Initial commit

1 files changed, 119 insertions, 0 deletions

diff --git a/main.py b/main.py
new file mode 100644
index 0000000..8187e1f
--- /dev/null
+++ b/main.py
@@ -0,0 +1,119 @@
+from banapedia.Ban import *
+from collections import Counter
+import json
+import pygal
+import numpy as np
+import urllib.request
+
+__author__ = 'pacien'
+
+
+BAN_MAP_FILE = "output/ban-map.svg"
+BAN_DURATION_MAP_FILE = "output/ban-duration-map.svg"
+HIST_FILE = "output/histogram.svg"
+
+BAN_FILE = "resources/ban_list.json"
+
+SAMPLES = 30000
+SAMPLES_BY_QUERY = 500
+
+
+def configure_proxy():
+    proxy = urllib.request.ProxyHandler(urllib.request.getproxies())
+    opener = urllib.request.build_opener(proxy)
+    urllib.request.install_opener(opener)
+
+
+def load_from_internet():
+    configure_proxy()
+    return fetch_multipart_ban_dict(SAMPLES, SAMPLES_BY_QUERY)
+
+
+def load_from_local():
+    with open(BAN_FILE, "r") as ban_dict_file:
+        return json.load(ban_dict_file)
+
+
+def write_to_local(ban_dict_list):
+    with open(BAN_FILE, "w") as ban_dict_file:
+        json.dump(ban_dict_list, ban_dict_file, indent="\t")
+
+
+# ban_dict_list = load_from_internet()
+# write_to_local(ban_dict_list)
+
+ban_dict_list = load_from_local()
+
+ban_list = map_bans(ban_dict_list)
+
+
+########## HISTOGRAM ##########
+
+ban_durations = [ban.get_duration() for ban in ban_list]
+(ban_durations, bins) = np.histogram(ban_durations, bins=[round(365/12*x) for x in range(1, 50+2)])
+
+print("[INFO]", "Generating histogram")
+bar_chart = pygal.Bar(legend_at_bottom=True)
+bar_chart.title = "Active Wikipedia bans by duration (%d samples)" % SAMPLES
+bar_chart.x_labels = map(str, range(1, len(ban_durations)+1))
+bar_chart.add("Number of active bans", ban_durations)
+bar_chart.render_to_file(HIST_FILE)
+print("[INFO]", "Histogram generation complete")
+
+
+########## NB BAN MAP ##########
+
+def count_by_country(ban_list):
+    country_ban_list = [ban.get_country_code() for ban in ban_list]
+    return Counter(country_ban_list)
+
+nb_bans_by_country = count_by_country(ban_list)
+
+print("[INFO]", "Generating ban map")
+worldmap_chart = pygal.Worldmap(legend_at_bottom=True)
+worldmap_chart.title = "World active Wikipedia bans by country (%d samples)" % SAMPLES
+worldmap_chart.add("Active bans", nb_bans_by_country)
+worldmap_chart.render_to_file(BAN_MAP_FILE)
+print("[INFO]", "Ban map generation complete")
+
+
+########## BAN DURATION MAP ##########
+
+def group_by_country(ban_list):
+    ban_duration_by_country = {}
+
+    for ban in ban_list:
+        country_code = ban.get_country_code()
+
+        if country_code not in ban_duration_by_country.keys():
+            ban_duration_by_country[country_code] = []
+
+        ban_duration_by_country[country_code].append(ban)
+
+    return ban_duration_by_country
+
+
+def calc_average_ban_by_country(ban_by_country_dict):
+    average_ban_duration_by_country = {}
+
+    for country, bans in ban_by_country_dict.items():
+        average = np.mean([ban.get_duration() for ban in bans])
+        average_ban_duration_by_country[country] = average
+
+    return average_ban_duration_by_country
+
+ban_duration_by_country = group_by_country(ban_list)
+average_ban_duration_by_country = calc_average_ban_by_country(ban_duration_by_country)
+
+average_ban_duration_by_country = {country: duration/30 for country, duration in average_ban_duration_by_country.items()}
+
+print("[INFO]", "Generating ban duration map")
+worldmap_chart = pygal.Worldmap(legend_at_bottom=True)
+worldmap_chart.title = "Average Wikipedia ban duration by country (%d samples)" % SAMPLES
+worldmap_chart.add("Average ban duration (months)", average_ban_duration_by_country)
+worldmap_chart.render_to_file(BAN_DURATION_MAP_FILE)
+print("[INFO]", "Ban duration map generation complete")
+
+print("\nTHIS WAS A TRIUMPH!")
+print("I'M MAKING A NOTE HERE:")
+print("HUGE [SUCCESS]")