
    cdoy = '{:03d}'.format(doy)
    cyy = '{:02d}'.format(year-2000)
    f= station + str(cdoy) + '0.' + cyy + '.snr' + str(snr)
    return f

def update_quick_plot(station, f):
    """
    input plt_screen integer value from gnssIR_lomb.
    (value of one means update the SNR and LSP plot)
    and values of the SNR data (x,y) and LSP (px,pz)
    """
    plt.subplot(212)
    plt.xlabel('reflector height (m)'); plt.title('SNR periodogram')
    plt.subplot(211)
    plt.xlabel('elev Angles (deg)')
    plt.title(station + ' SNR Data and Frequency L' + str(f))

    return True

def navfile_retrieve(navfile,cyyyy,cyy,cdoy):
    """
    inputs are navfile name and character strings for year, two character year, and 
    day of year.
    output: a boolean is returned if the file exists
    the code tries to find a file at unavco first, then sopac, then cddis.  it checks for illegal
    files at sopac. it stores the file as autoDDD0.YYn where DDD is day of year and YY
    is two charadter year irregardless of what the original name is.
    author: kristine larson, september 2019
    CHANGED April 2, 2020  Should not use the Sc02 file from UNAVCO.  This was a poor
    choice. Now will use CDDIS first, then SOPAC, 
    I have removed unavco as they do not seem to provide global nav files
    If they do, I do not know where they are kept.

    20jun14 added secure ftp for CDDIS
    """
    FileExists = True
    unavco= 'ftp://data-out.unavco.org'
    sopac = 'ftp://garner.ucsd.edu'
    cddis = 'ftp://ftp.cddis.eosdis.nasa.gov'
    navname = navfile
    navfile_sopac1 =  navfile   + '.Z' # regular nav file
    navfile_compressed = navfile + '.Z'
    url_sopac1 = sopac + '/pub/rinex/' + cyyyy + '/' + cdoy + '/' + navfile_sopac1

    navfile_sopac2 =  navfile
    url_sopac2 = sopac + '/pub/rinex/' + cyyyy + '/' + cdoy + '/' + navfile_sopac2

    navfile_unavco = 'sc02' + navfile[4:] + '.Z'
    url_unavco = unavco + '/pub/rinex/nav/' + cyyyy + '/' + cdoy + '/' + navfile_unavco

    navfile_cddis = 'brdc' + navfile[4:] + '.Z'
    url_cddis = cddis + '/gps/data/daily/' + cyyyy + '/' + cdoy + '/' +cyy + 'n/' + navfile_cddis
    print(url_cddis)
    try:
        # use a function instead
        get_cddis_navfile(navfile,cyyyy,cyy,cdoy) 
        #print('try cddis first')
        #wget.download(url_cddis,navfile_compressed)
        #subprocess.call(['uncompress',navfile_compressed])
        print('success at cddis')
    except:
        print('no success at cddis')
        try:
            print('try SOPAC next')
            wget.download(url_sopac1,navfile_sopac1)
            subprocess.call(['uncompress',navfile_sopac1])
        except:
            print('no success at SOPAC')
            if not os.path.exists(navname):
                print('get rid of corrupted SOPAC file, if any')
                subprocess.call(['rm','-f',navfile_sopac1])
                subprocess.call(['rm','-f',navname])

    if not os.path.isfile(navfile):
        FileExists = False

    return FileExists
def make_nav_dirs(yyyy):
    """
    input year and it makes sure output directories are created for orbits
    """
    n = os.environ['ORBITS']
    # if parent nav dir does not exist, exit
    if not os.path.isdir(n):
        print('You have not defined ORBITS environment variable properly. Exiting')
        print(n)
        sys.exit()
    cyyyy = '{:04d}'.format(yyyy)
    navfiledir = os.environ['ORBITS'] + '/' + cyyyy 
    if not os.path.exists(navfiledir):
        subprocess.call(['mkdir',navfiledir])
        print('making year directory')
    navfiledir1 = os.environ['ORBITS'] + '/' + cyyyy + '/nav' 
    if not os.path.exists(navfiledir1):
        subprocess.call(['mkdir',navfiledir1])
        print('making nav specific directory')
    navfiledir2 = os.environ['ORBITS'] + '/' + cyyyy + '/sp3'
    if not os.path.exists(navfiledir2):
        print('making sp3 specific directory')
        subprocess.call(['mkdir',navfiledir2])

    return True


def check_inputs(station,year,doy,snr_type):
    """
    inputs to Lomb Scargle and Rinex translation codes
    are checked for sensibility. Returns true or false to 
    code can exit. Error messages sent to the screen
    author: kristine m. larson
    2019sep22
    """
    exitSys = False
    if len(station) != 4:
        print('Station name must be four characters. Exiting')
        exitSys = True

