PTT gamesale 板的買賣文
# -*- coding: utf-8 -*-
# author: yinxuanh@gmail.com
# version: 1.0
# github: https://github.com/yin-xuanHuang/web_crawler_ptt_gamesale
# license: MIT license
import requests
import sqlite3
import re
from bs4 import BeautifulSoup
#from time import sleep
# 文章內容處理
def catch_content_return_list(content_list, sellORwanted):
# 找出目標內容，並擷取
find_all_body_list = content_list
content_wanted_index = 0
for body_string in find_all_body_list:
if u'【' in body_string: break
content_wanted_index = content_wanted_index + 1
try:
content_wanted = find_all_body_list[content_wanted_index]
except:
return []
while True:
try:
if find_all_body_list[content_wanted_index+1][0] == u"※": break
except:
break
content_wanted = content_wanted + find_all_body_list[content_wanted_index + 1]
content_wanted_index = content_wanted_index +1
# 當標題非售非徵
if sellORwanted != u'售' and sellORwanted != u'徵':
if u'徵' in content_wanted:
sellORwanted = u'徵'
else:
sellORwanted = u'售'
# 消除特定字元 TODO: 多遊戲同一文章之間隔區隔問題
body_content = content_wanted
body_content = body_content.replace("\n", "")
body_content = body_content.replace(" ", "")
body_content = body_content.replace(u'★', "")
body_content = body_content.replace(u' ', "")
# 消除可能的"範例文字"
# e.x.←可直接使用，無須編輯
if u"←可直接使用，無須編輯" in body_content:
body_content = body_content.replace(u"←可直接使用，無須編輯", "")
# e.x.（限制級、輔15級、輔12級、保護級、普遍級）
if u"（限制級、輔15級、輔12級、保護級、普遍級）" in body_content:
body_content = body_content.replace(u"（限制級、輔15級、輔12級、保護級、普遍級）", "")
# e.x.←分級提示
if u"←分級提示" in body_content:
body_content = body_content.replace(u"←分級提示", "")
# e.x.（未照左方分級者違規）
if u"（未照左方分級者違規）" in body_content:
body_content = body_content.replace(u"（未照左方分級者違規）", "")
# e.x.（徵求者、配件等無分級者可免填，販售國內未代理之遊戲，請直接填限制級即可。）
if u"（徵求者、配件等無分級者可免填，販售國內未代理之遊戲，請直接填限制級即可。）" in body_content:
body_content = body_content.replace(u"（徵求者、配件等無分級者可免填，販售國內未代理之遊戲，請直接填限制級即可。）", "")
# e.x.（遊戲有多語系版本必填，海外版請註明是哪個國家。）
if u"（遊戲有多語系版本必填，海外版請註明是哪個國家。）" in body_content:
body_content = body_content.replace(u"（遊戲有多語系版本必填，海外版請註明是哪個國家。）", "")
body_content = body_content.replace(u'。', "")
body_content = body_content[:-2] + u'【'
# 處理文章
try:
the_colon = body_content.index(u'：')
the_brackets = body_content.index(u'【')
next_brackets = body_content.index(u'【', the_brackets + 1)
content_dic = dict()
while True:
dic_name = body_content[the_brackets + 1:the_colon - 1]
content_dic[dic_name] = body_content[the_colon + 1:next_brackets]
if u'：' not in body_content[the_colon + 1:] or u'【' not in body_content[next_brackets + 1:]: break
the_brackets = next_brackets
next_brackets = body_content.index(u'【', next_brackets + 1)
the_colon = body_content.index(u'：', the_brackets)
except:
return []
# 排序內容，回傳有排序的序列
# TODO : "交換" 選項
content_list = list()
i = 0
while i < 9:
if i == 0: # 售or徵求 0
content_list.append(sellORwanted)
i = i + 1
check = False
for item in content_dic:
if u"名" in item and i == 1: # 物品名稱 1
content_list.append(content_dic.get(item," "))
check = True
break
elif u"分" in item and i == 2: # 遊戲分級 2
content_list.append(content_dic.get(item," "))
check = True
break
elif u"語" in item and i == 3: # 語系版本 3
content_list.append(content_dic.get(item," "))
check = True
break
elif u"價" in item and i == 4: # 徵求價 or 售價 4
content_list.append(content_dic.get(item," "))
check = True
break
elif u"易" in item and i == 5: # 交易方式 5
content_list.append(content_dic.get(item," "))
check = True
break
elif u"保" in item and i == 6: # 保存狀況 6
