add file cache

This commit is contained in:
j 2011-11-01 13:55:49 +01:00
parent 2a7b70c576
commit 8fe8822e09

View file

@ -1,6 +1,8 @@
# -*- coding: utf-8 -*- # -*- coding: utf-8 -*-
# vi:si:et:sw=4:sts=4:ts=4 # vi:si:et:sw=4:sts=4:ts=4
# GPL 2008 # GPL 2011
from __future__ import with_statement
import gzip import gzip
import zlib import zlib
import hashlib import hashlib
@ -12,12 +14,12 @@ import urllib2
import sqlite3 import sqlite3
import chardet import chardet
from ox.utils import json from utils import json
from .file import makedirs
import net import net
from net import DEFAULT_HEADERS, getEncoding from net import DEFAULT_HEADERS, getEncoding
cache_timeout = 30*24*60*60 # default is 30 days cache_timeout = 30*24*60*60 # default is 30 days
COMPRESS_TYPES = ( COMPRESS_TYPES = (
@ -54,12 +56,10 @@ def exists(url, data=None, headers=DEFAULT_HEADERS, timeout=cache_timeout):
return False return False
def getHeaders(url, data=None, headers=DEFAULT_HEADERS, timeout=cache_timeout): def getHeaders(url, data=None, headers=DEFAULT_HEADERS, timeout=cache_timeout):
url_headers = _readUrlCache(url, data, headers, timeout, "headers") url_headers = store.get(url, data, headers, timeout, "headers")
if url_headers: if not url_headers:
url_headers = json.loads(url_headers)
else:
url_headers = net.getHeaders(url, data, headers) url_headers = net.getHeaders(url, data, headers)
_saveUrlCache(url, data, -1, url_headers) store.set(url, data, -1, url_headers)
return url_headers return url_headers
class InvalidResult(Exception): class InvalidResult(Exception):
@ -80,7 +80,7 @@ def readUrl(url, data=None, headers=DEFAULT_HEADERS, timeout=cache_timeout, vali
#FIXME: send last-modified / etag from cache and only update if needed #FIXME: send last-modified / etag from cache and only update if needed
if isinstance(url, unicode): if isinstance(url, unicode):
url = url.encode('utf-8') url = url.encode('utf-8')
result = _readUrlCache(url, data, headers, timeout) result = store.get(url, data, headers, timeout)
if not result: if not result:
#print "get data", url #print "get data", url
try: try:
@ -92,7 +92,7 @@ def readUrl(url, data=None, headers=DEFAULT_HEADERS, timeout=cache_timeout, vali
if url_headers.get('content-encoding', None) == 'gzip': if url_headers.get('content-encoding', None) == 'gzip':
result = gzip.GzipFile(fileobj=StringIO.StringIO(result)).read() result = gzip.GzipFile(fileobj=StringIO.StringIO(result)).read()
if not valid or valid(result, url_headers): if not valid or valid(result, url_headers):
_saveUrlCache(url, data, result, url_headers) store.set(url, data, result, url_headers)
else: else:
raise InvalidResult(result, url_headers) raise InvalidResult(result, url_headers)
return result return result
@ -114,172 +114,203 @@ def saveUrl(url, filename, overwrite=False):
f.write(data) f.write(data)
f.close() f.close()
def _getCacheBase(): def cache_path():
'cache base is eather ~/.ox/cache or can set via env variable oxCACHE'
return os.environ.get('oxCACHE', os.path.expanduser('~/.ox/cache')) return os.environ.get('oxCACHE', os.path.expanduser('~/.ox/cache'))
def _getCacheDB(): class Cache:
path = _getCacheBase() def __init__(self):
if not os.path.exists(path): pass
os.makedirs(path)
return os.path.join(path, "cache.sqlite")
def _connectDb(): def get(self, url, data, headers=DEFAULT_HEADERS, timeout=-1, value="data"):
conn = sqlite3.connect(_getCacheDB(), timeout=10) '''
conn.text_factory = str if value == 'data' return data of url if its in the cache else None
return conn if value == 'headers' return headers for url
'''
pass