    if len(str(year)) != 4:
        print('Year must be four characters. Exiting')
        exitSys = True

    if (doy < 1) or (doy > 366):
        print('Day of year must be bewteen 1 and 366. Exiting')
        exitSys = True

    s = snr_type
    if (s == 99) or (s == 50) or (s==66) or (s==88) or (s==77):
        print('You have picked a proper SNR file format ending.' )
    else:
        print('You have picked an improper SNR file format ending: ' + str(s))
        print('Allowed values are 99, 66, 88, 77, or 50. Exiting')
        exitSys = True

    return exitSys


def rewrite_tseries_wrapids(station):
    """
    given a station name, look at a daily blewitt position (ENV)
    file and write a new file that is less insane to understand
    """
    siteid = station.upper()
    # NA12 env time series
    fname = 'tseries/' + siteid + '.NA12.tenv3'
    fname_rapid = 'tseries/' + siteid + '.NA12.rapid.tenv3'
    outputfile = 'tseries/' + station+ '_na12.env'
    print(fname,outputfile)
    try:
        x=np.genfromtxt(fname, skip_header=1, usecols = (3, 7, 8, 9, 10, 11, 12,13))
        y=np.genfromtxt(fname_rapid, skip_header=1, usecols = (3, 7, 8, 9, 10, 11, 12,13))
        N = len(x)
        N2 = len(y)
        print(N,'open outputfile',outputfile)
        f=open(outputfile,'w+')
        for i in range(0,N):
            mjd = x[i,0]
            yy,mm,dd = mjd_to_date(mjd)
            doy, cdoy, cyyyy, cyy = ymd2doy(yy,mm,dd)
            east = x[i,1] + x[i,2]
            north= x[i,3] + x[i,4]
            # adding in the antenna
            vert = x[i,5] + x[i,6] +  x[i,7]
            f.write(" {0:4.0f} {1:2.0f} {2:2.0f} {3:3.0f} {4:13.4f} {5:13.4f} {6:13.4f} \n".format(yy,mm,dd,doy,east,north,vert))
# write out the rapid numbers
        for i in range(0,N2):
            mjd = y[i,0]
            yy,mm,dd = mjd_to_date(mjd)
            doy, cdoy, cyyyy, cyy = ymd2doy(yy,mm,dd)
            east = y[i,1] + y[i,2]
            north= y[i,3] + y[i,4]
            # adding in the antenna
            vert = y[i,5] + y[i,6] +  y[i,7]
            f.write(" {0:4.0f} {1:2.0f} {2:2.0f} {3:3.0f} {4:13.4f} {5:13.4f} {6:13.4f} \n".format(yy,mm,dd,doy,east,north,vert))
#  then close it
        f.close()
    except:
        print('some problem writing the output')

def back2thefuture(iyear, idoy):
    """
    user inputs iyear and idoy
    and the code checks that this is not a day in the future
    also don't allow the dates to agree exactly because we 
    do not have the current day's orbit file available (though
    it could be chagned to allow it.

    reject data before the year 2000
    """
    # find out today's date
    year = int(date.today().strftime("%Y"));
    month = int(date.today().strftime("%m"));
    day = int(date.today().strftime("%d"));

    today=datetime.datetime(year,month,day)
    doy = (today - datetime.datetime(today.year, 1, 1)).days + 1

    badDay = False
    if (iyear > year):
        badDay = True
    elif (iyear == year) & (idoy >= doy):
        badDay = True
    elif (iyear < 2000):
        badDay = True

    return badDay

def rinex_ga_lowrate(station,year,month,day):
    """
    pick up lowrate data from Geoscience Australia
    only Rinex 2.11 for now
    kristine larson
    20jul10: change from year doy to year month day, because that is what the others are
    """
    fexists = False
    exedir = os.environ['EXE']
    crnxpath = exedir + '/CRX2RNX'
    if day == 0:
        doy=month
        d = doy2ymd(year,doy);
        month = d.month; day = d.day
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    ftpg = 'ftp://ftp.ga.gov.au/geodesy-outgoing/gnss/data/daily/' + cyyyy + '/' + cyy + cdoy  + '/'
    dnameZ = station + cdoy + '0.' + cyy + 'd.Z' 
    dname = station + cdoy + '0.' + cyy + 'd'
    rname = station + cdoy + '0.' + cyy + 'o'
    url = ftpg + dname + '.Z'
    print(url)
    #print(dname,rname)
    try:
        wget.download(url,dnameZ)
        if os.path.isfile(dnameZ):
            status = subprocess.call(['uncompress', dnameZ])
            status = subprocess.call([crnxpath, dname])
            status = subprocess.call(['rm', '-f', dname])
    except:
        print('download from GA did not work')

