Python3 cookiejar模块详解
Posted 肥宝Fable
tags:
篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了Python3 cookiejar模块详解相关的知识，希望对你有一定的参考价值。
本来想把整个模块的注释翻译一下，再加上自己的理解的。不过后来看着看着，实在没有达到那个水平。就只能理解多少写多少了。
对了，我用到时Python3.4。对于cookie不了解的同学，应该看看这个点击打开链接
首先来看看结构图：
CookieJar____
/ \\ \\
FileCookieJar \\ \\
/ | \\ \\ \\
MozillaCookieJar | LWPCookieJar \\ \\
| | \\
| ---MSIEBase | \\
| / | | \\
| / MSIEDBCookieJar BSDDBCookieJar
|/
MSIECookieJar剩下的MSIEBase，MSIEDBCookieJar， BSDDBCookieJar，MSIECookieJar都不是Python标准库。代码里面也没有，就不用管了。
整个模块内容很多。看着看着才发现应该从后面末尾倒着往上面看。越后面的内容就越接近使用者，越往前面就越是标准上面的东西。
所以从后面开始简单介绍一下。
MozillaCookieJar和LWPCookieJar类都是FileCookieJar的子类。实现了具体的把cookie内容保存为文件的方法。只是这两个类对应的标准不同而已。
FileCookieJar好明显，是为了统筹文件cookie而设计的。定义了一些文件方面的标准操作。
真正抽象的是CookieJar，看到这里，其实已经可能没必要再详细了解CookieJar了。边学边用就行了。
CookieJar是对于Cookie类的一个类似管理类的封装。
Cookie是具体的内容，更接近于数据结构，另外还有一些非常基础的函数。
然后就是策略类，CookiePolicy，定义了一些标准的策略。
DefaultCookiePolicy才是完全实现了标准的策略类。不过这些策略我也不是很懂。这里已经非常抽象了。
等以后用到在深入理解。
下面是代码的注释。要从后面看起才舒服。
r"""HTTP cookie handling for web clients.
This module has (now fairly distant) origins in Gisle Aas' Perl module
HTTP::Cookies, from the libwww-perl library.
Docstrings, comments and debug strings in this code refer to the
attributes of the HTTP cookie system as cookie-attributes, to distinguish
them clearly from Python attributes.
Class diagram (note that BSDDBCookieJar and the MSIE* classes are not
distributed with the Python standard library, but are available from
http://wwwsearch.sf.net/):
CookieJar____
/ \\ \\
FileCookieJar \\ \\
/ | \\ \\ \\
MozillaCookieJar | LWPCookieJar \\ \\
| | \\
| ---MSIEBase | \\
| / | | \\
| / MSIEDBCookieJar BSDDBCookieJar
|/
MSIECookieJar
"""
#导出的类、函数等
__all__ = ['Cookie', 'CookieJar', 'CookiePolicy', 'DefaultCookiePolicy',
'FileCookieJar', 'LWPCookieJar', 'LoadError', 'MozillaCookieJar']
import copy
import datetime
import re
import time
import urllib.parse
try:
import threading as _threading #应该是多个系统兼容处理
except ImportError:
import dummy_threading as _threading
import http.client # only for the default HTTP port
from calendar import timegm #日历
#debug模式会有一些特殊的信息打印出来
debug = False # set to True to enable debugging via the logging module
logger = None
#logging日志模块，可以设置不同的级别，不同的输出格式，输出目标等
def _debug(*args):
if not debug:
return
global logger
if not logger:
import logging
logger = logging.getLogger("http.cookiejar")
return logger.debug(*args)
#http端口
DEFAULT_HTTP_PORT = str(http.client.HTTP_PORT)
MISSING_FILENAME_TEXT = ("a filename was not supplied (nor was the CookieJar "
"instance initialised with one)")
#捕捉异常和发出警告
def _warn_unhandled_exception():
# There are a few catch-all except: statements in this module, for
# catching input that's bad in unexpected ways. Warn if any
# exceptions are caught there.
import io, warnings, traceback
f = io.StringIO()
traceback.print_exc(None, f)
msg = f.getvalue()
warnings.warn("http.cookiejar bug!\\n%s" % msg, stacklevel=2)
# Date/time conversion 日期转换
# -----------------------------------------------------------------------------
EPOCH_YEAR = 1970
def _timegm(tt):
year, month, mday, hour, min, sec = tt[:6]
if ((year >= EPOCH_YEAR) and (1 <= month <= 12) and (1 <= mday <= 31) and
(0 <= hour <= 24) and (0 <= min <= 59) and (0 <= sec <= 61)):
return timegm(tt)
else:
return None
DAYS = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
MONTHS = ["Jan", "Feb", "Mar", "Apr", "May", "Jun",
"Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]
MONTHS_LOWER = []
for month in MONTHS: MONTHS_LOWER.append(month.lower())
#按固定格式返回一个时间的字符串
def time2isoz(t=None):
"""Return a string representing time in seconds since epoch, t.
If the function is called without an argument, it will use the current
time.
The format of the returned string is like "YYYY-MM-DD hh:mm:ssZ",
representing Universal Time (UTC, aka GMT). An example of this format is:
1994-11-24 08:49:37Z
"""
if t is None:
dt = datetime.datetime.utcnow()
else:
dt = datetime.datetime.utcfromtimestamp(t)
return "%04d-%02d-%02d %02d:%02d:%02dZ" % (
dt.year, dt.month, dt.day, dt.hour, dt.minute, dt.second)
#按固定格式返回一个时间的字符串
def time2netscape(t=None):
"""Return a string representing time in seconds since epoch, t.
If the function is called without an argument, it will use the current
time.
