2c8595098c
git-svn-id: http://google-refine.googlecode.com/svn/trunk@883 7d457c2a-affb-35e4-300a-418c747d4874
481 lines
20 KiB
Python
481 lines
20 KiB
Python
# Copyright (C) 2004-2006 Python Software Foundation
|
||
# Authors: Baxter, Wouters and Warsaw
|
||
# Contact: email-sig@python.org
|
||
|
||
"""FeedParser - An email feed parser.
|
||
|
||
The feed parser implements an interface for incrementally parsing an email
|
||
message, line by line. This has advantages for certain applications, such as
|
||
those reading email messages off a socket.
|
||
|
||
FeedParser.feed() is the primary interface for pushing new data into the
|
||
parser. It returns when there's nothing more it can do with the available
|
||
data. When you have no more data to push into the parser, call .close().
|
||
This completes the parsing and returns the root message object.
|
||
|
||
The other advantage of this parser is that it will never throw a parsing
|
||
exception. Instead, when it finds something unexpected, it adds a 'defect' to
|
||
the current message. Defects are just instances that live on the message
|
||
object's .defects attribute.
|
||
"""
|
||
|
||
__all__ = ['FeedParser']
|
||
|
||
import re
|
||
|
||
from email import errors
|
||
from email import message
|
||
|
||
NLCRE = re.compile('\r\n|\r|\n')
|
||
NLCRE_bol = re.compile('(\r\n|\r|\n)')
|
||
NLCRE_eol = re.compile('(\r\n|\r|\n)$')
|
||
NLCRE_crack = re.compile('(\r\n|\r|\n)')
|
||
# RFC 2822 $3.6.8 Optional fields. ftext is %d33-57 / %d59-126, Any character
|
||
# except controls, SP, and ":".
|
||
headerRE = re.compile(r'^(From |[\041-\071\073-\176]{1,}:|[\t ])')
|
||
EMPTYSTRING = ''
|
||
NL = '\n'
|
||
|
||
NeedMoreData = object()
|
||
|
||
|
||
|
||
class BufferedSubFile(object):
|
||
"""A file-ish object that can have new data loaded into it.
|
||
|
||
You can also push and pop line-matching predicates onto a stack. When the
|
||
current predicate matches the current line, a false EOF response
|
||
(i.e. empty string) is returned instead. This lets the parser adhere to a
|
||
simple abstraction -- it parses until EOF closes the current message.
|
||
"""
|
||
def __init__(self):
|
||
# The last partial line pushed into this object.
|
||
self._partial = ''
|
||
# The list of full, pushed lines, in reverse order
|
||
self._lines = []
|
||
# The stack of false-EOF checking predicates.
|
||
self._eofstack = []
|
||
# A flag indicating whether the file has been closed or not.
|
||
self._closed = False
|
||
|
||
def push_eof_matcher(self, pred):
|
||
self._eofstack.append(pred)
|
||
|
||
def pop_eof_matcher(self):
|
||
return self._eofstack.pop()
|
||
|
||
def close(self):
|
||
# Don't forget any trailing partial line.
|
||
self._lines.append(self._partial)
|
||
self._partial = ''
|
||
self._closed = True
|
||
|
||
def readline(self):
|
||
if not self._lines:
|
||
if self._closed:
|
||
return ''
|
||
return NeedMoreData
|
||
# Pop the line off the stack and see if it matches the current
|
||
# false-EOF predicate.
|
||
line = self._lines.pop()
|
||
# RFC 2046, section 5.1.2 requires us to recognize outer level
|
||
# boundaries at any level of inner nesting. Do this, but be sure it's
|
||
# in the order of most to least nested.
|
||
for ateof in self._eofstack[::-1]:
|
||
if ateof(line):
|
||
# We're at the false EOF. But push the last line back first.
|
||
self._lines.append(line)
|
||
return ''
|
||
return line
|
||
|
||
def unreadline(self, line):
|
||
# Let the consumer push a line back into the buffer.
|
||
assert line is not NeedMoreData
|
||
self._lines.append(line)
|
||
|
||
def push(self, data):
|
||
"""Push some new data into this object."""
