Gets all the F1 race results from 1950 to 2017. The results are flatten so that they are easily utilized in mapreduce.

f1-results.py

#!/usr/bin/python

import sys
import urllib2
import time
import json

def get_results(output_dir):
    for year in range(1950, 2017 + 1):
        for rnd in range(1, 50):
            results_url = "http://ergast.com/api/f1/{0}/{1}/results.json".format(year, rnd)
            results = urllib2.urlopen(results_url).read()

            # Flatten results.
            race_json = json.loads(results)
            if len(race_json["MRData"]["RaceTable"]["Races"]) is 0:
                print "Round {0} is Invalid".format(rnd)
                break;
                
            circuit = race_json["MRData"]["RaceTable"]["Races"][0]["Circuit"]["circuitId"]
            results_json = race_json["MRData"]["RaceTable"]["Races"][0]["Results"]

            results = ""
            for result_json in results_json:
                #print race_json
                constructor_json = result_json["Constructor"]
                driver_json = result_json["Driver"]
                status = result_json["status"]

                time_milli = "-1" # Invalid
                if "Time" in result_json:
                    time_milli = result_json["Time"]["millis"]

                position = result_json["position"]
                results += "{0},{1},{2},{3},{4},{5},{6}\n".format(
                    year,
                    circuit,
                    constructor_json["constructorId"],
                    driver_json["driverId"],
                    status,
                    time_milli,
                    position);

            output_path = "{0}/{1}-{2}.txt".format(output_dir, year, circuit)
            print output_path

            output_file = open(output_path, "w")
            output_file.write(results)
            output_file.close()

            time.sleep(2)  # Avoid DOS signature.
        

if __name__ == "__main__":
    output_dir = str(sys.argv[1])
    get_results(output_dir)