scientific-programming-exer.../exam/ex01/receive.py

109 lines
3.2 KiB
Python

import logging
from collections import deque
from cfg import config
from url import construct_url
from proxy import get_data_with_proxy, NoMoreProxiesException
from db_util import get_page_id, get_page_title
logger = logging.getLogger(__name__)
def ignore_title(title):
ignore_starters = ["Help:"
, "Wikipedia:"
, "Template:"
, "Template_talk:"
, "Category:"
]
for ignore in ignore_starters:
if(title.startswith(ignore)):
return True
return False
def _receive_links(page, connection):
title = get_page_title(page, connection)
url = construct_url(title)
result = get_data_with_proxy(url, connection)
# This is basically because we don't know the page ID.
for k, page_data in result["query"]["pages"].items():
cursor = connection.cursor()
if(not "links" in page_data):
destination_title = page_data["title"].replace(" ", "_")
# avoid 1-loops
if(destination_title == title):
continue
if(ignore_title(destination_title)):
continue
destination = get_page_id(destination_title, connection)
cursor.execute("INSERT INTO links(source, destination) VALUES(%s, %s)", (page, destination))
yield destination
else:
for destination in page_data["links"]:
destination_title = destination["title"].replace(" ", "_")
if(ignore_title(destination_title)):
continue
destination = get_page_id(destination_title, connection)
cursor.execute("INSERT INTO links(source, destination) VALUES(%s, %s)", (page, destination))
yield destination
connection.commit()
def receive_links(title, connection):
return list(_receive_links(title, connection))
def receive_link_graph(title, connection, depth):
page = get_page_id(title, connection)
do_receive_link_graph(page, connection, depth, fetch_missing=True)
cursor = connection.cursor()
cursor.execute("SELECT COUNT(page) FROM failed_to_fetch")
if(cursor.fetchone()[0]):
do_receive_link_graph(page, connection, depth, fetch_missing=True)
def do_receive_link_graph(page, connection, depth, fetch_missing=False):
if(depth < 0):
# end of recursion
return
logger.info("do_receive_link_graph(%d, <connection>, %d)" % (page, depth))
cursor = connection.cursor()
# Fetch the missing links.
if(fetch_missing):
delete_cursor = connection.cursor()
cursor.execute('''SELECT failed_to_fetch.depth, failed_to_fetch.page
FROM failed_to_fetch
''')
for d, p in cursor:
do_receive_link_graph(p, connection, d, fetch_missing=False)
delete_cursor.execute("DELETE FROM failed_to_fetch WHERE page=%s", (p,))
cursor = connection.cursor()
cursor.execute("SELECT COUNT(source) FROM links WHERE source=%s", (page,))
if(cursor.fetchone()[0] != 0):
# we fetched that title already
return
logger.info("fetching links for {}".format(page))
for link in _receive_links(page, connection):
try:
do_receive_link_graph(link, connection, depth - 1)
except NoMoreProxiesException as e:
logger.exception("All proxies are blocked")
# Wikipedia blocked all our proxies.
# Retry later, so we have to store our list that is still to fetch.
cursor = connection.cursor()
cursor.execute("INSERT INTO failed_to_fetch(page, depth) VALUES(%s, %s)", (link, depth - 1))
connection.commit()