    if os.path.isfile(rname):
        print('file found')
        fexists = True

    return fexists

def rinex_ga_highrate(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a higrate RINEX file from Geoscience Australia
    you can input day =0 and it will assume month is day of year
    not sure if it merges them ...
    """
    exedir = os.environ['EXE']
    crnxpath = exedir + '/CRX2RNX'
    teqcpath = exedir + '/teqc'
    alpha='abcdefghijklmnopqrstuvwxyz'
    # if doy is input
    if day == 0:
        doy=month
        d = doy2ymd(year,doy);
        month = d.month; day = d.day
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    gns = 'ftp://ftp.ga.gov.au/geodesy-outgoing/gnss/data/highrate/' + cyyyy + '/' + cyy + cdoy 
    for h in range(0,24):
        # subdirectory
        ch = '{:02d}'.format(h)
        for e in ['00', '15', '30', '45']:
            dname = station + cdoy + alpha[h] + e + '.' + cyy + 'd.Z'
            dname1 = station + cdoy + alpha[h] + e + '.' + cyy + 'd'
            dname2 = station + cdoy + alpha[h] + e + '.' + cyy + 'o'
            url = gns + '/' + ch + '/' + dname
            print(url)
            try:
                wget.download(url,dname)
                subprocess.call(['uncompress',dname])
                subprocess.call([crnxpath, dname1])
                subprocess.call([teqcpath])
            except:
                print('did not work')


def highrate_nz(station, year, month, day):
    """
    author: kristine larson
    inputs: station name, year, month, day
    picks up a RINEX file from GNS New zealand
    you can input day =0 and it will assume month is day of year
    """
    doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)
    gns = 'ftp://ftp.geonet.org.nz/gnss/event.highrate/1hz/raw/' + cyyyy +'/' + cdoy + '/'
    print(gns)
    stationU=station.upper()
    # will not work for all days but life is short
    cmm  = '{:02d}'.format(month)
    cdd  = '{:02d}'.format(day)
    exedir = os.environ['EXE']
    trimbleexe = exedir + '/runpkr00' 
    teqc = exedir + '/teqc' 
    for h in range(0,24):
        # subdirectory
        chh = '{:02d}'.format(h)
        file1= stationU + cyyyy + cmm + cdd  + chh + '00b.T02'
        file1out= stationU + cyyyy + cmm + cdd  + chh + '00b.tgd'
        file2= stationU + cyyyy + cmm + cdd  + chh + '00b.rnx'
        url = gns + file1
        try:
            wget.download(url,file1)
            subprocess.call([trimbleexe, '-g','-d',file1])
            print(file1out, file2)
            f = open(file2, 'w')
            subprocess.call([teqc, '-week', '2083', '-O.obs', 'S1+S2+C2+L2', file1out], stdout=f)
            f.close()
            print('successful download from GeoNet New Zealand')
        except:
            print('some kind of problem with download',file1)

def llh2xyz(lat,lon,height):
    """
    inputs lat,lon (in degrees) and height in meters
    returns x,y,z in meters
    # not sure where I got this
    """
    A_EARTH = 6378137;
    f= 1/298.257223563;
    NAV_E2 = (2-f)*f; # also e^2
    deg2rad = math.pi/180.0

    slat = math.sin(lat*deg2rad);
    clat = math.cos(lat*deg2rad);
    #    print(username,lat,slat)
    r_n = A_EARTH/(math.sqrt(1 - NAV_E2*slat*slat))
    x= (r_n + height)*clat*math.cos(lon*deg2rad)
    y= (r_n + height)*clat*math.sin(lon*deg2rad)
    z= (r_n*(1 - NAV_E2) + height)*slat
    return x,y,z

