Added script for weather API assignment

README_api_fetcher.txt

@@ -0,0 +1,101 @@
+--------- Short Description ----------
+
+This R script downloads, cleans and creates a data frame out of all the .json files derived from the weather API automatically, then prints a .csv file. Should work on every machine.
+
+The variables are selected from the "historic hourly" data frame are only date and temperature.
+
+The selected period is from 13/12/2021 at 00:00:00 to 09/01/22 23:00:00.
+
+The files are bound together by the union() function at every cycle.
+
+REMEMBER to delete the # at the last line of code and change the path and name to get a proper .csv file at the end.
+
++++++ ONLY in case you need to modify anything read below ++++++
+
+---------- Starting and ending period -----------
+
+What determines the starting period is the date determined by the various url pieces plus the "num1" and "num2" counters.
+
+For example:
+
+url1 <- as.character("https://raw.githubusercontent.com/data-hydenv/data/master/extra/weather/data/2021_12_")
+
+num1 <- as.character("12")
+
+url2 <- as.character("_raw_dump.json")
+
+makes it so that the 
+
+urlt <- paste(url1, num1, url2, sep= "")
+
+yields https://raw.githubusercontent.com/data-hydenv/data/master/extra/weather/data/2021_12_12_raw_dump.json
+
+where the date is 2021_12_12.
+
+So to modify the starting period you have to adjust url1 and num1. Same goes for url3 and num2 regarding January or any months different than the one in url1.
+
+The starting number is 12 and the cycles run further than the 2022-01-09 since the historic data are compiled in such a manner that you need the previous and the next days to get a full "present" day.
+
+That is nothing to worry about, as at the end of the cycles, the data frame is filtered so that it stays within to the desired dates.
+
+It has to be noted that also the while conditions have to be changed in order to keep the cycles running for the desired amount of covered days.
+
+NOTE that as the file for 2021-12-27 is missing, a condition has to setup in order not to get error messages and keep the while machine running.
+
+num1 <- ifelse(num1 == 26, as.integer(num1) + 2, as.integer(num1) + 1)
+
+That means that num1 = 27 is skipped.
+
+To adapt it to a different period you also have to get rid of it in order to get it working properly.
+
+VARIABLES and CODE TO MODIFY in order to make it run on a different period
+
+url1
+
+num1
+
+url3
+
+num2
+
+line 52
+
+line 66
+
+--------- Variables from .json ---------
+
+dttm and th (temperature)
+
+Those are selected from the "historic hourly" dataframe by:
+
+jdf <- js$historic$hourly %>% as.data.frame()
+
+then mutate the date and select just those two:
+
+ dfor <- jdf %>% select(dt, temp)%>% 
+    mutate(th = temp) %>% 
+    mutate(dttm = as.POSIXct(dt, origin ="1970-01-01")) %>% 
+    select(dttm, th)
+
+"dfor" is the single dataframe yielded by the current .json file.
+
+By modifying jdf one can access to different nested dataframes within the json.
+
+VARIABLES and CODE TO MODIFY to get different variables
+
+jdf
+
+dfor
+
+- Yielded Data Frame -
+
+Before the cycles we create an empty df "dtt".
+
+dtt <- data.frame(dttm =as.Date(character()), th = as.double(double())) 
+
+Every cycle it unites with itself plus "dfor", thus containing every .json desired data.
+
+dtt <- union(dtt, dfor)
+
+
+

api_weather_data_fetcher.R

@@ -0,0 +1,70 @@
+#---- libraries ----
+
+library(jsonlite)
+library(dplyr)
+
+#---- set up url ----
+
+url1 <- as.character("https://raw.githubusercontent.com/data-hydenv/data/master/extra/weather/data/2021_12_")
+
+num1 <- as.character("12")
+
+url2 <- as.character("_raw_dump.json")
+
+url3 <- as.character("https://raw.githubusercontent.com/data-hydenv/data/master/extra/weather/data/2022_1_")
+
+urlt <- paste(url1, num1, url2, sep= "")
+urlt
+
+# for january
+
+num2 <- as.character("1")
+
+# ---- empty data frame
+
+dtt <- data.frame(dttm =as.Date(character()), th = as.double(double())) 
+
+##---- while machine ----
+
+while (as.integer(num1) < 42) {
+
+  #---- json importer  ----  
+
+  js <- fromJSON(urlt, flatten =TRUE)
+
+  jdf <- js$historic$hourly %>% as.data.frame()
+
+  dfor <- jdf %>% select(dt, temp)%>% 
+    mutate(th = temp) %>% 
+    mutate(dttm = as.POSIXct(dt, origin ="1970-01-01")) %>% 
+    select(dttm, th)
+
+  #---- create data frame ----
+
+  dtt <- union(dtt, dfor)
+
+  #---- names and url update ----
+
+  ifelse(num1 < 31, #condition yes = december
+
+         {
+
+           num1 <- ifelse(num1 == 26, as.integer(num1) + 2, as.integer(num1) + 1)
+
+           urlt <- paste(url1, as.character(num1), url2, sep= "")}, # yes
+
+         { # no = january
+
+           num1 <- as.integer(num1) + 1
+
+           urlt <- paste(url3, num2, url2, sep= "")
+
+           num2 <- as.integer(num2) + 1} # no
+
+  )}
+
+dtt <- dtt %>% filter(dttm > "2021-12-13", dttm < "2022-01-10")
+
+# ++++++++ remember to change the path and file name below +++++++++
+
+#write.csv(dtt, file = '~/your_path/file_name.csv', row.names = FALSE, quote = FALSE)