The format of the returned string is like this:
Wed, DD-Mon-YYYY HH:MM:SS GMT
"""
if t is None:
dt = datetime.datetime.utcnow()
else:
dt = datetime.datetime.utcfromtimestamp(t)
return "%s %02d-%s-%04d %02d:%02d:%02d GMT" % (
DAYS[dt.weekday()], dt.day, MONTHS[dt.month-1],
dt.year, dt.hour, dt.minute, dt.second)
UTC_ZONES = "GMT": None, "UTC": None, "UT": None, "Z": None
TIMEZONE_RE = re.compile(r"^([-+])?(\\d\\d?):?(\\d\\d)?$", re.ASCII)
def offset_from_tz_string(tz):
offset = None
if tz in UTC_ZONES:
offset = 0
else:
m = TIMEZONE_RE.search(tz)
if m:
offset = 3600 * int(m.group(2))
if m.group(3):
offset = offset + 60 * int(m.group(3))
if m.group(1) == '-':
offset = -offset
return offset
#把字符串编程时间
def _str2time(day, mon, yr, hr, min, sec, tz):
# translate month name to number
# month numbers start with 1 (January)
try:
mon = MONTHS_LOWER.index(mon.lower())+1
except ValueError:
# maybe it's already a number
try:
imon = int(mon)
except ValueError:
return None
if 1 <= imon <= 12:
mon = imon
else:
return None
# make sure clock elements are defined
if hr is None: hr = 0
if min is None: min = 0
if sec is None: sec = 0
yr = int(yr)
day = int(day)
hr = int(hr)
min = int(min)
sec = int(sec)
if yr < 1000:
# find "obvious" year
cur_yr = time.localtime(time.time())[0]
m = cur_yr % 100
tmp = yr
yr = yr + cur_yr - m
m = m - tmp
if abs(m) > 50:
if m > 0: yr = yr + 100
else: yr = yr - 100
# convert UTC time tuple to seconds since epoch (not timezone-adjusted)
t = _timegm((yr, mon, day, hr, min, sec, tz))
if t is not None:
# adjust time using timezone string, to get absolute time since epoch
if tz is None:
tz = "UTC"
tz = tz.upper()
offset = offset_from_tz_string(tz)
if offset is None:
return None
t = t - offset
return t
STRICT_DATE_RE = re.compile(
r"^[SMTWF][a-z][a-z], (\\d\\d) ([JFMASOND][a-z][a-z]) "
"(\\d\\d\\d\\d) (\\d\\d):(\\d\\d):(\\d\\d) GMT$", re.ASCII)
WEEKDAY_RE = re.compile(
r"^(?:Sun|Mon|Tue|Wed|Thu|Fri|Sat)[a-z]*,?\\s*", re.I | re.ASCII)
LOOSE_HTTP_DATE_RE = re.compile(
r"""^
(\\d\\d?) # day
(?:\\s+|[-\\/])
(\\w+) # month
(?:\\s+|[-\\/])
(\\d+) # year
(?:
(?:\\s+|:) # separator before clock
(\\d\\d?):(\\d\\d) # hour:min
(?::(\\d\\d))? # optional seconds
)? # optional clock
\\s*
([-+]?\\d2,4|(?![APap][Mm]\\b)[A-Za-z]+)? # timezone
\\s*
(?:\\(\\w+\\))? # ASCII representation of timezone in parens.
\\s*$""", re.X | re.ASCII)
def http2time(text):
"""Returns time in seconds since epoch of time represented by a string.
Return value is an integer.
None is returned if the format of str is unrecognized, the time is outside
the representable range, or the timezone string is not recognized. If the
string contains no timezone, UTC is assumed.
The timezone in the string may be numerical (like "-0800" or "+0100") or a
string timezone (like "UTC", "GMT", "BST" or "EST"). Currently, only the
timezone strings equivalent to UTC (zero offset) are known to the function.
The function loosely parses the following formats:
Wed, 09 Feb 1994 22:23:32 GMT -- HTTP format
Tuesday, 08-Feb-94 14:15:29 GMT -- old rfc850 HTTP format
Tuesday, 08-Feb-1994 14:15:29 GMT -- broken rfc850 HTTP format
09 Feb 1994 22:23:32 GMT -- HTTP format (no weekday)
08-Feb-94 14:15:29 GMT -- rfc850 format (no weekday)
08-Feb-1994 14:15:29 GMT -- broken rfc850 format (no weekday)
The parser ignores leading and trailing whitespace. The time may be
absent.
If the year is given with only 2 digits, the function will select the
century that makes the year closest to the current date.
"""
# fast exit for strictly conforming string
m = STRICT_DATE_RE.search(text)
if m:
g = m.groups()
mon = MONTHS_LOWER.index(g[1].lower()) + 1
tt = (int(g[2]), mon, int(g[0]),
int(g[3]), int(g[4]), float(g[5]))
return _timegm(tt)
# No, we need some messy parsing...
# clean up
text = text.lstrip()
text = WEEKDAY_RE.sub("", text, 1) # Useless weekday
# tz is time zone specifier string
day, mon, yr, hr, min, sec, tz = [None]*7
# loose regexp parse
m = LOOSE_HTTP_DATE_RE.search(text)
if m is not None:
day, mon, yr, hr, min, sec, tz = m.groups()
else:
return None # bad format
return _str2time(day, mon, yr, hr, min, sec, tz)
ISO_DATE_RE = re.compile(
"""^
(\\d4) # year
[-\\/]?
(\\d\\d?) # numerical month
[-\\/]?
(\\d\\d?) # day
(?:
(?:\\s+|[-:Tt]) # separator before clock
(\\d\\d?):?(\\d\\d) # hour:min
(?::?(\\d\\d(?:\\.\\d*)?))? # optional seconds (and fractional)
)? # optional clock
\\s*
([-+]?\\d\\d?:?(:?\\d\\d)?
|Z|z)? # timezone (Z is "zero meridian", i.e. GMT)
\\s*$""", re.X | re. ASCII)
def iso2time(text):
"""
As for http2time, but parses the ISO 8601 formats:
1994-02-03 14:15:29 -0100 -- ISO 8601 format
1994-02-03 14:15:29 -- zone is optional
1994-02-03 -- only date
1994-02-03T14:15:29 -- Use T as separator
19940203T141529Z -- ISO 8601 compact format
19940203 -- only date
"""
# clean up
text = text.lstrip()
# tz is time zone specifier string
day, mon, yr, hr, min, sec, tz = [None]*7
# loose regexp parse
m = ISO_DATE_RE.search(text)
if m is not None:
# XXX there's an extra bit of the timezone I'm ignoring here: is
# this the right thing to do?
yr, mon, day, hr, min, sec, tz, _ = m.groups()
else:
return None # bad format
return _str2time(day, mon, yr, hr, min, sec, tz)
# Header parsing
# -----------------------------------------------------------------------------
#返回匹配对象不匹配的部分
def unmatched(match):
"""Return unmatched part of re.Match object."""
