python-ox/ox/cache.py

496 lines
15 KiB
Python
Raw Permalink Normal View History

2008-04-27 16:54:37 +00:00
# -*- coding: utf-8 -*-
2008-06-19 09:21:21 +00:00
# vi:si:et:sw=4:sts=4:ts=4
2011-11-01 12:55:49 +00:00
# GPL 2011
from __future__ import print_function
2011-11-01 12:55:49 +00:00
import gzip
2009-03-16 17:15:14 +00:00
import hashlib
2008-04-27 16:54:37 +00:00
import os
2016-06-08 13:30:25 +00:00
import sqlite3
2008-04-27 16:54:37 +00:00
import time
2016-06-08 13:30:25 +00:00
import zlib
2023-07-27 11:07:13 +00:00
from io import BytesIO
import urllib
2023-07-27 16:07:49 +00:00
import requests
2023-07-27 16:21:21 +00:00
from requests.structures import CaseInsensitiveDict
2023-07-27 16:07:49 +00:00
2008-04-27 16:54:37 +00:00
2014-09-30 19:04:46 +00:00
from .utils import json
2011-11-01 12:55:49 +00:00
from .file import makedirs
2014-09-30 19:04:46 +00:00
from . import net
from .net import DEFAULT_HEADERS, detect_encoding
2008-04-27 16:54:37 +00:00
2016-06-08 13:30:25 +00:00
cache_timeout = 30*24*60*60 # default is 30 days
2023-07-27 16:07:49 +00:00
requests_session = requests.Session()
2008-04-27 16:54:37 +00:00
2011-03-28 18:27:41 +00:00
COMPRESS_TYPES = (
'text/html',
'text/plain',
'text/xml',
'text/x-wiki',
2014-10-02 08:28:22 +00:00
'application/json',
2011-03-28 18:27:41 +00:00
'application/xhtml+xml',
'application/x-javascript',
'application/javascript',
'application/ecmascript',
'application/rss+xml'
)
2009-08-20 17:27:11 +00:00
2016-06-08 13:30:25 +00:00
def status(url, data=None, headers=None, timeout=cache_timeout):
2008-06-19 09:21:21 +00:00
'''
>>> status('http://google.com')
200
>>> status('http://google.com/mysearch')
404
'''
headers = get_headers(url, data, headers)
2008-06-19 09:21:21 +00:00
return int(headers['status'])
2016-06-08 13:30:25 +00:00
def exists(url, data=None, headers=None, timeout=cache_timeout):
2008-06-19 09:21:21 +00:00
'''
>>> exists('http://google.com')
True
>>> exists('http://google.com/mysearch')
False
'''
s = status(url, data, headers, timeout)
if s >= 200 and s < 400:
return True
return False
2016-06-08 13:30:25 +00:00
def get_headers(url, data=None, headers=None, timeout=cache_timeout):
2011-11-01 12:55:49 +00:00
url_headers = store.get(url, data, headers, timeout, "headers")
if not url_headers:
url_headers = net.get_headers(url, data, headers)
2011-11-01 12:55:49 +00:00
store.set(url, data, -1, url_headers)
2023-07-27 16:21:21 +00:00
return CaseInsensitiveDict(url_headers)
2008-04-27 16:54:37 +00:00
2016-06-08 13:30:25 +00:00
def get_json(url, data=None, headers=None, timeout=cache_timeout):
return json.loads(read_url(url, data, headers, timeout).decode('utf-8'))
class InvalidResult(Exception):
"""Base class for exceptions in this module."""