def _getSetting(c, key, default=None): def set(self, url, post_data, data, headers):
c.execute('SELECT value FROM setting WHERE key = ?', (key, )) pass
for row in c:
return row[0]
return default
def _setSetting(c, key, value): class SQLiteCache(Cache):
c.execute(u'INSERT OR REPLACE INTO setting values (?, ?)', (key, str(value))) def __init__(self):
path = cache_path()
if not os.path.exists(path):
os.makedirs(path)
self.db = os.path.join(path, "cache.sqlite")
def _createDb(conn, c): def connect(self):
# Create table and indexes self.conn = sqlite3.connect(self.db, timeout=10)
c.execute('''CREATE TABLE IF NOT EXISTS cache (url_hash varchar(42) unique, domain text, url text, self.conn.text_factory = str
post_data text, headers text, created int, data blob, only_headers int)''') self.create()
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)''') def create(self):
if int(_getSetting(c, 'version', 0)) < 1: c = self.conn.cursor()
_setSetting(c, 'version', 1) # Create table and indexes
c.execute('''ALTER TABLE cache ADD compressed INT DEFAULT 0''') c.execute('''CREATE TABLE IF NOT EXISTS cache (url_hash varchar(42) unique, domain text, url text,
conn.commit() 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)''')
def _readUrlCache(url, data, headers=DEFAULT_HEADERS, timeout=-1, value="data"): c.execute('''CREATE TABLE IF NOT EXISTS setting (key varchar(1024) unique, value text)''')
r = None if int(self.get_setting(c, 'version', 0)) < 1:
if timeout == 0: self.set_setting(c, 'version', 1)
c.execute('''ALTER TABLE cache ADD compressed INT DEFAULT 0''')
self.conn.commit()
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
def set_setting(self, c, key, value):
c.execute(u'INSERT OR REPLACE INTO setting values (?, ?)', (key, str(value)))
def get(self, url, data={}, headers=DEFAULT_HEADERS, timeout=-1, value="data"):
r = None
if timeout == 0:
return r
if data:
url_hash = hashlib.sha1(url + '?' + data).hexdigest()
else:
url_hash = hashlib.sha1(url).hexdigest()
self.connect()
c = self.conn.cursor()
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)
for row in c:
r = row[0]
if value == 'headers':
r = json.loads(r)
elif value == 'data':
if row[1] == 1:
r = zlib.decompress(r)
else:
r = str(r)
break
c.close()
self.conn.close()
return r return r
if data: def set(self, url, post_data, data, headers):
url_hash = hashlib.sha1(url + '?' + data).hexdigest() if post_data:
else: url_hash = hashlib.sha1(url + '?' + post_data).hexdigest()
url_hash = hashlib.sha1(url).hexdigest() else:
url_hash = hashlib.sha1(url).hexdigest()
conn = _connectDb() domain = ".".join(urlparse.urlparse(url)[1].split('.')[-2:])
c = conn.cursor()
_createDb(conn, c)
sql = 'SELECT %s, compressed FROM cache WHERE url_hash=?' % value self.connect()
if timeout > 0: c = self.conn.cursor()
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)
for row in c:
r = row[0]
if value == 'data':
if row[1] == 1:
r = zlib.decompress(r)
else:
r = str(r)
break
c.close() # Insert a row of data
conn.close() if not post_data: post_data=""
return r only_headers = 0
if data == -1:
def _saveUrlCache(url, post_data, data, headers): only_headers = 1
if post_data: data = ""
url_hash = hashlib.sha1(url + '?' + post_data).hexdigest() created = time.mktime(time.localtime())
else:
url_hash = hashlib.sha1(url).hexdigest()
domain = ".".join(urlparse.urlparse(url)[1].split('.')[-2:])
conn = _connectDb()
c = conn.cursor()
# Create table if not exists
_createDb(conn, c)
# Insert a row of data
if not post_data: post_data=""
only_headers = 0
if data == -1:
only_headers = 1
data = ""
created = time.mktime(time.localtime())
