update extractor base classes

pull/13/head
Mike Fährmann 10 years ago
parent 63bf0d1725
commit 41f00809ff

@ -1,51 +1,81 @@
# -*- coding: utf-8 -*-
# Copyright 2014, 2015 Mike Fährmann
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
"""Common classes and constants used by extractor modules."""
import queue import queue
import threading import threading
import requests import requests
from ..util import safe_request # from ..util import safe_request
class BasicExtractor(): class Message():
def __init__(self, config): Version = 1
self.session = requests.Session() Directory = 2
self.category = "" Url = 3
self.directory = ""
class Extractor():
def __init__(self):
self.session = requests.Session()
def __iter__(self): def __iter__(self):
return self.images() return self.items()
def items(self):
yield Message.Version, 1
return
def request(self, url, *args, **kwargs): def request(self, url, *args, **kwargs):
return safe_request(self.session, url, *args, **kwargs) return safe_request(self.session, url, *args, **kwargs)
def enable_useragent(self): def enable_useragent(self):
self.session.headers["User-Agent"] = "Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Firefox/24.0" self.session.headers["User-Agent"] = (
"Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Firefox/24.0"
)
@staticmethod @staticmethod
def extract(txt, begin, end, pos=0): def extract(txt, begin, end, pos=0):
try: try:
first = txt.index(begin, pos) + len(begin) first = txt.index(begin, pos) + len(begin)
last = txt.index(end, first) last = txt.index(end, first)
return txt[first:last], last+len(end) return txt[first:last], last+len(end)
except: except ValueError:
return None, pos return None, pos
@staticmethod @staticmethod
def extract_all(txt, begin, end, pos=0): def extract_all(txt, begin, end, pos=0):
try: try:
first = txt.index(begin, pos) first = txt.index(begin, pos)
last = txt.index(end, first + len(begin)) + len(end) last = txt.index(end, first + len(begin)) + len(end)
return txt[first:last], last return txt[first:last], last
except: except ValueError:
return None, pos return None, pos
class AsyncExtractor(BasicExtractor):
class SequentialExtractor(Extractor):
def __init__(self, _):
Extractor.__init__(self)
class AsyncExtractor(Extractor):
def __init__(self, config): def __init__(self, config):
super().__init__(config) Extractor.__init__(self)
self.__queue = queue.Queue(maxsize=5) queue_size = int(config.get("queue-size", 5))
self.__thread = threading.Thread(target=self.async_images, daemon=True) self.__queue = queue.Queue(maxsize=queue_size)
self.__thread = threading.Thread(target=self.async_items)
# self.__thread = threading.Thread(target=self.async_images, daemon=True)
def __iter__(self): def __iter__(self):
get = self.__queue.get get = self.__queue.get
done = self.__queue.task_done done = self.__queue.task_done
self.__thread.start() self.__thread.start()
@ -56,12 +86,12 @@ class AsyncExtractor(BasicExtractor):
yield task yield task
done() done()
def async_images(self): def async_items(self):
put = self.__queue.put put = self.__queue.put
try: try:
for task in self.images(): for task in self.items():
put(task) put(task)
except: except Exception:
import traceback import traceback
print(traceback.format_exc()) print(traceback.format_exc())
put(None) put(None)

Loading…
Cancel
Save