|
||
# Handle any previous leftovers
|
||
data, self._partial = self._partial + data, ''
|
||
# Crack into lines, but preserve the newlines on the end of each
|
||
parts = NLCRE_crack.split(data)
|
||
# The *ahem* interesting behaviour of re.split when supplied grouping
|
||
# parentheses is that the last element of the resulting list is the
|
||
# data after the final RE. In the case of a NL/CR terminated string,
|
||
# this is the empty string.
|
||
self._partial = parts.pop()
|
||
# parts is a list of strings, alternating between the line contents
|
||
# and the eol character(s). Gather up a list of lines after
|
||
# re-attaching the newlines.
|
||
lines = []
|
||
for i in range(len(parts) // 2):
|
||
lines.append(parts[i*2] + parts[i*2+1])
|
||
self.pushlines(lines)
|
||
|
||
def pushlines(self, lines):
|
||
# Reverse and insert at the front of the lines.
|
||
self._lines[:0] = lines[::-1]
|
||
|
||
def is_closed(self):
|
||
return self._closed
|
||
|
||
def __iter__(self):
|
||
return self
|
||
|
||
def next(self):
|
||
line = self.readline()
|
||
if line == '':
|
||
raise StopIteration
|
||
return line
|
||
|
||
|
||
|
||
class FeedParser:
|
||
"""A feed-style parser of email."""
|
||
|
||
def __init__(self, _factory=message.Message):
|
||
"""_factory is called with no arguments to create a new message obj"""
|
||
self._factory = _factory
|
||
self._input = BufferedSubFile()
|
||
self._msgstack = []
|
||
self._parse = self._parsegen().next
|
||
self._cur = None
|
||
self._last = None
|
||
self._headersonly = False
|
||
|
||
# Non-public interface for supporting Parser's headersonly flag
|
||
def _set_headersonly(self):
|
||
self._headersonly = True
|
||
|
||
def feed(self, data):
|
||
"""Push more data into the parser."""
|
||
self._input.push(data)
|
||
self._call_parse()
|
||
|
||
def _call_parse(self):
|
||
try:
|
||
self._parse()
|
||
except StopIteration:
|
||
pass
|
||
|
||
def close(self):
|
||
"""Parse all remaining data and return the root message object."""
|
||
self._input.close()
|
||
self._call_parse()
|
||
root = self._pop_message()
|
||
assert not self._msgstack
|
||
# Look for final set of defects
|
||
if root.get_content_maintype() == 'multipart' \
|
||
and not root.is_multipart():
|
||
root.defects.append(errors.MultipartInvariantViolationDefect())
|
||
return root
|
||
|
||
def _new_message(self):
|
||
msg = self._factory()
|
||
if self._cur and self._cur.get_content_type() == 'multipart/digest':
|
||
msg.set_default_type('message/rfc822')
|
||
if self._msgstack:
|
||
self._msgstack[-1].attach(msg)
|
||
self._msgstack.append(msg)
|
||
self._cur = msg
|
||
self._last = msg
|
||
|
||
def _pop_message(self):
|
||
retval = self._msgstack.pop()
|
||
if self._msgstack:
|
||
self._cur = self._msgstack[-1]
|
||
else:
|
||
self._cur = None
|
||
return retval
|
||
|
||
def _parsegen(self):
|
||
# Create a new message and start by parsing headers.
|
||
self._new_message()
|
||
headers = []
|
||
# Collect the headers, searching for a line that doesn't match the RFC
|
||
# 2822 header or continuation pattern (including an empty line).
|
||
for line in self._input:
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
if not headerRE.match(line):
|
||
# If we saw the RFC defined header/body separator
|
||
# (i.e. newline), just throw it away. Otherwise the line is
|
||
# part of the body so push it back.
|
||
if not NLCRE.match(line):
|
||
self._input.unreadline(line)
|
||
break
|
||
headers.append(line)
|
||
# Done with the headers, so parse them and figure out what we're
|
||
# supposed to see in the body of the message.
|
||
self._parse_headers(headers)
|
||
# Headers-only parsing is a backwards compatibility hack, which was
|
||
# necessary in the older parser, which could throw errors. All
|
||
# remaining lines in the input are thrown into the message body.
|
||
if self._headersonly:
|
||
lines = []
|
||
while True:
|
||
line = self._input.readline()
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
if line == '':
|
||
break
|
||
lines.append(line)
|
||
self._cur.set_payload(EMPTYSTRING.join(lines))
|
||
return
|
||
if self._cur.get_content_type() == 'message/delivery-status':
|
||
# message/delivery-status contains blocks of headers separated by
|
||
# a blank line. We'll represent each header block as a separate
|
||
# nested message object, but the processing is a bit different
|
||
# than standard message/* types because there is no body for the
|
||
# nested messages. A blank line separates the subparts.
|
||
while True:
|
||
self._input.push_eof_matcher(NLCRE.match)
|
||
for retval in self._parsegen():
|
||
if retval is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
break
|
||
msg = self._pop_message()