start, end = match.span(0)
return match.string[:start]+match.string[end:]
HEADER_TOKEN_RE = re.compile(r"^\\s*([^=\\s;,]+)")
HEADER_QUOTED_VALUE_RE = re.compile(r"^\\s*=\\s*\\"([^\\"\\\\]*(?:\\\\.[^\\"\\\\]*)*)\\"")
HEADER_VALUE_RE = re.compile(r"^\\s*=\\s*([^\\s;,]*)")
HEADER_ESCAPE_RE = re.compile(r"\\\\(.)")
def split_header_words(header_values):
r"""Parse header values into a list of lists containing key,value pairs.
The function knows how to deal with ",", ";" and "=" as well as quoted
values after "=". A list of space separated tokens are parsed as if they
were separated by ";".
If the header_values passed as argument contains multiple values, then they
are treated as if they were a single value separated by comma ",".
This means that this function is useful for parsing header fields that
follow this syntax (BNF as from the HTTP/1.1 specification, but we relax
the requirement for tokens).
headers = #header
header = (token | parameter) *( [";"] (token | parameter))
token = 1*<any CHAR except CTLs or separators>
separators = "(" | ")" | "<" | ">" | "@"
| "," | ";" | ":" | "\\" | <">
| "/" | "[" | "]" | "?" | "="
| "" | "" | SP | HT
quoted-string = ( <"> *(qdtext | quoted-pair ) <"> )
qdtext = <any TEXT except <">>
quoted-pair = "\\" CHAR
parameter = attribute "=" value
attribute = token
value = token | quoted-string
Each header is represented by a list of key/value pairs. The value for a
simple token (not part of a parameter) is None. Syntactically incorrect
headers will not necessarily be parsed as you would want.
This is easier to describe with some examples:
>>> split_header_words(['foo="bar"; port="80,81"; discard, bar=baz'])
[[('foo', 'bar'), ('port', '80,81'), ('discard', None)], [('bar', 'baz')]]
>>> split_header_words(['text/html; charset="iso-8859-1"'])
[[('text/html', None), ('charset', 'iso-8859-1')]]
>>> split_header_words([r'Basic realm="\\"foo\\bar\\""'])
[[('Basic', None), ('realm', '"foobar"')]]
"""
assert not isinstance(header_values, str)
result = []
for text in header_values:
orig_text = text
pairs = []
while text:
m = HEADER_TOKEN_RE.search(text)
if m:
text = unmatched(m)
name = m.group(1)
m = HEADER_QUOTED_VALUE_RE.search(text)
if m: # quoted value
text = unmatched(m)
value = m.group(1)
value = HEADER_ESCAPE_RE.sub(r"\\1", value)
else:
m = HEADER_VALUE_RE.search(text)
if m: # unquoted value
text = unmatched(m)
value = m.group(1)
value = value.rstrip()
else:
# no value, a lone token
value = None
pairs.append((name, value))
elif text.lstrip().startswith(","):
# concatenated headers, as per RFC 2616 section 4.2
text = text.lstrip()[1:]
if pairs: result.append(pairs)
pairs = []
else:
# skip junk
non_junk, nr_junk_chars = re.subn("^[=\\s;]*", "", text)
assert nr_junk_chars > 0, (
"split_header_words bug: '%s', '%s', %s" %
(orig_text, text, pairs))
text = non_junk
if pairs: result.append(pairs)
return result
HEADER_JOIN_ESCAPE_RE = re.compile(r"([\\"\\\\])")
def join_header_words(lists):
"""Do the inverse (almost) of the conversion done by split_header_words.
Takes a list of lists of (key, value) pairs and produces a single header
value. Attribute values are quoted if needed.
>>> join_header_words([[("text/plain", None), ("charset", "iso-8859/1")]])
'text/plain; charset="iso-8859/1"'
>>> join_header_words([[("text/plain", None)], [("charset", "iso-8859/1")]])
'text/plain, charset="iso-8859/1"'
"""
headers = []
for pairs in lists:
attr = []
for k, v in pairs:
if v is not None:
if not re.search(r"^\\w+$", v):
v = HEADER_JOIN_ESCAPE_RE.sub(r"\\\\\\1", v) # escape " and \\
v = '"%s"' % v
k = "%s=%s" % (k, v)
attr.append(k)
if attr: headers.append("; ".join(attr))
return ", ".join(headers)
def strip_quotes(text):
if text.startswith('"'):
text = text[1:]
if text.endswith('"'):
text = text[:-1]
return text
def parse_ns_headers(ns_headers):
"""Ad-hoc parser for Netscape protocol cookie-attributes.
The old Netscape cookie format for Set-Cookie can for instance contain
an unquoted "," in the expires field, so we have to use this ad-hoc
parser instead of split_header_words.
XXX This may not make the best possible effort to parse all the crap
that Netscape Cookie headers contain. Ronald Tschalar's HTTPClient
parser is probably better, so could do worse than following that if
this ever gives any trouble.
Currently, this is also used for parsing RFC 2109 cookies.
"""
known_attrs = ("expires", "domain", "path", "secure",
# RFC 2109 attrs (may turn up in Netscape cookies, too)
"version", "port", "max-age")
result = []
for ns_header in ns_headers:
pairs = []
version_set = False
for ii, param in enumerate(re.split(r";\\s*", ns_header)):
param = param.rstrip()
if param == "": continue
if "=" not in param:
k, v = param, None
else:
k, v = re.split(r"\\s*=\\s*", param, 1)
k = k.lstrip()
if ii != 0:
lc = k.lower()
if lc in known_attrs:
k = lc
if k == "version":
# This is an RFC 2109 cookie.
v = strip_quotes(v)
version_set = True
if k == "expires":
# convert expires date to seconds since epoch
v = http2time(strip_quotes(v)) # None if invalid
pairs.append((k, v))
if pairs:
if not version_set:
pairs.append(("version", "0"))
result.append(pairs)
return result
IPV4_RE = re.compile(r"\\.\\d+$", re.ASCII)
def is_HDN(text):
"""Return True if text is a host domain name."""