def __init__(self, result, headers):
self.result = result
self.headers = headers
2012-08-17 20:20:35 +00:00
def _fix_unicode_url(url):
2014-09-30 19:04:46 +00:00
if not isinstance(url, bytes):
2012-08-17 20:20:35 +00:00
url = url.encode('utf-8')
return url
2016-06-08 13:30:25 +00:00
def read_url(url, data=None, headers=None, timeout=cache_timeout, valid=None, unicode=False):
'''
url - url to load
data - possible post data
headers - headers to send with request
timeout - get from cache if cache not older than given seconds, -1 to get from cache
valid - function to check if result is ok, its passed result and headers
if this function fails, InvalidResult will be raised deal with it in your code
'''
2012-08-21 06:41:25 +00:00
if net.DEBUG:
2014-09-30 19:04:46 +00:00
print('ox.cache.read_url', url)
2016-06-08 13:30:25 +00:00
# FIXME: send last-modified / etag from cache and only update if needed
# url = _fix_unicode_url(url)
2012-08-17 20:20:35 +00:00
result = store.get(url, data, headers, timeout)
2014-09-30 19:04:46 +00:00
url_headers = {}
2012-08-17 20:20:35 +00:00
if not result:
2023-07-27 16:07:49 +00:00
if headers is None:
headers = DEFAULT_HEADERS.copy()
if data:
r = requests_session.post(url, data=data, headers=headers)
else:
r = requests_session.get(url, headers=headers)
for key in r.headers:
url_headers[key.lower()] = r.headers[key]
result = r.content
url_headers['Status'] = "%s" % r.status_code
if not valid or valid(result, url_headers):
store.set(url, post_data=data, data=result, headers=url_headers)
else:
2023-07-27 16:07:49 +00:00
raise InvalidResult(result, url_headers)
if unicode:
2014-09-30 19:04:46 +00:00
ctype = url_headers.get('content-type', '').lower()
if 'charset' in ctype:
encoding = ctype.split('charset=')[-1]
else:
encoding = detect_encoding(result)
if not encoding:
encoding = 'latin-1'
2012-08-17 20:20:35 +00:00
result = result.decode(encoding)
return result
2016-06-08 13:30:25 +00:00
get_url = read_url
def save_url(url, filename, overwrite=False):
2009-11-16 22:43:17 +00:00
if not os.path.exists(filename) or overwrite:
dirname = os.path.dirname(filename)
2014-12-09 12:11:09 +00:00
if dirname and not os.path.exists(dirname):
2017-08-31 13:49:19 +00:00
makedirs(dirname)
data = read_url(url)
2014-10-02 08:34:04 +00:00
with open(filename, 'wb') as f:
f.write(data)
2009-11-16 22:43:17 +00:00
2011-11-01 12:55:49 +00:00
def cache_path():
2008-06-19 09:21:21 +00:00
return os.environ.get('oxCACHE', os.path.expanduser('~/.ox/cache'))
2008-04-27 16:54:37 +00:00
2011-11-01 12:55:49 +00:00
class Cache:
def __init__(self):
pass
2016-06-08 13:30:25 +00:00
def get(self, url, data, headers=None, timeout=-1, value="data"):
2011-11-01 12:55:49 +00:00
'''
if value == 'data' return data of url if its in the cache else None
if value == 'headers' return headers for url
'''
pass
def set(self, url, post_data, data, headers):
pass
2015-12-11 19:00:05 +00:00