def get_orbits_setexe(year,month,day,orbtype):
    """
    helper script to make snr files
    takes year, month, day and orbit type
    picks up and stores
    also sets executable location (gpsonly vs gnss)
    returns whether file was found, its name, the directory, and name of snrexe
    20apr15:  check for xz compression
    kristine larson
    """
    #default values
    foundit = False
    f=''; orbdir=''
    # define directory for the conversion executables
    exedir = os.environ['EXE']
    snrexe = gpsSNR_version()
    # this means you are using multi-GNSS and GFZ
    #f,orbdir,foundit=getsp3file_mgex(year,month,day,'gbm')
    #snrexe = gnssSNR_version()
    #warn_and_exit(snrexe)
    if orbtype == 'sha':
        # SHANGHAI multi gnss
        f,orbdir,foundit=getsp3file_mgex(year,month,day,'sha')
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif (orbtype == 'grg'):
        # French multi gnss, but not Beidou
        f,orbdir,foundit=getsp3file_mgex(year,month,day,'grg')
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif (orbtype == 'sp3'):
        print('uses default IGS orbits, so only GPS ?')
        f,orbdir,foundit=getsp3file_flex(year,month,day,'igs')
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif (orbtype == 'gfz'):
        print('using gfz sp3 file, GPS and GLONASS') # though I advocate gbm
        f,orbdir,foundit=getsp3file_flex(year,month,day,'gfz')
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif (orbtype == 'igr'):
        print('using rapid orbits, so only GPS')
        f,orbdir,foundit=getsp3file_flex(year,month,day,'igr') # use default
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif (orbtype == 'igs'):
        print('using IGS final orbits, so only GPS')
        f,orbdir,foundit=getsp3file_flex(year,month,day,'igs') # use default
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif (orbtype == 'gbm'):
        # this uses GFZ multi-GNSS and is rapid, but not super rapid
        f,orbdir,foundit=getsp3file_mgex(year,month,day,'gbm')
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif (orbtype == 'wum'):
        # this uses WUHAN multi-GNSS which is ultra, but is not rapid ??
        # but only hour 00:00
        f,orbdir,foundit=getsp3file_mgex(year,month,day,'wum')
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif orbtype == 'jax':
        # this uses JAXA, has GPS and GLONASS and appears to be quick and reliable
        f,orbdir,foundit=getsp3file_mgex(year,month,day,'jax')
        snrexe = gnssSNR_version()
        warn_and_exit(snrexe)
    elif orbtype == 'nav':
        f,orbdir,foundit=getnavfile(year, month, day) # use default version, which is gps only
        warn_and_exit(snrexe)
    else:
        print('I do not recognize the orbit type you tried to use: ', orbtype)

    return foundit, f, orbdir, snrexe

def warn_and_exit(snrexe):
    """
    if snr executable does not exist, exit
    """
    if (os.path.isfile(snrexe) == False):
        print('The translation executable does not exist:' + snrexe + ' Exiting ')
        sys.exit()

def quick_rinex_snrC(year, doy, station, option, orbtype,receiverrate,dec_rate,archive):
    """
    inputs: year and day of year (integers) and station name
    option is for the snr creation ??? integer or character?
    orbtype can be nav or sp3.  if the former, then gpsSNR is used.
    if the later, then gnssSNR
    what are receiverrate and dec_rate defaults?
    this assumes you follow my definitions for where things go,
    i.e. REFL_CODE and ORBITS
    it currently checks Unavco, SOPAC, and SONEL. I should add CDDIS
    author: kristine m. larson
    19may20, added decimation
    19sep12, I got tired of code crashing for files > 20 observables.  I am thus using teqc
    20apr15, xz compression added but also try to streamline it.
    20jul10, added arvchive setting. default is 'all'

    """
    # define directory for the conversion executables
    print('receiver rate:',receiverrate)
    print('decimation: ', dec_rate)
    print('archive: ', archive)
    exedir = os.environ['EXE']
    snrname_full, snrname_compressed, snre = define_and_xz_snr(station,year,doy,option)
    if (snre == True):
        print('snrfile already exists:', snrname_full)
    else:
        print('the snrfile does not exist ', snrname_full)
        d = doy2ymd(year,doy); 
        month = d.month; day = d.day
        # new function to do the whole orbit thing
        foundit, f, orbdir, snrexe = get_orbits_setexe(year,month,day,orbtype) 
        # if you have the orbit file, you can get the rinex file
        if foundit:
            # now you can look for a rinex file
            rinexfile,rinexfiled = rinex_name(station, year, month, day)
            # This goes to find the rinex file. I am changing it to allow 
            # an archive preference 
             
            go_get_rinex_flex(station,year,month,day,receiverrate,archive)
# define booleans
            oexist = os.path.isfile(orbdir + '/' + f) == True
            rexist = os.path.isfile(rinexfile) == True
            exc = exedir + '/teqc' 
            texist = os.path.isfile(exc) == True
            if rexist:
                # and teqc executable exists, then 
                # get rid of all the observables i do not need
                if texist:
                    print('teqc executable exists, so let us use it')
                    foutname = 'tmp.' + rinexfile
                    fout = open(foutname,'w')
                    subprocess.call([exc, '-O.obs','S1+S2+S5+S6+S7+S8', '-n_GLONASS', '27', rinexfile],stdout=fout)
                    fout.close()
                # store it in the original rinex filename
                    subprocess.call(['rm','-f',rinexfile])
                    subprocess.call(['mv','-f',foutname, rinexfile])
                # decimate this new rinex file
                    if (rexist and dec_rate > 0): 
                        print('decimate using teqc ', dec_rate, ' seconds')
                        rinexout = rinexfile + '.tmp'; cdec = str(dec_rate)
                        fout = open(rinexout,'w')
                        subprocess.call([exc, '-O.dec', cdec, rinexfile],stdout=fout)
                        fout.close() # needed?
                        status = subprocess.call(['mv','-f', rinexout, rinexfile])
            if (oexist and rexist):
            #convert to SNR file
                snrname = snr_name(station, year,month,day,option)
                orbfile = orbdir + '/' + f
                try:
                    subprocess.call([snrexe, rinexfile, snrname, orbfile, str(option)])
                    status = subprocess.call(['rm','-f', rinexfile ])
                    print('xz compress the orbit files here')
                    status = subprocess.call(['xz', orbfile])
                except:
                    print('no success making SNR file')
                if os.path.isfile(snrname): 
#                make sure it exists and is non-zero size before moving it
                    if (os.stat(snrname).st_size == 0):
                        print('you created a zero file size which could mean a lot of things')
                        print('bad exe, bad snr option, do not really have the orbit file')
                        status = subprocess.call(['rm','-f', snrname ])
                    else:
                        print('a SNR file was created and it is non-zero in length')
                        print(snrname_full)
                        store_snrfile(snrname,year,station) 
            else:
                print('Either the rinex file or orbit file does not exist, so there is nothing to convert')

