WAQWAQProject

next49 さんがWAQWAQプロジェクト2014で進行状況チェックに 困ってるらしいことを見つけた。

ちょっとPythonのWebスクレイピングの実験も兼ねて書いてみた。

# coding: utf-8

import datetime
import re
import sys
import codecs
import urllib
from bs4 import BeautifulSoup
import xlsxwriter

reload(sys)
sys.setdefaultencoding('utf-8')


def get_wikipedia_history(username):
    urlparts = u"特別:投稿記録"
    urlparts_quote = urllib.quote(urlparts.encode("utf-8"))
    username_quote = urllib.quote_plus(username.encode("utf-8"))
    url = u"http://ja.wikipedia.org/w/index.php?limit=1000&tagfilter=&title=%s&contribs=user&target=%s&namespace=0&tagfilter=&year=2014&month=-1" % (
        urlparts_quote, username_quote)
    req = urllib.urlopen(url)
    soup = BeautifulSoup(req)

    result = []

    print("****************************************")
    for li in soup.ul.find_all("li"):
        item = {}

        print(li)

        a_date = li.find(class_=["mw-changeslist-date", "history-deleted"])
        a_date_str = a_date.string
        # m = re.match(u"(\d+)年(\d+)月(\d+)日（.*）(\d+):(\d+)",a_date_str)
        m = re.match(u"(\d+)年(\d+)月(\d+)日.*(\d+):(\d+)", a_date_str)
        a_y = int(m.group(1))
        a_m = int(m.group(2))
        a_d = int(m.group(3))
        a_h = int(m.group(4))
        a_min = int(m.group(5))
        item["dt"] = datetime.datetime(a_y, a_m, a_d, a_h, a_min)

        abbr = li.abbr
        if abbr is None:
            if li.find(class_="history-deleted") is not None:
                item["type"] = "Delete"
            else:
                item["type"] = "Edit"
        else:
            if abbr["class"][0] == "newpage":
                item["type"] = "New"
            elif abbr["class"][0] == "minoredit":
                item["type"] = "MinorEdit"
        a_title = li.find(class_="mw-contributions-title")
        item["title"] = a_title.string
        plusminus = li.find(
            class_=["mw-plusminus-pos", "mw-plusminus-null", "mw-plusminus-neg"])
        m = re.match(r"\([+-]?([0-9,]+)\)", plusminus.string)
        item["pm"] = int(m.group(1).replace(",", ""))
        print item
        print("--------------------------------------")
        result.append(item)
    result.reverse()
    return result


def write_worksheet(wb, username, changes, date_format):
    ws = wb.add_worksheet(username)
    ws_c = wb.add_chartsheet(u"%s_c" % username)

    ws.write("A1", u"変更日時")
    ws.write("B1", u"変更種別")
    ws.write("C1", u"変更バイト数")
    ws.write("D1", u"ページ名")
    ws.write("E1", u"スコア")

    scores = {"New": 5, "Edit": 3, "MinorEdit": 1, "Delete": 2}

    row = 0
    score = 0
    for change in changes:
        row += 1
        score += scores[change["type"]]
        ws.write(row, 0, change["dt"], date_format)
        ws.write(row, 1, change["type"])
        ws.write(row, 2, change["pm"])
        ws.write(row, 3, change["title"])
        ws.write(row, 4, score)

    row += 1
    ws.write(row, 0, datetime.datetime.now(), date_format)
    print "row = %d" % row

    chart = wb.add_chart({"type": "line"})
    chart.add_series({
        "name": username,
        "categories": [ws.get_name(), 1, 0, row, 0],
        "values": [ws.get_name(), 1, 4, row, 4],
#        "data_labels": {"value": True},
        })
    chart.set_legend({"none": True})
    ws_c.set_chart(chart)
    return

if __name__ == "__main__":
    wb = xlsxwriter.Workbook("waqwaq.xlsx")
    date_format = wb.add_format({"num_format": "yyyy/mm/dd hh:mm:ss"})
    #names = [u"Next49", u"Mishika", u"Theta K", u"蒋龍"]
    names = [u"Next49"]
    for username in names:
        result = get_wikipedia_history(username)
        write_worksheet(wb, username, result, date_format)

    wb.close()

waqwaq.xlsx こんなファイルが出来上がる。

TODO:

• コメント書く
• MS-Excel で表示内容を確認する（Libre Office でしか見てない）
• matplotlib を使った png/svg の出力を試してみる
• 期間を限定する
• スコアの重み付けの見直しをする
• バイト数や、新規ページ数などでもグラフを作る

といった微修正が残っているが、まあ、まずはプロトタイプとしてこんな感じでどうだろうか。