def get_domain(self, url):
return ".".join(urllib.parse.urlparse(url)[1].split('.')[-2:])
def get_url_hash(self, url, data=None):
if data:
url_hash = hashlib.sha1((url + '?' + data).encode('utf-8')).hexdigest()
else:
url_hash = hashlib.sha1(url.encode('utf-8')).hexdigest()
return url_hash
2011-11-01 12:55:49 +00:00
class SQLiteCache(Cache):
def __init__(self):
path = cache_path()
if not os.path.exists(path):
2017-08-31 13:49:19 +00:00
makedirs(path)
2011-11-01 12:55:49 +00:00
self.db = os.path.join(path, "cache.sqlite")
self.create()
2011-11-01 12:55:49 +00:00
def connect(self):
2014-09-30 19:04:46 +00:00
self.conn = sqlite3.connect(self.db, timeout=10)
return self.conn
2011-11-01 12:55:49 +00:00
def create(self):
conn = self.connect()
c = conn.cursor()
2011-11-01 12:55:49 +00:00
# Create table and indexes
c.execute('''CREATE TABLE IF NOT EXISTS cache (url_hash varchar(42) unique, domain text, url text,
post_data text, headers text, created int, data blob, only_headers int)''')
c.execute('''CREATE INDEX IF NOT EXISTS cache_domain ON cache (domain)''')
c.execute('''CREATE INDEX IF NOT EXISTS cache_url ON cache (url)''')
c.execute('''CREATE INDEX IF NOT EXISTS cache_url_hash ON cache (url_hash)''')
c.execute('''CREATE TABLE IF NOT EXISTS setting (key varchar(1024) unique, value text)''')
if int(self.get_setting(c, 'version', 0)) < 1:
self.set_setting(c, 'version', 1)
c.execute('''ALTER TABLE cache ADD compressed INT DEFAULT 0''')
conn.commit()
2015-12-11 19:00:05 +00:00
2011-11-01 12:55:49 +00:00
def get_setting(self, c, key, default=None):
c.execute('SELECT value FROM setting WHERE key = ?', (key, ))
for row in c:
return row[0]
return default
2009-08-20 17:27:11 +00:00
2011-11-01 12:55:49 +00:00
def set_setting(self, c, key, value):
c.execute(u'INSERT OR REPLACE INTO setting values (?, ?)', (key, str(value)))
2016-06-08 13:30:25 +00:00
def get(self, url, data={}, headers=None, timeout=-1, value="data"):
2011-11-01 12:55:49 +00:00
r = None
if timeout == 0:
return r
2015-12-11 19:00:05 +00:00
url_hash = self.get_url_hash(url, data)
2011-11-01 12:55:49 +00:00
conn = self.connect()
c = conn.cursor()
2011-11-01 12:55:49 +00:00
sql = 'SELECT %s, compressed FROM cache WHERE url_hash=?' % value
if timeout > 0:
now = time.mktime(time.localtime())
t = (url_hash, now-timeout)
sql += ' AND created > ?'
else:
t = (url_hash, )
if value != "headers":
sql += ' AND only_headers != 1 '
c.execute(sql, t)
2011-03-28 18:27:41 +00:00
for row in c:
2011-11-01 12:55:49 +00:00
r = row[0]
if value == 'headers':
r = json.loads(r)
elif value == 'data':
if row[1] == 1:
r = zlib.decompress(r)
break
c.close()
conn.close()
2011-11-01 12:55:49 +00:00
return r
2016-06-08 13:30:25 +00:00
def delete(self, url, data=None, headers=None):
2015-12-11 19:00:05 +00:00
url_hash = self.get_url_hash(url, data)
2015-12-11 18:27:54 +00:00
conn = self.connect()
c = conn.cursor()
sql = 'DELETE FROM cache WHERE url_hash=?'