def go_get_rinex(station,year,month,day,receiverrate):
    """
    function to do the dirty work of getting a rinex file
    inputs station name, year, month, day
    """
    print('seeking ', receiverrate)
    rinexfile,rinexfiled = rinex_name(station, year, month, day)
    print('Name of the rinexfile should be:', rinexfile)
    if (os.path.isfile(rinexfile) == True):
        print('rinex file exists')
    else:
        print('go get the rinex file')
        if receiverrate == 'low':
            print('seeking low rate at unavco')
            rinex_unavco(station, year, month, day)
        else:
            print('seeking high rate at unavco')
            rinex_unavco_highrate(station, year, month, day)
        if os.path.isfile(rinexfile):
            print('you now have the rinex file')
        else:
          # only keep looking if you are seeking lowrate data
            if receiverrate == 'low':
                try:
                    rinex_sopac(station, year, month, day)
                except:
                    print('SOPAC did not work')
                if not os.path.isfile(rinexfile):
                    try:
                        rinex_cddis(station, year, month, day)
                    except:
                        print('CDDIS did not work')
                if not os.path.isfile(rinexfile):
                    try:
                        rinex_sonel(station, year, month, day)
                    except:
                        print('SONEL did not work')
                if not os.path.isfile(rinexfile):
                    print('no RINEX')
#
def go_get_rinex_flex(station,year,month,day,receiverrate,archive):
    """
    function to do the dirty work of getting a rinex file
    inputs station name, year, month, day
    20jul10 preferred RINEX archive can be set (all is everything)
    added geoscience australia and nz archives
    """
    print('requested data rate: ', receiverrate)
    rinexfile,rinexfiled = rinex_name(station, year, month, day)
    print('Name of the rinexfile should be:', rinexfile)

    if (os.path.isfile(rinexfile) == True):
        print('RINEX file exists')
    else:
        print('go get the RINEX file')
        if archive == 'all':
        # use the old code
            go_get_rinex(station,year,month,day,receiverrate) 
        else:
            if archive == 'unavco':
                if receiverrate == 'low':
                    print('seeking low rate at unavco')
                    rinex_unavco(station, year, month, day)
                else:
                    print('seeking high rate at unavco')
                    rinex_unavco_highrate(station, year, month, day)
          # only keep looking if you are seeking lowrate data
            elif (archive == 'sopac'):
                print('try sopac')
                rinex_sopac(station, year, month, day)
            elif (archive == 'cddis'):
                print('try cddis')
                rinex_cddis(station, year, month, day)
            elif (archive == 'sonel'):
                print('try sonel')
                rinex_sonel(station, year, month, day)
            elif (archive == 'nz'):
                print('try new zealand')
                rinex_nz(station, year, month, day)
            elif (archive == 'ga'):
                print('try geoscience australia, low rate')
                rinex_ga_lowrate(station,year,month,day)
            elif (archive == 'bkg'):
                print('try BKG, low rate')
                rinex_bkg(station,year,month,day)
            elif (archive == 'jeff'):
                pickup_pbay(year,month,day)
            else:
                print('eek - I have run out of archives')

def cddis_rinex3(station9ch, year, doy,srate,orbtype):
    """
    attempt to download Rinex3 files from CDDIS
    and then translate them to something useful (Rinex 2.11)

    inputs: 9 character station name (should be all capitals)
    year, day of year (doy) and sample rate (in seconds)
    station9ch can be lower or upper case - code changes it to upper case

    sending orbit type so that if nav file is used, no point writing out the non-GPS data

    returns file existence boolean and name of the RINEX 3 file (so it can be cleaned up)
    author: kristine larson
    """
#    https://cddis.nasa.gov/Data_and_Derived_Products/GNSS/RINEX_Version_3.html
    fexists = False 
    print(orbtype)
    ftp = 'ftp://cddis.nasa.gov/gnss/data/daily/'
    cdoy = '{:03d}'.format(doy)
    cyy = '{:02d}'.format(year-2000)
    cyyyy = str(year)