# XXX
# This may well be wrong. Which RFC is HDN defined in, if any (for
# the purposes of RFC 2965)?
# For the current implementation, what about IPv6? Remember to look
# at other uses of IPV4_RE also, if change this.
if IPV4_RE.search(text):
return False
if text == "":
return False
if text[0] == "." or text[-1] == ".":
return False
return True
def domain_match(A, B):
"""Return True if domain A domain-matches domain B, according to RFC 2965.
A and B may be host domain names or IP addresses.
RFC 2965, section 1:
Host names can be specified either as an IP address or a HDN string.
Sometimes we compare one host name with another. (Such comparisons SHALL
be case-insensitive.) Host A's name domain-matches host B's if
* their host name strings string-compare equal; or
* A is a HDN string and has the form NB, where N is a non-empty
name string, B has the form .B', and B' is a HDN string. (So,
x.y.com domain-matches .Y.com but not Y.com.)
Note that domain-match is not a commutative operation: a.b.c.com
domain-matches .c.com, but not the reverse.
"""
# Note that, if A or B are IP addresses, the only relevant part of the
# definition of the domain-match algorithm is the direct string-compare.
A = A.lower()
B = B.lower()
if A == B:
return True
if not is_HDN(A):
return False
i = A.rfind(B)
if i == -1 or i == 0:
# A does not have form NB, or N is the empty string
return False
if not B.startswith("."):
return False
if not is_HDN(B[1:]):
return False
return True
def liberal_is_HDN(text):
"""Return True if text is a sort-of-like a host domain name.
For accepting/blocking domains.
"""
if IPV4_RE.search(text):
return False
return True
def user_domain_match(A, B):
"""For blocking/accepting domains.
A and B may be host domain names or IP addresses.
"""
A = A.lower()
B = B.lower()
if not (liberal_is_HDN(A) and liberal_is_HDN(B)):
if A == B:
# equal IP addresses
return True
return False
initial_dot = B.startswith(".")
if initial_dot and A.endswith(B):
return True
if not initial_dot and A == B:
return True
return False
cut_port_re = re.compile(r":\\d+$", re.ASCII)
def request_host(request):
"""Return request-host, as defined by RFC 2965.
Variation from RFC: returned value is lowercased, for convenient
comparison.
"""
url = request.get_full_url()
host = urllib.parse.urlparse(url)[1]
if host == "":
host = request.get_header("Host", "")
# remove port, if present
host = cut_port_re.sub("", host, 1)
return host.lower()
def eff_request_host(request):
"""Return a tuple (request-host, effective request-host name).
As defined by RFC 2965, except both are lowercased.
"""
erhn = req_host = request_host(request)
if req_host.find(".") == -1 and not IPV4_RE.search(req_host):
erhn = req_host + ".local"
return req_host, erhn
def request_path(request):
"""Path component of request-URI, as defined by RFC 2965."""
url = request.get_full_url()
parts = urllib.parse.urlsplit(url)
path = escape_path(parts.path)
if not path.startswith("/"):
# fix bad RFC 2396 absoluteURI
path = "/" + path
return path
def request_port(request):
host = request.host
i = host.find(':')
if i >= 0:
port = host[i+1:]
try:
int(port)
except ValueError:
_debug("nonnumeric port: '%s'", port)
return None
else:
port = DEFAULT_HTTP_PORT
return port
# Characters in addition to A-Z, a-z, 0-9, '_', '.', and '-' that don't
# need to be escaped to form a valid HTTP URL (RFCs 2396 and 1738).
HTTP_PATH_SAFE = "%/;:@&=+$,!~*'()"
ESCAPED_CHAR_RE = re.compile(r"%([0-9a-fA-F][0-9a-fA-F])")
def uppercase_escaped_char(match):
return "%%%s" % match.group(1).upper()
def escape_path(path):
"""Escape any invalid characters in HTTP URL, and uppercase all escapes."""
# There's no knowing what character encoding was used to create URLs
# containing %-escapes, but since we have to pick one to escape invalid
# path characters, we pick UTF-8, as recommended in the HTML 4.0
# specification:
# http://www.w3.org/TR/REC-html40/appendix/notes.html#h-B.2.1
# And here, kind of: draft-fielding-uri-rfc2396bis-03
# (And in draft IRI specification: draft-duerst-iri-05)
# (And here, for new URI schemes: RFC 2718)
path = urllib.parse.quote(path, HTTP_PATH_SAFE)
path = ESCAPED_CHAR_RE.sub(uppercase_escaped_char, path)
return path
def reach(h):
"""Return reach of host h, as defined by RFC 2965, section 1.
The reach R of a host name H is defined as follows:
* If
- H is the host domain name of a host; and,
- H has the form A.B; and
- A has no embedded (that is, interior) dots; and
- B has at least one embedded dot, or B is the string "local".
then the reach of H is .B.
* Otherwise, the reach of H is H.
>>> reach("www.acme.com")
'.acme.com'
>>> reach("acme.com")
'acme.com'
>>> reach("acme.local")
'.local'
"""
i = h.find(".")
if i >= 0:
#a = h[:i] # this line is only here to show what a is
b = h[i+1:]
i = b.find(".")
if is_HDN(h) and (i >= 0 or b == "local"):
return "."+b
return h
def is_third_party(request):
"""
RFC 2965, section 3.3.6:
An unverifiable transaction is to a third-party host if its request-
host U does not domain-match the reach R of the request-host O in the
origin transaction.
"""
req_host = request_host(request)
if not domain_match(req_host, reach(request.origin_req_host)):
return True
else:
return False
#cookie类，非常简单，主要是cookie的属性
class Cookie:
"""HTTP Cookie.
This class represents both Netscape and RFC 2965 cookies.
This is deliberately a very simple class. It just holds attributes. It's
possible to construct Cookie instances that don't comply with the cookie
standards. CookieJar.make_cookies is the factory function for Cookie
objects -- it deals with cookie parsing, supplying defaults, and
normalising to the representation used in this class. CookiePolicy is
responsible for checking them to see whether they should be accepted from
and returned to the server.