t = (url_hash, )
c.execute(sql, t)
2015-12-11 19:00:05 +00:00
conn.commit()
c.close()
conn.close()
2015-12-11 18:27:54 +00:00
2011-11-01 12:55:49 +00:00
def set(self, url, post_data, data, headers):
2015-12-11 19:00:05 +00:00
url_hash = self.get_url_hash(url, post_data)
domain = self.get_domain(url)
2011-11-01 12:55:49 +00:00
conn = self.connect()
c = conn.cursor()
2011-11-01 12:55:49 +00:00
# Insert a row of data
2016-06-08 13:30:25 +00:00
if not post_data:
post_data = ""
2011-11-01 12:55:49 +00:00
only_headers = 0
if data == -1:
only_headers = 1
data = ""
created = time.mktime(time.localtime())
2011-03-28 18:27:41 +00:00
content_type = headers.get('content-type', '').split(';')[0].strip()
if content_type in COMPRESS_TYPES:
2011-11-01 12:55:49 +00:00
compressed = 1
2011-03-28 18:27:41 +00:00
data = zlib.compress(data)
2011-11-01 12:55:49 +00:00
else:
compressed = 0
2023-07-27 16:21:21 +00:00
if isinstance(data, str):
data = data.encode("utf-8")
2011-11-01 12:55:49 +00:00
data = sqlite3.Binary(data)
2014-04-23 13:38:38 +00:00
#fixme: this looks wrong
try:
_headers = json.dumps(headers)
except:
for h in headers:
headers[h] = headers[h].decode(detect_encoding(headers[h]))
_headers = json.dumps(headers)
t = (url_hash, domain, url, post_data, _headers, created,
2011-11-01 12:55:49 +00:00
data, only_headers, compressed)
c.execute(u"""INSERT OR REPLACE INTO cache values (?, ?, ?, ?, ?, ?, ?, ?, ?)""", t)
# Save (commit) the changes and clean up
conn.commit()
2011-11-01 12:55:49 +00:00
c.close()
conn.close()
2011-11-01 12:55:49 +00:00
class FileCache(Cache):
def __init__(self):
f, self.root = cache_path().split(':')
def files(self, domain, h):
prefix = os.path.join(self.root, domain, h[:2], h[2:4], h[4:6], h[6:8])
2017-03-14 17:24:11 +00:00
i = os.path.join(prefix, '%s.json' % h)
f = os.path.join(prefix, '%s.dat' % h)
2011-11-01 12:55:49 +00:00
return prefix, i, f
2015-12-11 19:00:05 +00:00
2016-06-08 13:30:25 +00:00
def get(self, url, data={}, headers=None, timeout=-1, value="data"):
2011-11-01 12:55:49 +00:00
r = None
if timeout == 0:
return r
2015-12-11 19:00:05 +00:00
url_hash = self.get_url_hash(url, data)
domain = self.get_domain(url)
2011-11-01 12:55:49 +00:00
prefix, i, f = self.files(domain, url_hash)
if os.path.exists(i):
with open(i) as _i:
try:
info = json.load(_i)
except:
return r
2011-11-01 12:55:49 +00:00
now = time.mktime(time.localtime())
expired = now-timeout
if value != 'headers' and info['only_headers']:
return None
if timeout < 0 or info['created'] > expired:
if value == 'headers':
r = info['headers']
else:
2016-06-08 13:30:25 +00:00
with open(f, 'rb') as fd:
r = fd.read()
2011-11-01 12:55:49 +00:00
if info['compressed']:
r = zlib.decompress(r)
return r
2016-06-08 13:30:25 +00:00
def delete(self, url, data=None, headers=None):
2015-12-11 19:00:05 +00:00
url_hash = self.get_url_hash(url, data)
domain = self.get_domain(url)
2015-12-11 18:27:54 +00:00
prefix, i, f = self.files(domain, url_hash)
if os.path.exists(i):
os.unlink(i)
2011-11-01 12:55:49 +00:00
def set(self, url, post_data, data, headers):
2015-12-11 19:00:05 +00:00
url_hash = self.get_url_hash(url, post_data)
domain = self.get_domain(url)
2011-11-01 12:55:49 +00:00
prefix, i, f = self.files(domain, url_hash)
makedirs(prefix)
created = time.mktime(time.localtime())
content_type = headers.get('content-type', '').split(';')[0].strip()
info = {
'compressed': content_type in COMPRESS_TYPES,
'only_headers': data == -1,
'created': created,
'headers': headers,
2011-11-01 21:53:36 +00:00
'url': url,
2011-11-01 12:55:49 +00:00
}
if post_data:
info['post_data'] = post_data
if not info['only_headers']:
if info['compressed']:
data = zlib.compress(data)
2016-09-07 13:47:12 +00:00
elif not isinstance(data, bytes):
2014-10-02 08:34:04 +00:00
data = data.encode('utf-8')
with open(f, 'wb') as _f:
2011-11-01 12:55:49 +00:00
_f.write(data)
2016-09-07 13:47:12 +00:00
with open(i, 'w') as _i:
2011-11-01 12:55:49 +00:00
json.dump(info, _i)
2017-03-14 17:24:11 +00:00
class KVCache(Cache):
_bytes_only = False
def _keys(self, url, data, headers=None):
url_hash = self.get_url_hash(url, data)
domain = self.get_domain(url)
key = 'ox:%s:%s' % (domain, url_hash)
return key, key + ':data'
def get(self, url, data={}, headers=None, timeout=-1, value="data"):
if timeout == 0:
return None
r = None
info_key, data_key = self._keys(url, data, headers)
info = self.backend.get(info_key)
if info:
if self._bytes_only:
info = json.loads(info.decode())
now = time.mktime(time.localtime())
expired = now-timeout
if value != 'headers' and info['only_headers']:
return None
if timeout < 0 or info['created'] > expired:
if value == 'headers':
r = info['headers']
else:
r = self.backend.get(data_key)
if r and info['compressed']:
r = zlib.decompress(r)
2017-03-14 17:24:11 +00:00
return r
def delete(self, url, data=None, headers=None):
for key in self._keys(url, data, headers):
self.backend.delete(key)
def set(self, url, post_data, data, headers):
info_key, data_key = self._keys(url, post_data, headers)
created = time.mktime(time.localtime())
content_type = headers.get('content-type', '').split(';')[0].strip()
info = {
'compressed': content_type in COMPRESS_TYPES,
'only_headers': data == -1,
'created': created,
'headers': headers,
'url': url,
}
if post_data:
info['post_data'] = post_data
if not info['only_headers']:
if info['compressed']:
data = zlib.compress(data)
elif not isinstance(data, bytes):
data = data.encode('utf-8')
self.backend.set(data_key, data)
if self._bytes_only:
info = json.dumps(info, ensure_ascii=False).encode('utf-8')
self.backend.set(info_key, info)
class MemCache(KVCache):
_bytes_only = False
def __init__(self):
import pylibmc
f, self.host = cache_path().split(':', 1)
self.backend = pylibmc.Client([self.host])
2017-03-23 09:48:42 +00:00
self.backend.behaviors['connect_timeout'] = 60000
2017-03-14 17:24:11 +00:00
class RedisCache(KVCache):
_bytes_only = True
def __init__(self):
import redis
f, self.url = cache_path().split(':', 1)
self.backend = redis.from_url(self.url)
2019-04-29 10:22:27 +00:00
class FallbackCache(KVCache):
caches = []
def __init__(self):
fallback = cache_path()
for path in fallback.split('|'):
os.environ['oxCACHE'] = path
if path.startswith('redis:'):
store = RedisCache()
elif path.startswith('memcache:'):
store = MemCache()
self.caches.append(store)
os.environ['oxCACHE'] = fallback
def get(self, url, data, headers=None, timeout=-1, value="data"):
if timeout == 0:
return None
info_key, data_key = self._keys(url, data, headers)
for cache in self.caches:
try:
info = cache.backend.get(info_key)
except:
info = None
if info:
return cache.get(url, data, headers, timeout, value)
return None
def set(self, url, post_data, data, headers):
self.caches[0].set(url, post_data, data, headers)
for cache in self.caches[1:]:
cache.delete(url, post_data, headers)
def delete(self, url, data=None, headers=None):
for cache in self.caches:
cache.delete(url, data, headers)
if '|' in cache_path():
store = FallbackCache()
elif cache_path().startswith('fs:'):
2011-11-01 12:55:49 +00:00
store = FileCache()
2017-03-14 17:24:11 +00:00
elif cache_path().startswith('redis:'):
store = RedisCache()
elif cache_path().startswith('memcache:'):
store = MemCache()
2011-11-01 12:55:49 +00:00
else:
store = SQLiteCache()