content_type = headers.get('content-type', '').split(';')[0].strip()
if content_type in COMPRESS_TYPES:
compressed = 1
data = zlib.compress(data)
else:
compressed = 0
data = sqlite3.Binary(data)
t = (url_hash, domain, url, post_data, json.dumps(headers), created, data, only_headers, compressed)
c.execute(u"""INSERT OR REPLACE INTO cache values (?, ?, ?, ?, ?, ?, ?, ?, ?)""", t)
# Save (commit) the changes and clean up
conn.commit()
c.close()
conn.close()
def migrate_to_db():
import re
import os
import sqlite3
import glob
conn = _connectDb()
c = conn.cursor()
_createDb(conn, c)
files = glob.glob(_getCacheBase() + "/*/*/*/*/*")
_files = filter(lambda x: not x.endswith(".headers"), files)
for f in _files:
info = re.compile("%s/(.*?)/../../../(.*)" % _getCacheBase()).findall(f)
domain = url = info[0][0]
url_hash = info[0][1]
post_data = ""
created = os.stat(f).st_ctime
fd = open(f, "r")
data = fd.read()
fd.close()
fd = open(f + ".headers", "r")
headers = fd.read()
fd.close()
t = (url_hash, domain, url, post_data, headers, created, sqlite3.Binary(data), 0)
c.execute(u"""INSERT OR REPLACE INTO cache values (?, ?, ?, ?, ?, ?, ?, ?)""", t)
conn.commit()
c.close()
conn.close()
def compress_db():
conn = _connectDb()
c = conn.cursor()
_createDb(conn, c)
c.execute(u"""SELECT url_hash FROM cache WHERE compressed = 0""")
ids = [row[0] for row in c]
for url_hash in ids:
c.execute(u"""SELECT headers, data FROM cache WHERE url_hash = ?""", (url_hash, ))
headers = {}
for row in c:
headers = json.loads(row[0])
data = row[1]
content_type = headers.get('content-type', '').split(';')[0].strip() content_type = headers.get('content-type', '').split(';')[0].strip()
if content_type in COMPRESS_TYPES: if content_type in COMPRESS_TYPES:
compressed = 1
data = zlib.compress(data) data = zlib.compress(data)
t = (sqlite3.Binary(data), url_hash) else:
print url_hash, 'update' compressed = 0
c.execute('UPDATE cache SET compressed = 1, data = ? WHERE url_hash = ?', t) data = sqlite3.Binary(data)
t = (url_hash, domain, url, post_data, json.dumps(headers), created,
data, only_headers, compressed)
c.execute(u"""INSERT OR REPLACE INTO cache values (?, ?, ?, ?, ?, ?, ?, ?, ?)""", t)
# Save (commit) the changes and clean up
self.conn.commit()
c.close()
self.conn.close()
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])
i = os.path.join(prefix, '%s.json'%h)
f = os.path.join(prefix, '%s.dat'%h)
return prefix, i, f
def get(self, url, data={}, headers=DEFAULT_HEADERS, timeout=-1, value="data"):
r = None
if timeout == 0:
return r
if data:
url_hash = hashlib.sha1(url + '?' + data).hexdigest()
else:
url_hash = hashlib.sha1(url).hexdigest()
domain = ".".join(urlparse.urlparse(url)[1].split('.')[-2:])
prefix, i, f = self.files(domain, url_hash)
if os.path.exists(i):
with open(i) as _i:
info = json.load(_i)
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:
with open(f) as data:
r = data.read()
if info['compressed']:
r = zlib.decompress(r)
return r
def set(self, url, post_data, data, headers):
if post_data:
url_hash = hashlib.sha1(url + '?' + post_data).hexdigest()
else:
url_hash = hashlib.sha1(url).hexdigest()
domain = ".".join(urlparse.urlparse(url)[1].split('.')[-2:])
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,
}
if post_data:
info['post_data'] = post_data
if not info['only_headers']:
if info['compressed']:
data = zlib.compress(data)
with open(f, 'w') as _f:
_f.write(data)
with open(i, 'w') as _i:
json.dump(info, _i)
if cache_path().startswith('fs:'):
store = FileCache()
else:
store = SQLiteCache()
conn.commit()
print "optimizing database"
c.execute('VACUUM')
conn.commit()
c.close()
conn.close()