Protoball:Count Game Tab Sources.py
Jump to navigation
Jump to search
def count_region_sources(region_text): regions = region_text.split("\n") from urllib import request from bs4 import BeautifulSoup import re ref_item = re.compile("\(\d+\)") total_sources = 0 for region in regions: r = request.urlopen("http://protoball.org/Game_Tab:" + region.replace(" ", "_")) t = r.read() soup = BeautifulSoup(t) tables = soup.select("table.nice") rows = [] for table in tables: rows.extend(table.select("tr")[1:]) sources = 0 for row in rows: source_cell = row.find_all("td")[-1] text = source_cell.get_text() sources += len(ref_item.findall(text)) print(region, ": ", sources) total_sources += sources print("Total Sources", ": ", total_sources)