|
||
# We need to pop the EOF matcher in order to tell if we're at
|
||
# the end of the current file, not the end of the last block
|
||
# of message headers.
|
||
self._input.pop_eof_matcher()
|
||
# The input stream must be sitting at the newline or at the
|
||
# EOF. We want to see if we're at the end of this subpart, so
|
||
# first consume the blank line, then test the next line to see
|
||
# if we're at this subpart's EOF.
|
||
while True:
|
||
line = self._input.readline()
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
break
|
||
while True:
|
||
line = self._input.readline()
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
break
|
||
if line == '':
|
||
break
|
||
# Not at EOF so this is a line we're going to need.
|
||
self._input.unreadline(line)
|
||
return
|
||
if self._cur.get_content_maintype() == 'message':
|
||
# The message claims to be a message/* type, then what follows is
|
||
# another RFC 2822 message.
|
||
for retval in self._parsegen():
|
||
if retval is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
break
|
||
self._pop_message()
|
||
return
|
||
if self._cur.get_content_maintype() == 'multipart':
|
||
boundary = self._cur.get_boundary()
|
||
if boundary is None:
|
||
# The message /claims/ to be a multipart but it has not
|
||
# defined a boundary. That's a problem which we'll handle by
|
||
# reading everything until the EOF and marking the message as
|
||
# defective.
|
||
self._cur.defects.append(errors.NoBoundaryInMultipartDefect())
|
||
lines = []
|
||
for line in self._input:
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
lines.append(line)
|
||
self._cur.set_payload(EMPTYSTRING.join(lines))
|
||
return
|
||
# Create a line match predicate which matches the inter-part
|
||
# boundary as well as the end-of-multipart boundary. Don't push
|
||
# this onto the input stream until we've scanned past the
|
||
# preamble.
|
||
separator = '--' + boundary
|
||
boundaryre = re.compile(
|
||
'(?P<sep>' + re.escape(separator) +
|
||
r')(?P<end>--)?(?P<ws>[ \t]*)(?P<linesep>\r\n|\r|\n)?$')
|
||
capturing_preamble = True
|
||
preamble = []
|
||
linesep = False
|
||
while True:
|
||
line = self._input.readline()
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
if line == '':
|
||
break
|
||
mo = boundaryre.match(line)
|
||
if mo:
|
||
# If we're looking at the end boundary, we're done with
|
||
# this multipart. If there was a newline at the end of
|
||
# the closing boundary, then we need to initialize the
|
||
# epilogue with the empty string (see below).
|
||
if mo.group('end'):
|
||
linesep = mo.group('linesep')
|
||
break
|
||
# We saw an inter-part boundary. Were we in the preamble?
|
||
if capturing_preamble:
|
||
if preamble:
|
||
# According to RFC 2046, the last newline belongs
|
||
# to the boundary.
|
||
lastline = preamble[-1]
|
||
eolmo = NLCRE_eol.search(lastline)
|
||
if eolmo:
|
||
preamble[-1] = lastline[:-len(eolmo.group(0))]
|
||
self._cur.preamble = EMPTYSTRING.join(preamble)
|
||
capturing_preamble = False
|
||
self._input.unreadline(line)
|
||
continue
|
||
# We saw a boundary separating two parts. Consume any
|
||
# multiple boundary lines that may be following. Our
|
||
# interpretation of RFC 2046 BNF grammar does not produce
|
||
# body parts within such double boundaries.
|
||
while True:
|
||
line = self._input.readline()
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
mo = boundaryre.match(line)
|
||
if not mo:
|
||
self._input.unreadline(line)
|
||
break
|
||
# Recurse to parse this subpart; the input stream points
|
||
# at the subpart's first line.
|
||
self._input.push_eof_matcher(boundaryre.match)
|
||
for retval in self._parsegen():
|
||
if retval is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
break
|
||
# Because of RFC 2046, the newline preceding the boundary
|
||
# separator actually belongs to the boundary, not the
|
||
# previous subpart's payload (or epilogue if the previous
|
||
# part is a multipart).
|
||
if self._last.get_content_maintype() == 'multipart':
|
||
epilogue = self._last.epilogue
|
||
if epilogue == '':
|
||
self._last.epilogue = None
|
||
elif epilogue is not None:
|
||
mo = NLCRE_eol.search(epilogue)
|
||
if mo:
|
||
end = len(mo.group(0))
|
||
self._last.epilogue = epilogue[:-end]
|
||
else:
|
||
payload = self._last.get_payload()
|
||
if isinstance(payload, basestring):
|
||
mo = NLCRE_eol.search(payload)
|
||
if mo:
|
||
payload = payload[:-len(mo.group(0))]
|
||
self._last.set_payload(payload)
|
||
self._input.pop_eof_matcher()
|
||
self._pop_message()