Note that the port may be present in the headers, but unspecified ("Port"
rather than"Port=80", for example); if this is the case, port is None.
"""
def __init__(self, version, name, value,
port, port_specified,
domain, domain_specified, domain_initial_dot,
path, path_specified,
secure,
expires,
discard,
comment,
comment_url,
rest,
rfc2109=False,
):
if version is not None: version = int(version)#版本
if expires is not None: expires = int(expires)#过期时间
if port is None and port_specified is True:#端口
raise ValueError("if port is None, port_specified must be false")
self.version = version#版本
self.name = name#名称
self.value = value#值
self.port = port#端口
self.port_specified = port_specified#特殊端口
# normalise case, as per RFC 2965 section 3.3.3
self.domain = domain.lower()#域
self.domain_specified = domain_specified#特殊域
# Sigh. We need to know whether the domain given in the
# cookie-attribute had an initial dot, in order to follow RFC 2965
# (as clarified in draft errata). Needed for the returned $Domain
# value.
self.domain_initial_dot = domain_initial_dot#是否以.开头域名
self.path = path#路径
self.path_specified = path_specified#特殊路径
self.secure = secure#安全级别
self.expires = expires#过期时间
self.discard = discard#丢弃
self.comment = comment#
self.comment_url = comment_url
self.rfc2109 = rfc2109
self._rest = copy.copy(rest)
#有非标准属性
def has_nonstandard_attr(self, name):
return name in self._rest
#获得某个非标准属性
def get_nonstandard_attr(self, name, default=None):
return self._rest.get(name, default)
#设置非标准属性
def set_nonstandard_attr(self, name, value):
self._rest[name] = value
#是否过期
def is_expired(self, now=None):
if now is None: now = time.time()
if (self.expires is not None) and (self.expires <= now):
return True
return False
def __str__(self):
if self.port is None: p = ""
else: p = ":"+self.port
limit = self.domain + p + self.path
if self.value is not None:
namevalue = "%s=%s" % (self.name, self.value)
else:
namevalue = self.name
return "<Cookie %s for %s>" % (namevalue, limit)
def __repr__(self):
args = []
for name in ("version", "name", "value",
"port", "port_specified",
"domain", "domain_specified", "domain_initial_dot",
"path", "path_specified",
"secure", "expires", "discard", "comment", "comment_url",
):
attr = getattr(self, name)
args.append("%s=%s" % (name, repr(attr)))
args.append("rest=%s" % repr(self._rest))
args.append("rfc2109=%s" % repr(self.rfc2109))
return "Cookie(%s)" % ", ".join(args)
#cookie策略类，应该是做一些标准的公用的处理，
#还有一个默认策略类
class CookiePolicy:
"""Defines which cookies get accepted from and returned to server.
May also modify cookies, though this is probably a bad idea.
The subclass DefaultCookiePolicy defines the standard rules for Netscape
and RFC 2965 cookies -- override that if you want a customised policy.
"""
#设置可以从request中获得cookie，就返回true。为实现
def set_ok(self, cookie, request):
"""Return true if (and only if) cookie should be accepted from server.
Currently, pre-expired cookies never get this far -- the CookieJar
class deletes such cookies itself.
"""
raise NotImplementedError()
def return_ok(self, cookie, request):
"""Return true if (and only if) cookie should be returned to server."""
raise NotImplementedError()
def domain_return_ok(self, domain, request):
"""Return false if cookies should not be returned, given cookie domain.
"""
return True
def path_return_ok(self, path, request):
"""Return false if cookies should not be returned, given cookie path.
"""
return True
#实现了标准规则的策略类
class DefaultCookiePolicy(CookiePolicy):
"""Implements the standard rules for accepting and returning cookies."""
DomainStrictNoDots = 1
DomainStrictNonDomain = 2
DomainRFC2965Match = 4
DomainLiberal = 0
DomainStrict = DomainStrictNoDots|DomainStrictNonDomain
def __init__(self,
blocked_domains=None, allowed_domains=None,
netscape=True, rfc2965=False,
rfc2109_as_netscape=None,
hide_cookie2=False,
strict_domain=False,
strict_rfc2965_unverifiable=True,
strict_ns_unverifiable=False,
strict_ns_domain=DomainLiberal,
strict_ns_set_initial_dollar=False,
strict_ns_set_path=False,
):
"""Constructor arguments should be passed as keyword arguments only."""
self.netscape = netscape
self.rfc2965 = rfc2965
self.rfc2109_as_netscape = rfc2109_as_netscape
self.hide_cookie2 = hide_cookie2
self.strict_domain = strict_domain
self.strict_rfc2965_unverifiable = strict_rfc2965_unverifiable
self.strict_ns_unverifiable = strict_ns_unverifiable
self.strict_ns_domain = strict_ns_domain
self.strict_ns_set_initial_dollar = strict_ns_set_initial_dollar
self.strict_ns_set_path = strict_ns_set_path
if blocked_domains is not None:
self._blocked_domains = tuple(blocked_domains)
else:
self._blocked_domains = ()
if allowed_domains is not None:
allowed_domains = tuple(allowed_domains)
self._allowed_domains = allowed_domains
def blocked_domains(self):
"""Return the sequence of blocked domains (as a tuple)."""
return self._blocked_domains
def set_blocked_domains(self, blocked_domains):
"""Set the sequence of blocked domains."""
self._blocked_domains = tuple(blocked_domains)
def is_blocked(self, domain):
for blocked_domain in self._blocked_domains:
if user_domain_match(domain, blocked_domain):
return True
return False
def allowed_domains(self):
"""Return None, or the sequence of allowed domains (as a tuple)."""
return self._allowed_domains
def set_allowed_domains(self, allowed_domains):
"""Set the sequence of allowed domains, or None."""
if allowed_domains is not None:
allowed_domains = tuple(allowed_domains)
self._allowed_domains = allowed_domains
def is_not_allowed(self, domain):
if self._allowed_domains is None:
return False
for allowed_domain in self._allowed_domains:
if user_domain_match(domain, allowed_domain):
return False
return True
def set_ok(self, cookie, request):
"""
If you override .set_ok(), be sure to call this method. If it returns
false, so should your subclass (assuming your subclass wants to be more
strict about which cookies to accept).
