graphing.py

import networkx as nx
import utils
import csv
from math import sqrt
import matplotlib.pyplot as plt

""" Creates a list of tuples of voting data from a CSV and creates a NetworkX graph with the according nodes and edge weights """
def create_graph(csvFileName):
    tupleList = utils.from_csv(csvFileName)
    graphyBoi = nx.Graph()
    for tupleBoi in tupleList:
        graphyBoi.add_edge(tupleBoi[0], tupleBoi[1], weight=float(tupleBoi[4]))
    return graphyBoi

def extract_member_map(file="data/115nom.csv"):
    """
    returns a dictionary from member bioguide_id's to dw_nominate tuples
    """
    f = open(file, "r")
    csv_reader = csv.DictReader(f)
    member_map = {}
    for row in csv_reader:
        bioguide_id = row["bioguide_id"]
        if (row["bioname"] and row["nominate_dim1"] and row["nominate_dim2"]):
            member_map[bioguide_id] = (float(row["nominate_dim1"]), float(row["nominate_dim2"]))
    f.close()
    return member_map

def create_naive_graph(congress, chamber):
    assert chamber in utils.CHAMBERS
    path = "naive_{}_{}_metric.csv".format(congress, chamber)
    return create_graph(path)

def create_dw_graph():
    """
    creates a graph with dw_nominate similarity scores
    """
    g = nx.Graph()
    euclid = lambda pair1,pair2: sqrt((pair1[0]-pair2[0])**2 + (pair1[1]-pair2[1])**2)
    member_map = extract_member_map()
    for memb1 in member_map:
        for memb2 in member_map:
            if memb1 != memb2:
                scores1, scores2 = member_map[memb1], member_map[memb2]
                dist = euclid(scores1,scores2)
                #weight = 1/(1 + dist)
                if (dist != 0):
                    weight = 1/dist
                else:
                    weight = 1/(.001**2)
                #if weight > .5:
                g.add_edge(memb1, memb2, weight=weight)
    """minn = 1/(1+sqrt(2)*2)
    maxx = 1 - minn
    for u, v, d in g.edges(data=True):
        d['weight'] -= minn
        d['weight'] /= maxx"""
    return g

def plot_clustering(clustering, g):
    values = [clustering.get(node) for node in g.nodes]
    nx.draw_spring(g, cmap=plt.get_cmap('jet'), node_color = values, node_size=30, with_labels=False)
    plt.show()

def pad_graph(g):
    for v1 in g.nodes:
        for v2 in g.nodes:
            if v1 == v2:
                continue
            if not g.has_edge(v1,v2):
                g.add_edge(v1,v2, weight=0)