|
||
# Set the multipart up for newline cleansing, which will
|
||
# happen if we're in a nested multipart.
|
||
self._last = self._cur
|
||
else:
|
||
# I think we must be in the preamble
|
||
assert capturing_preamble
|
||
preamble.append(line)
|
||
# We've seen either the EOF or the end boundary. If we're still
|
||
# capturing the preamble, we never saw the start boundary. Note
|
||
# that as a defect and store the captured text as the payload.
|
||
# Everything from here to the EOF is epilogue.
|
||
if capturing_preamble:
|
||
self._cur.defects.append(errors.StartBoundaryNotFoundDefect())
|
||
self._cur.set_payload(EMPTYSTRING.join(preamble))
|
||
epilogue = []
|
||
for line in self._input:
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
self._cur.epilogue = EMPTYSTRING.join(epilogue)
|
||
return
|
||
# If the end boundary ended in a newline, we'll need to make sure
|
||
# the epilogue isn't None
|
||
if linesep:
|
||
epilogue = ['']
|
||
else:
|
||
epilogue = []
|
||
for line in self._input:
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
epilogue.append(line)
|
||
# Any CRLF at the front of the epilogue is not technically part of
|
||
# the epilogue. Also, watch out for an empty string epilogue,
|
||
# which means a single newline.
|
||
if epilogue:
|
||
firstline = epilogue[0]
|
||
bolmo = NLCRE_bol.match(firstline)
|
||
if bolmo:
|
||
epilogue[0] = firstline[len(bolmo.group(0)):]
|
||
self._cur.epilogue = EMPTYSTRING.join(epilogue)
|
||
return
|
||
# Otherwise, it's some non-multipart type, so the entire rest of the
|
||
# file contents becomes the payload.
|
||
lines = []
|
||
for line in self._input:
|
||
if line is NeedMoreData:
|
||
yield NeedMoreData
|
||
continue
|
||
lines.append(line)
|
||
self._cur.set_payload(EMPTYSTRING.join(lines))
|
||
|
||
def _parse_headers(self, lines):
|
||
# Passed a list of lines that make up the headers for the current msg
|
||
lastheader = ''
|
||
lastvalue = []
|
||
for lineno, line in enumerate(lines):
|
||
# Check for continuation
|
||
if line[0] in ' \t':
|
||
if not lastheader:
|
||
# The first line of the headers was a continuation. This
|
||
# is illegal, so let's note the defect, store the illegal
|
||
# line, and ignore it for purposes of headers.
|
||
defect = errors.FirstHeaderLineIsContinuationDefect(line)
|
||
self._cur.defects.append(defect)
|
||
continue
|
||
lastvalue.append(line)
|
||
continue
|
||
if lastheader:
|
||
# XXX reconsider the joining of folded lines
|
||
lhdr = EMPTYSTRING.join(lastvalue)[:-1].rstrip('\r\n')
|
||
self._cur[lastheader] = lhdr
|
||
lastheader, lastvalue = '', []
|
||
# Check for envelope header, i.e. unix-from
|
||
if line.startswith('From '):
|
||
if lineno == 0:
|
||
# Strip off the trailing newline
|
||
mo = NLCRE_eol.search(line)
|
||
if mo:
|
||
line = line[:-len(mo.group(0))]
|
||
self._cur.set_unixfrom(line)
|
||
continue
|
||
elif lineno == len(lines) - 1:
|
||
# Something looking like a unix-from at the end - it's
|
||
# probably the first line of the body, so push back the
|
||
# line and stop.
|
||
self._input.unreadline(line)
|
||
return
|
||
else:
|
||
# Weirdly placed unix-from line. Note this as a defect
|
||
# and ignore it.
|
||
defect = errors.MisplacedEnvelopeHeaderDefect(line)
|
||
self._cur.defects.append(defect)
|
||
continue
|
||
# Split the line on the colon separating field name from value.
|
||
i = line.find(':')
|
||
if i < 0:
|
||
defect = errors.MalformedHeaderDefect(line)
|
||
self._cur.defects.append(defect)
|
||
continue
|
||
lastheader = line[:i]
|
||
lastvalue = [line[i+1:].lstrip()]
|
||
# Done with all the lines, so handle the last header.
|
||
if lastheader:
|
||
# XXX reconsider the joining of folded lines
|
||
self._cur[lastheader] = EMPTYSTRING.join(lastvalue).rstrip('\r\n')
|