"""
_debug(" - checking cookie %s=%s", cookie.name, cookie.value)
assert cookie.name is not None
for n in "version", "verifiability", "name", "path", "domain", "port":
fn_name = "set_ok_"+n
fn = getattr(self, fn_name)
if not fn(cookie, request):
return False
return True
def set_ok_version(self, cookie, request):
if cookie.version is None:
# Version is always set to 0 by parse_ns_headers if it's a Netscape
# cookie, so this must be an invalid RFC 2965 cookie.
_debug(" Set-Cookie2 without version attribute (%s=%s)",
cookie.name, cookie.value)
return False
if cookie.version > 0 and not self.rfc2965:
_debug(" RFC 2965 cookies are switched off")
return False
elif cookie.version == 0 and not self.netscape:
_debug(" Netscape cookies are switched off")
return False
return True
def set_ok_verifiability(self, cookie, request):
if request.unverifiable and is_third_party(request):
if cookie.version > 0 and self.strict_rfc2965_unverifiable:
_debug(" third-party RFC 2965 cookie during "
"unverifiable transaction")
return False
elif cookie.version == 0 and self.strict_ns_unverifiable:
_debug(" third-party Netscape cookie during "
"unverifiable transaction")
return False
return True
def set_ok_name(self, cookie, request):
# Try and stop servers setting V0 cookies designed to hack other
# servers that know both V0 and V1 protocols.
if (cookie.version == 0 and self.strict_ns_set_initial_dollar and
cookie.name.startswith("$")):
_debug(" illegal name (starts with '$'): '%s'", cookie.name)
return False
return True
def set_ok_path(self, cookie, request):
if cookie.path_specified:
req_path = request_path(request)
if ((cookie.version > 0 or
(cookie.version == 0 and self.strict_ns_set_path)) and
not req_path.startswith(cookie.path)):
_debug(" path attribute %s is not a prefix of request "
"path %s", cookie.path, req_path)
return False
return True
def set_ok_domain(self, cookie, request):
if self.is_blocked(cookie.domain):
_debug(" domain %s is in user block-list", cookie.domain)
return False
if self.is_not_allowed(cookie.domain):
_debug(" domain %s is not in user allow-list", cookie.domain)
return False
if cookie.domain_specified:
req_host, erhn = eff_request_host(request)
domain = cookie.domain
if self.strict_domain and (domain.count(".") >= 2):
# XXX This should probably be compared with the Konqueror
# (kcookiejar.cpp) and Mozilla implementations, but it's a
# losing battle.
i = domain.rfind(".")
j = domain.rfind(".", 0, i)
if j == 0: # domain like .foo.bar
tld = domain[i+1:]
sld = domain[j+1:i]
if sld.lower() in ("co", "ac", "com", "edu", "org", "net",
"gov", "mil", "int", "aero", "biz", "cat", "coop",
"info", "jobs", "mobi", "museum", "name", "pro",
"travel", "eu") and len(tld) == 2:
# domain like .co.uk
_debug(" country-code second level domain %s", domain)
return False
if domain.startswith("."):
undotted_domain = domain[1:]
else:
undotted_domain = domain
embedded_dots = (undotted_domain.find(".") >= 0)
if not embedded_dots and domain != ".local":
_debug(" non-local domain %s contains no embedded dot",
domain)
return False
if cookie.version == 0:
if (not erhn.endswith(domain) and
(not erhn.startswith(".") and
not ("."+erhn).endswith(domain))):
_debug(" effective request-host %s (even with added "
"initial dot) does not end with %s",
erhn, domain)
return False
if (cookie.version > 0 or
(self.strict_ns_domain & self.DomainRFC2965Match)):
if not domain_match(erhn, domain):
_debug(" effective request-host %s does not domain-match "
"%s", erhn, domain)
return False
if (cookie.version > 0 or
(self.strict_ns_domain & self.DomainStrictNoDots)):
host_prefix = req_host[:-len(domain)]
if (host_prefix.find(".") >= 0 and
not IPV4_RE.search(req_host)):
_debug(" host prefix %s for domain %s contains a dot",
host_prefix, domain)
return False
return True
def set_ok_port(self, cookie, request):
if cookie.port_specified:
req_port = request_port(request)
if req_port is None:
req_port = "80"
else:
req_port = str(req_port)
for p in cookie.port.split(","):
try:
int(p)
except ValueError:
_debug(" bad port %s (not numeric)", p)
return False
if p == req_port:
break
else:
_debug(" request port (%s) not found in %s",
req_port, cookie.port)
return False
return True
def return_ok(self, cookie, request):
"""
If you override .return_ok(), be sure to call this method. If it
returns false, so should your subclass (assuming your subclass wants to
be more strict about which cookies to return).
"""