    f = cyyyy + '/' + cdoy + '/' + cyy + 'd'  + '/'
    new_way_f = '/gnss/data/daily/' + f
    ff = station9ch.upper() +   '_R_' + cyyyy + cdoy + '0000_01D_' + str(srate) + 'S_MO'
    smallff = station9ch[0:4].lower() + cdoy + '0.' + cyy + 'o'
    ending = '.crx' # compressed rinex3 ending
    rending = '.rnx' # rinex3 ending
    gzfilename = ff+ending + '.gz' # the crx.gz file
    filename = ff+ending  # the crx file
    rfilename = ff+rending # the rnx file
    url = ftp + f + gzfilename  
    print(url)
    # not sure i still neeed this
    exedir = os.environ['EXE']
    gexe = gfz_version()
    crnxpath = hatanaka_version()
    if orbtype == 'nav':
        # added this bevcause weird Glonass data were making teqc unhappy
        gobblygook = 'G:S1C,S2X,S2L,S2S,S5'
    else:
    # I hate S2W 
    # should add beidou here???
        gobblygook = 'G:S1C,S2X,S2L,S2S,S5+R:S1P,S1C,S2P,S2C+E:S1,S5,S6,S7,S8'
    print(gobblygook)
    if os.path.isfile(rfilename):
        print('rinex3 file already exists')
    else:
        print(gzfilename)
        print(new_way_f)
        try:
            #wget.download(url,gzfilename)
            # using new protocol
            cddis_download(gzfilename,new_way_f)
            #wget.download(url,gzfilename)
            # unzip and Hatanaka decompress
            subprocess.call(['gunzip',gzfilename])
            subprocess.call([crnxpath,filename])
            # remove the crx file
            subprocess.call(['rm',filename])
        except:
            print('no file at CDDIS')

    if os.path.isfile(rfilename) and os.path.isfile(gexe):
        print('making rinex 2.11')
        try:
            subprocess.call([gexe,'-finp', rfilename, '-fout', smallff, '-vo','2','-ot', gobblygook, '-f'])
            print('woohoo!')
            print('look for the rinex 2.11 file here: ', smallff)
            fexists = True
        except:
            print('some kind of problem in translation to 2.11')
    else:
        print('either the rinex3 file does not exist OR the gfzrnx executable does not exist')

    return fexists, rfilename
def bkg_rinex3(station9ch, year, doy,srate):
    """
    download rinex 3 from BKG
    inputs: 9 character station name, year, day of year and srate  
    is sample rate in seconds
    note: not sure this works
    """
    cdoy = '{:03d}'.format(doy)
    cyy = '{:02d}'.format(year-2000)
    csrate = '{:02d}'.format(srate)
    cyyyy = str(year)
    url = 'ftp://igs.bkg.bund.de/EUREF/obs/' + cyyyy + '/' + cdoy + '/'
    ff = station9ch.upper() +   '_R_' + cyyyy + cdoy + '0000_01D_' + csrate + 'S_MO' + '.crx.gz'
    url = url + ff
    try:
        wget.download(url,ff)
    except:
        print('problem with bkg download')

def rinex3_rinex2(gzfilename,v2_filename):
    """
    input gzfile name
    gunzip, de-hatanaka, then convert to RINEX 2.11
    WARNING: this doesn't work yet
    """
    fexists = False
    gexe = gfz_version()
    hexe = hatanaka_version()
    # I hate S2W for this version, only GPS
    # / OBS TYPES AUBG claims to have these
    # S1C   S2S S2W S5Q
    gobblygook = 'G:S1C,S5'
    gobblygook = 'G:S1C,S2X,S2L,S2S,S5'
    l=len(gzfilename)
    cnxfilename = gzfilename[0:l-3]
    rnxfilename = gzfilename[0:l-6] + 'rnx'
    print(cnxfilename)
    print(rnxfilename)
    if os.path.isfile(gzfilename):
        # unzip and Hatanaka decompress
        subprocess.call(['gunzip',gzfilename])
        subprocess.call([hexe,cnxfilename])
    if os.path.isfile(rnxfilename) and os.path.isfile(gexe):
        print('making rinex 2.11 of this file')
        try:
            subprocess.call([gexe,'-finp', rnxfilename, '-fout', v2_filename, '-vo','2','-ot', gobblygook, '-f'])
            print('woohoo!')
            print('look for the rinex 2.11 file here: ', v2_filename)
            fexists = True
        except:
            print('some kind of problem in translation to 2.11')
    else:
        print('either the rinex3 file does not exist OR the gfzrnx executable does not exist')