content_list.append(content_dic.get(item," "))
check = True
break
elif u"地" in item and i == 7: # 地區 7
content_list.append(content_dic.get(item," "))
check = True
break
elif u"附" in item and i == 8: # 附註 8
content_list.append(content_dic.get(item," "))
check = True
break
if not check:
content_list.append(" ")
i = i + 1
return content_list
# 主程式開始
# connect database
conn = sqlite3.connect('ptt_gamesale.sqlite')
cur = conn.cursor()
# TODO 適當資料型態
# set database
# author, ok
# date, ok
# gamestation, ok
# title_id, titleID ok
# sellORwanted 0 ok
# 物品名稱 1 item_name ok
# 遊戲分級 2 law_limited ok
# 語系版本 3 language ok
# 徵求價 or 售價 4 price ok
# 交易方式 5 howtotrade ok
# 保存狀況 6 status ok
# 地區 7 address ok
# 附註 8 message ok
cur.executescript('''
CREATE TABLE IF NOT EXISTS Content(
id INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT UNIQUE,
titleID TEXT UNIQUE,
author_id INTEGER,
sellORwanted_id INTEGER,
gamestation_id INTEGER,
date TEXT,
item_name TEXT,
law_limited_id INTEGER,
language_id INTEGER,
price TEXT,
howtotrade TEXT,
status TEXT,
address TEXT,
message TEXT
);
CREATE TABLE IF NOT EXISTS Author (
id INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT UNIQUE,
name TEXT UNIQUE
);
CREATE TABLE IF NOT EXISTS SellORwanted (
id INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT UNIQUE,
name TEXT UNIQUE
);
CREATE TABLE IF NOT EXISTS Gamestation (
id INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT UNIQUE,
name TEXT UNIQUE
);
CREATE TABLE IF NOT EXISTS Law_limited (
id INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT UNIQUE,
name TEXT UNIQUE
);
CREATE TABLE IF NOT EXISTS Language (
id INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT UNIQUE,
name TEXT UNIQUE
);
CREATE TABLE IF NOT EXISTS Fail_title (
title_id TEXT NOT NULL PRIMARY KEY UNIQUE
);
CREATE TABLE IF NOT EXISTS End_page (
pre_page TEXT NOT NULL PRIMARY KEY UNIQUE
);
''')
ptt_url = "https://www.ptt.cc/bbs/Gamesale/index.html"
#from_page = raw_input("輸入開始頁數（index3311.html），不需要就直接按Enter:")
#if from_page == "":
# res = requests.get(ptt_url)
#else:
# try:
# res = requests.get(ptt_url[:32] + from_page, "html.parser")
# except:
# print "輸入錯誤：", from_page
# exit()
#print res.encoding
res = requests.get(ptt_url)
soup = BeautifulSoup(res.text, "html.parser")
#print soup.prettify().encode('utf-8')
#title_list_per_page = list()
# 處理的文章條數
count = 0
# 處理的頁數
page = 0
# 問題文章數
fail_count = 0
# 發現重複處理的文章
#find_title_id = False
# 是否使用database 抓取上次寫入的頁數
is_read_database = False
# 遊戲主機白名單
gamestation_list = ["PC","XONE","X360",u"其他","WII","NDS","GBA","XBOX","PS","NGC","DC","DS"]
while True: # page loop
try:
page = page + 1
# page上的每一條文章 list row_list
row_list = soup.select("body div.title a")
row_list.reverse()
# 抓page上一頁的網址
pre_page = str(soup.find_all("a", text="‹ 上頁")[0].get('href'))
# 處理每條文章開始
for row in row_list: # title loop
# 文章的連結與其識別碼 str title_url title_id
title_id = row.get('href').encode('utf-8')
title_url = ptt_url[0:18] + title_id
title_id = title_url[32:-5]
# TODO 設定第二判斷（有爬完一次為第二次判斷，以免被覆蓋）
# 判斷文章識別碼是否已經在資料庫中
cur.execute('SELECT titleID FROM Content WHERE titleID =?',(title_id,))
r = cur.fetchone()
if r is not None :
#find_title_id = True
if not is_read_database:
cur.execute('SELECT pre_page FROM End_page')
try:
pre_page_database = cur.fetchone()[0]
pre_page = pre_page_database
is_read_database = True
except:
pass
#print title_id, "is already in database."