# Path has already been checked by .path_return_ok(), and domain
# blocking done by .domain_return_ok().
_debug(" - checking cookie %s=%s", cookie.name, cookie.value)
for n in "version", "verifiability", "secure", "expires", "port", "domain":
fn_name = "return_ok_"+n
fn = getattr(self, fn_name)
if not fn(cookie, request):
return False
return True
def return_ok_version(self, cookie, request):
if cookie.version > 0 and not self.rfc2965:
_debug(" RFC 2965 cookies are switched off")
return False
elif cookie.version == 0 and not self.netscape:
_debug(" Netscape cookies are switched off")
return False
return True
def return_ok_verifiability(self, cookie, request):
if request.unverifiable and is_third_party(request):
if cookie.version > 0 and self.strict_rfc2965_unverifiable:
_debug(" third-party RFC 2965 cookie during unverifiable "
"transaction")
return False
elif cookie.version == 0 and self.strict_ns_unverifiable:
_debug(" third-party Netscape cookie during unverifiable "
"transaction")
return False
return True
def return_ok_secure(self, cookie, request):
if cookie.secure and request.type != "https":
_debug(" secure cookie with non-secure request")
return False
return True
def return_ok_expires(self, cookie, request):
if cookie.is_expired(self._now):
_debug(" cookie expired")
return False
return True
def return_ok_port(self, cookie, request):
if cookie.port:
req_port = request_port(request)
if req_port is None:
req_port = "80"
for p in cookie.port.split(","):
if p == req_port:
break
else:
_debug(" request port %s does not match cookie port %s",
req_port, cookie.port)
return False
return True
def return_ok_domain(self, cookie, request):
req_host, erhn = eff_request_host(request)
domain = cookie.domain
# strict check of non-domain cookies: Mozilla does this, MSIE5 doesn't
if (cookie.version == 0 and
(self.strict_ns_domain & self.DomainStrictNonDomain) and
not cookie.domain_specified and domain != erhn):
_debug(" cookie with unspecified domain does not string-compare "
"equal to request domain")
return False
if cookie.version > 0 and not domain_match(erhn, domain):
_debug(" effective request-host name %s does not domain-match "
"RFC 2965 cookie domain %s", erhn, domain)
return False
if cookie.version == 0 and not ("."+erhn).endswith(domain):
_debug(" request-host %s does not match Netscape cookie domain "
"%s", req_host, domain)
return False
return True
def domain_return_ok(self, domain, request):
# Liberal check of. This is here as an optimization to avoid
# having to load lots of MSIE cookie files unless necessary.
req_host, erhn = eff_request_host(request)
if not req_host.startswith("."):
req_host = "."+req_host
if not erhn.startswith("."):
erhn = "."+erhn
if not (req_host.endswith(domain) or erhn.endswith(domain)):
#_debug(" request domain %s does not match cookie domain %s",
# req_host, domain)
return False
if self.is_blocked(domain):
_debug(" domain %s is in user block-list", domain)
return False
if self.is_not_allowed(domain):
_debug(" domain %s is not in user allow-list", domain)
return False
return True
def path_return_ok(self, path, request):
_debug("- checking cookie path=%s", path)
req_path = request_path(request)
if not req_path.startswith(path):
_debug(" %s does not path-match %s", req_path, path)
return False
return True
def vals_sorted_by_key(adict):
keys = sorted(adict.keys())
return map(adict.get, keys)
def deepvalues(mapping):
"""Iterates over nested mapping, depth-first, in sorted order by key."""
values = vals_sorted_by_key(mapping)
for obj in values:
mapping = False
try:
obj.items
except AttributeError:
pass
else:
mapping = True
yield from deepvalues(obj)
if not mapping:
yield obj
# Used as second parameter to dict.get() method, to distinguish absent
# dict key from one with a None value.
class Absent: pass
class CookieJar:
"""Collection of HTTP cookies.
如果没有需求，可以不使用这方面的知识
You may not need to know about this class: try
urllib.request.build_opener(HTTPCookieProcessor).open(url).
"""
#各种正则表达式
non_word_re = re.compile(r"\\W")
quote_re = re.compile(r"([\\"\\\\])")
strict_domain_re = re.compile(r"\\.?[^.]*")
domain_re = re.compile(r"[^.]*")
dots_re = re.compile(r"^\\.+")
magic_re = re.compile(r"^\\#LWP-Cookies-(\\d+\\.\\d+)", re.ASCII)
#初始化
def __init__(self, policy=None):
if policy is None:
policy = DefaultCookiePolicy()#使用默认的cookie策略
self._policy = policy
self._cookies_lock = _threading.RLock()#cookie锁，对这个线程加锁解锁
self._cookies = #按照[domain][path][name]组织的多重字典结构
#设置cookie策略
def set_policy(self, policy):
self._policy = policy
def _cookies_for_domain(self, domain, request):
cookies = []
if not self._policy.domain_return_ok(domain, request):
return []
_debug("Checking %s for cookies to return", domain)
cookies_by_path = self._cookies[domain]
for path in cookies_by_path.keys():
if not self._policy.path_return_ok(path, request):
continue
cookies_by_name = cookies_by_path[path]
for cookie in cookies_by_name.values():
if not self._policy.return_ok(cookie, request):
_debug(" not returning cookie")
continue
_debug(" it's a match")
cookies.append(cookie)
return cookies
def _cookies_for_request(self, request):
"""Return a list of cookies to be returned to server."""
#返回服务端获得的cookies列表
cookies = []
for domain in self._cookies.keys():
cookies.extend(self._cookies_for_domain(domain, request))
return cookies
def _cookie_attrs(self, cookies):
"""Return a list of cookie-attributes to be returned to server.
like ['foo="bar"; $Path="/"', ...]
The $Version attribute is also added when appropriate (currently only
once per request).
"""
# add cookies in order of most specific (ie. longest) path first
cookies.sort(key=lambda a: len(a.path), reverse=True)
version_set = False
attrs = []
for cookie in cookies:
# set version of Cookie header
# XXX
# What should it be if multiple matching Set-Cookie headers have
# different versions themselves?
# Answer: there is no answer; was supposed to be settled by
# RFC 2965 errata, but that may never appear...
version = cookie.version
if not version_set:
version_set = True
if version > 0:
attrs.append("$Version=%s" % version)
# quote cookie value if necessary
# (not for Netscape protocol, which already has any quotes
# intact, due to the poorly-specified Netscape Cookie: syntax)
if ((cookie.value is not None) and
self.non_word_re.search(cookie.value) and version > 0):
value = self.quote_re.sub(r"\\\\\\1", cookie.value)
else:
value = cookie.value
# add cookie-attributes to be returned in Cookie header
if cookie.value is None:
attrs.append(cookie.name)
else:
attrs.append("%s=%s" % (cookie.name, value))
if version > 0:
if cookie.path_specified:
attrs.append('$Path="%s"' % cookie.path)
if cookie.domain.startswith("."):
domain = cookie.domain
if (not cookie.domain_initial_dot and
domain.startswith(".")):
domain = domain[1:]
attrs.append('$Domain="%s"' % domain)
if cookie.port is not None:
p = "$Port"
if cookie.port_specified:
p = p + ('="%s"' % cookie.port)
attrs.append(p)
return attrs
#把cookie附加到表头中
def add_cookie_header(self, request):
"""Add correct Cookie: header to request (urllib.request.Request object).