    return fexists

def hatanaka_version():
    """
    return string with location of hatanaka executable
    """
    exedir = os.environ['EXE']
    hatanakav = exedir + '/CRX2RNX'
    # heroku version should be in the main area
    if not os.path.exists(hatanakav):
        hatanakav = './CRX2RNX'
    return hatanakav

def gfz_version():
    """
    return string with location of gfzrnx executable
    """
    exedir = os.environ['EXE']
    gfzv = exedir + '/gfzrnx'
    # heroku version should be in the main area
    if not os.path.exists(gfzv):
        gfzv = './gfzrnx'
    return gfzv

def gpsSNR_version():
    """
    return string with location of gpsSNR executable
    """
    exedir = os.environ['EXE']
    gpse = exedir + '/gpsSNR.e'
    # heroku version should be in the main area
    if not os.path.exists(gpse):
        gpse = './gpsSNR.e'
    return gpse

def gnssSNR_version():
    """
    return string with location of gnssSNR executable
    """
    exedir = os.environ['EXE']
    gpse = exedir + '/gnssSNR.e'
    # heroku version should be in the main area
    if not os.path.exists(gpse):
        gpse = './gnssSNR.e'
    return gpse

def teqc_version():
    """
    return string with location of teqcexecutable
    author: kristine larson
    """
    exedir = os.environ['EXE']
    gpse = exedir + '/teqc'
    # heroku version should be in the main area
    if not os.path.exists(gpse):
        gpse = './teqc'
    return gpse

def snr_exist(station,year,doy,snrEnd):
    """
    given station name, year, doy, snr type
    returns whether snr file exists on your machine
    bizarrely snrEnd is a character string
    year and doy are integers, which makes sense!
    author: Kristine Larson

    """
    xdir = os.environ['REFL_CODE']
    cdoy = '{:03d}'.format(doy)
    cyy = '{:02d}'.format(year-2000)
    f= station + cdoy + '0.' + cyy + '.snr' + snrEnd
    fname = xdir + '/' + str(year) + '/snr/' + station + '/' + f
    fname2 = xdir + '/' + str(year) + '/snr/' + station + '/' + f  + '.xz'
    snre = False
    # check for both
    if os.path.isfile(fname):
        snre = True
    if os.path.isfile(fname2):
        snre = True

    return snre 


def unavco_rinex3(station9ch, year, doy,srate,orbtype):
    """
    attempt to download Rinex3 files from UNAVCO
    and then translate them to something useful (Rinex 2.11)
    inputs: 9 character station name (can be upper or lower, as code will change to all 
    uppercase)

    year, day of year (doy) and sample rate (in seconds)

    orbtype - nav, sp3 etc

    if nav file orbtype is used, no point writing out the non-GPS data

    returns file existence boolean and name of the RINEX 3 file (so it can be cleaned up)


    warning - i do not write out the L2P data - only the good L2C signals.

    author: kristine larson
    """
    fexists = False 
    print(orbtype)
    cdoy = '{:03d}'.format(doy)
    cyy = '{:02d}'.format(year-2000)
    csrate = '{:02d}'.format(srate)
    cyyyy = str(year)
    ftp = 'ftp://data-out.unavco.org/pub/rinex3/obs/' 
    #ftp = 'ftp://cddis.nasa.gov/gnss/data/daily/'

    f = cyyyy + '/' + cdoy + '/' 
    ff = station9ch.upper() +   '_R_' + cyyyy + cdoy + '0000_01D_' + csrate + 'S_MO'
    smallff = station9ch[0:4].lower() + cdoy + '0.' + cyy + 'o'
    ending = '.crx' # compressed rinex3 ending
    rending = '.rnx' # rinex3 ending
    gzfilename = ff+ending + '.gz' # the crx.gz file
    filename = ff+ending  # the crx file
    rfilename = ff+rending # the rnx file
    url = ftp + f + gzfilename  
    print(url)
    # not sure i still neeed this
    exedir = os.environ['EXE']
    gexe = gfz_version()
    crnxpath = hatanaka_version()
    if orbtype == 'nav':
        # added this bevcause weird Glonass data were making teqc unhappy
        gobblygook = 'G:S1C,S2X,S2L,S2S,S5'
    else:
    # I hate S2W  - so it is not written out
        gobblygook = 'G:S1C,S2X,S2L,S2S,S5+R:S1P,S1C,S2P,S2C+E:S1,S5,S6,S7,S8'
    print(gobblygook)
    if os.path.isfile(rfilename):
        print('rinex3 file already exists')
    else:
        try:
            wget.download(url,gzfilename)
            # unzip and Hatanaka decompress
            subprocess.call(['gunzip',gzfilename])
            subprocess.call([crnxpath,filename])
            # remove the crx file
            subprocess.call(['rm',filename])
        except:
            print('no file at CDDIS')

    if os.path.isfile(rfilename) and os.path.isfile(gexe):
        print('making rinex 2.11')
        try:
            subprocess.call([gexe,'-finp', rfilename, '-fout', smallff, '-vo','2','-ot', gobblygook, '-f'])
            print('woohoo!')
            print('look for the rinex 2.11 file here: ', smallff)
            fexists = True
        except:
            print('some kind of problem in translation to 2.11')
    else:
        print('either the rinex3 file does not exist OR the gfzrnx executable does not exist')

    return fexists, rfilename

def get_cddis_navfile(navfile,cyyyy,cyy,cdoy):
    """
    kristine larson
    inputs navfile name with character string verisons of year, 2ch year and doy
    tries to download from CDDIS archive