#print "title_url=", title_url
continue
# 擷取文章名
title_content = row.get_text()
# 徵求文或販賣文
try:
sellORwanted = re.search("\]\s*([\S+])\s*", title_content).group(1)
except:
cur.execute('''INSERT OR IGNORE INTO Fail_title (title_id)
VALUES ( ? )''', ( title_id, ) )
print title_id, "skiped.(Fail)"
fail_count = fail_count + 1
continue
# page上的遊戲主機 str gamestation
if '[' not in title_content or ']' not in title_content: continue
pre_bracket = title_content.index('[')
back_bracket = title_content.index(']')
gamestation = title_content[pre_bracket + 1: back_bracket].upper()
gamestation = gamestation.rstrip().lstrip()
# 文章的遊戲主機 是否在白名單中
in_gamestation_list = False
for station in gamestation_list:
if station in gamestation:
in_gamestation_list = True
if not in_gamestation_list:
print title_id, "skiped.(Not in white list)"
continue
# 爬進文章中 TODO:
# 網址會壞掉嗎？例如剛(被)砍的文章.ans:Yes(try and except?) 利用回應時間來解決？
res = requests.get(title_url)
soup = BeautifulSoup(res.text, "html.parser")
# 擷取文章資訊
title_head = soup.select(".article-meta-value")
try:
author = title_head[0].text
date = title_head[3].text
date = date[4:16] + date[19:]
except:
cur.execute('''INSERT OR IGNORE INTO Fail_title (title_id)
VALUES ( ? )''', ( title_id, ) )
print title_id, "skiped.(Fail)"
fail_count = fail_count + 1
continue
# 找出目標內容，並擷取
content_list = catch_content_return_list(soup.body.find_all(string=True), sellORwanted)
if content_list == []:
cur.execute('''INSERT OR IGNORE INTO Fail_title (title_id)
VALUES ( ? )''', ( title_id, ) )
print title_id, "skiped.(Fail)"
fail_count = fail_count + 1
continue
# print "---------------------------------------------------"
# print "author=", author
# print "date =", date
# print "GS =", gamestation
# print "ID =", title_id
# print "售或徵求 =", content_list[0]
# print "物品名稱 =", content_list[1]
# print "遊戲分級 =", content_list[2]
# print "語系版本 =", content_list[3]
# print "目標價錢 =", content_list[4]
# print "交易方式 =", content_list[5]
# print "保存狀況 =", content_list[6]
# print "所在地區 =", content_list[7]
# print "其他附註 =", content_list[8]
# print "---------------------------------------------------"
# database 存入
cur.execute('''INSERT OR IGNORE INTO Language (name)
VALUES ( ? )''', ( content_list[3], ) )
cur.execute('SELECT id FROM Language WHERE name = ? ', (content_list[3], ))
language_id = cur.fetchone()[0]
cur.execute('''INSERT OR IGNORE INTO Law_limited (name)
VALUES ( ? )''', ( content_list[2], ) )
cur.execute('SELECT id FROM Law_limited WHERE name = ? ', (content_list[2], ))
law_limited_id = cur.fetchone()[0]
cur.execute('''INSERT OR IGNORE INTO Gamestation (name)
VALUES ( ? )''', ( gamestation, ) )
cur.execute('SELECT id FROM Gamestation WHERE name = ? ', (gamestation, ))
gamestation_id = cur.fetchone()[0]
cur.execute('''INSERT OR IGNORE INTO SellORwanted (name)
VALUES ( ? )''', ( content_list[0], ) )
cur.execute('SELECT id FROM SellORwanted WHERE name = ? ', (content_list[0], ))
sellORwanted_id = cur.fetchone()[0]
cur.execute('''INSERT OR IGNORE INTO Author (name)
VALUES ( ? )''', ( author, ) )
cur.execute('SELECT id FROM Author WHERE name = ? ', (author, ))
author_id = cur.fetchone()[0]
cur.execute('''
INSERT OR REPLACE INTO Content
( titleID, author_id, sellORwanted_id, gamestation_id, date, item_name,
law_limited_id, language_id, price, howtotrade, status, address, message)
VALUES ( ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ? )''',
( title_id, author_id, sellORwanted_id, gamestation_id, date, content_list[1],
law_limited_id, language_id, content_list[4], content_list[5], content_list[6], content_list[7], content_list[8] ) )
conn.commit()
print title_id, "success!"
is_read_database = False
count = count + 1
# 發現重複處理的文章後處置
#if find_title_id: break
# want be saved to database
pre_page_save = pre_page
# 爬進上一頁，直到第一頁
if pre_page == "None":
cur.execute('''DELETE FROM End_page''')
cur.execute('''INSERT OR REPLACE INTO End_page(pre_page)
VALUES ( ? )''', ("None", ) )
conn.commit()
print "pre_page=", pre_page
break
pre_page = ptt_url[:18] + pre_page
res = requests.get(pre_page)
soup = BeautifulSoup(res.text, "html.parser")
print "DO", pre_page
except:
print ""
print "Program interrupted!!!"
print "NOW TitleID:", title_id
print "NOW Page url:", pre_page_save
print "NOW titles:", count
cur.execute('''DELETE FROM End_page''')
cur.execute('''INSERT OR REPLACE INTO End_page(pre_page)
VALUES ( ? )''', ( pre_page_save, ) )
conn.commit()
print "Record pre_page to database."
exit()
print "Total pages:", page
print "Total titles:", count
print "Total fail pages", fail_count
print "Done!"
爬蟲目標是PTT gamesale 板的買賣文，
只要是在白名單的文章分類大部分會爬進去 （文章標題不符會跳過）。
將爬到的資料都用字串存在sqlite資料庫裡，
有用關聯式的資料庫。
內容有過濾特殊字串，
像是★、空白、大的空白、換行、發文提醒的範例文字，
最後也有過濾"【"、"：" --->用這些判斷區間，
像是 "物品名稱" 存字串由"："到下個"【"之間，
內容分：八個區段：
- 物品名稱
- 遊戲分級
- 語系版本
- 徵求價 or 售價
- 交易方式
- 保存狀況
- 地區
- 附註
书籍推荐