The Cookie2 header is also added unless policy.hide_cookie2 is true.
"""
_debug("add_cookie_header")
self._cookies_lock.acquire()
try:
self._policy._now = self._now = int(time.time())
cookies = self._cookies_for_request(request)
attrs = self._cookie_attrs(cookies)
if attrs:
if not request.has_header("Cookie"):
request.add_unredirected_header(
"Cookie", "; ".join(attrs))
# if necessary, advertise that we know RFC 2965
if (self._policy.rfc2965 and not self._policy.hide_cookie2 and
not request.has_header("Cookie2")):
for cookie in cookies:
if cookie.version != 1:
request.add_unredirected_header("Cookie2", '$Version="1"')
break
finally:
self._cookies_lock.release()
self.clear_expired_cookies()
#返回元组，其中包含cookie的信息
#attrs_set是Set-Cookie或者Set-Cookie2头中的键值对
#元组中包含名称，值，标准(discard, secure, version, expires or max-age, domain, path and port) ，以及剩余部分
def _normalized_cookie_tuples(self, attrs_set):
"""Return list of tuples containing normalised cookie information.
attrs_set is the list of lists of key,value pairs extracted from
the Set-Cookie or Set-Cookie2 headers.
Tuples are name, value, standard, rest, where name and value are the
cookie name and value, standard is a dictionary containing the standard
cookie-attributes (discard, secure, version, expires or max-age,
domain, path and port) and rest is a dictionary containing the rest of
the cookie-attributes.
"""
cookie_tuples = []
boolean_attrs = "discard", "secure"#布尔值的属性
value_attrs = ("version",
"expires", "max-age",
"domain", "path", "port",
"comment", "commenturl")
for cookie_attrs in attrs_set:
name, value = cookie_attrs[0]
# Build dictionary of standard cookie-attributes (standard) and
# dictionary of other cookie-attributes (rest).
# Note: expiry time is normalised to seconds since epoch. V0
# cookies should have the Expires cookie-attribute, and V1 cookies
# should have Max-Age, but since V1 includes RFC 2109 cookies (and
# since V0 cookies may be a mish-mash of Netscape and RFC 2109), we
# accept either (but prefer Max-Age).
max_age_set = False
bad_cookie = False
standard =
rest =
for k, v in cookie_attrs[1:]:
lc = k.lower()
# don't lose case distinction for unknown fields
if lc in value_attrs or lc in boolean_attrs:
k = lc
if k in boolean_attrs and v is None:
# boolean cookie-attribute is present, but has no value
# (like "discard", rather than "port=80")
v = True
if k in standard:
# only first value is significant
continue
if k == "domain":
if v is None:
_debug(" missing value for domain attribute")
bad_cookie = True
break
# RFC 2965 section 3.3.3
v = v.lower()
if k == "expires":
if max_age_set:
# Prefer max-age to expires (like Mozilla)
continue
if v is None:
_debug(" missing or invalid value for expires "
"attribute: treating as session cookie")
continue
if k == "max-age":
max_age_set = True
try:
v = int(v)
except ValueError:
_debug(" missing or invalid (non-numeric) value for "
"max-age attribute")
bad_cookie = True
break
# convert RFC 2965 Max-Age to seconds since epoch
# XXX Strictly you're supposed to follow RFC 2616
# age-calculation rules. Remember that zero Max-Age is a
# is a request to discard (old and new) cookie, though.
k = "expires"
v = self._now + v
if (k in value_attrs) or (k in boolean_attrs):
if (v is None and
k not in ("port", "comment", "commenturl")):
_debug(" missing value for %s attribute" % k)
bad_cookie = True
break
standard[k] = v
else:
rest[k] = v
if bad_cookie:
continue
cookie_tuples.append((name, value, standard, rest))
return cookie_tuples
def _cookie_from_cookie_tuple(self, tup, request):
# standard is dict of standard cookie-attributes, rest is dict of the
# rest of them
name, value, standard, rest = tup
domain = standard.get("domain", Absent)
path = standard.get("path", Absent)
port = standard.get("port", Absent)
expires = standard.get("expires", Absent)
# set the easy defaults
version = standard.get("version", None)
if version is not None:
try:
version = int(version)
except ValueError:
return None # invalid version, ignore cookie
secure = standard.get("secure", False)
# (discard is also set if expires is Absent)
discard = standard.get("discard", False)
comment = standard.get("comment", None)
comment_url = standard.get("commenturl", None)
# set default path
if path is not Absent and path != "":
path_specified = True
path = escape_path(path)
else:
path_specified = False
path = request_path(request)
i = path.rfind("/")
if i != -1:
if version == 0:
# Netscape spec parts company from reality here
path = path[:i]
else:
path = path[:i+1]
if len(path) == 0: path = "/"
# set default domain
domain_specified = domain is not Absent
# but first we have to remember whether it starts with a dot
domain_initial_dot = False
if domain_specified:
domain_initial_dot = bool(domain.startswith("."))
if domain is Absent:
req_host, erhn = eff_request_host(request)
domain = erhn
elif not domain.startswith("."):
domain = "."+domain
# set default port
port_specified = False
if port is not Absent:
if port is None:
# Port attr present, but has no value: default to request port.
# Cookie should then only be sent back on that port.
port = request_port(request)
else:
port_specified = True
port = re.sub(r"\\s+", "", port)
else:
# No port attr present. Cookie can be sent back on any port.
port = None
# set default expires and discard
if expires is Absent:
expires = None
discard = True
elif expires <= self._now:
# Expiry date in past is request to delete cookie. This can't be
# in DefaultCookiePolicy, because can't delete cookies there.
try:
self.clear(domain, path, name)
except KeyError:
pass
_debug("Expiring cookie, domain='%s', path='%s', name='%s'",
domain, path, name)
return None
return Cookie(version,
name, value,
port, port_specified,
domain, domain_specified, domain_initial_dot,
以上是关于Python3 cookiejar模块详解的主要内容，如果未能解决你的问题，请参考以下文章
urllib中的保存cookie使用，运用cookiejar来模拟登录人人网