    20jun11, implemented new CDDIS security requirements
    """
    # ths old way
    # just in case you sent it the navfile with auto instead of brdc
    cddisfile = 'brdc' + navfile[4:] + '.Z'
    cddis = 'ftp://cddis.nasa.gov'
    # navfile will continue to be called auto
    navfile_compressed = cddisfile
    url_cddis = cddis + '/gps/data/daily/' + cyyyy + '/' + cdoy + '/' +cyy + 'n/' + navfile_compressed
    #print(url_cddis)

    # new way
    cddisfile = 'brdc' + cdoy + '0.' +cyy  +'n'
    navfile_compressed = cddisfile + '.Z'
    # where the file should be at CDDIS ....
    mdir = '/gps/data/daily/' + cyyyy + '/' + cdoy + '/' +cyy + 'n/'

    try:
        #wget.download(url_cddis,navfile_compressed)
        cddis_download(navfile_compressed,mdir)
        subprocess.call(['uncompress',navfile_compressed])
        print('success at cddis')
        subprocess.call(['mv',cddisfile,navfile])
    except:
        print('no success at cddis')
    return navfile



def cddis_download(filename, directory):
    """
    https://cddis.nasa.gov/Data_and_Derived_Products/CDDIS_Archive_Access.html
    attempt to use more secure download protocol that is CDDIS compliant

    input: filename and directory (without leading location)

    this will replace using wget.download when CDDIS turns off anonymous ftp

    was supposed to returns whether file was created but now it just returns true


    """
    filename = 'ftps://gdc.cddis.eosdis.nasa.gov' + directory + filename 
    callit = ['wget', '--ftp-user','anonymous','--ftp-password', 'kristine@colorado.edu', filename]
    #print(callit)
    #print(filename)
    #print(filename)
    subprocess.call(callit)
    return True 

def pickup_pbay(year,month, day):
    """
    jeff freymueller's site. L2C, but GPS only.
    changed to year,month,day to be consistent with other
    """
    print('downloading highrate PBAY')
    station = 'pbay'
    if (day == 0):
        doy = month
        year, month, day, cyyyy,cdoy, YMD = ydoy2useful(year,doy)
        cyy = cyyyy[2:4]
    else:
        doy,cdoy,cyyyy,cyy = ymd2doy(year,month,day)

    url = 'ftp://gps.alaska.edu/pub/gpsdata/permanent/C2/' + cyyyy + '/' + cdoy + '/'
    fname = station + cdoy + '0.' + cyyyy[2:4] + 'o.gz'
    url = url + fname
    try:
        wget.download(url,fname)
        subprocess.call(['gunzip', fname])
    except:
        print('problems downloading pbay')

    return True

def ydoy2useful(year, doy):
    """
    inputs: year and day of year (doy), integers
    returns: useful stuff, like month and day and character 
    strings for year and doy and YMD as character string ....
    """

    d = datetime.datetime(year, 1, 1) + datetime.timedelta(days=(doy-1))
#    print('ymd',d)
#   not sure you need to do this int step
    month = int(d.month)
    day = int(d.day)
    cdoy = '{:03d}'.format(doy)
    cyyyy = '{:04d}'.format(year)
    cyy = '{:02d}'.format(year-2000)
    cdd = '{:02d}'.format(day)
    cmonth = char_month_converter(month)
    YMD = cyy + cmonth + cdd
    return year, month, day, cyyyy,cdoy, YMD

def prevdoy(year,doy):
    """
    given year and doy, return previous year and doy
    """
    if (doy == 1):
        pyear = year -1
        doyx,cdoyx,cyyyy,cyy = ymd2doy(pyear,12,31)
        pdoy = doyx
    else:
#       doy is decremented by one and year stays the same
        pdoy = doy - 1
        pyear = year

    return pyear, pdoy

def nextdoy(year,doy):
    """
    given year and doy, return next year and doy
    """
    dec31,cdoy,ctmp,ctmp2 = ymd2doy(year,12,31)
    if (doy == dec31):
        nyear = year + 1
        ndoy = 1
    else:
        nyear = year
        ndoy = doy + 1

    return nyear, ndoy