wsgi/request.py
author Julien Cristau <julien.cristau@logilab.fr>
Tue, 23 Sep 2014 14:19:11 +0200
changeset 9973 bbe05c74eb92
parent 9821 2077c8da1893
child 9990 c84ad981fc4a
permissions -rw-r--r--
[doc] proofreading CWEP002 section
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
7879
9aae456abab5 [pylint] fix pylint detected errors and tweak it so that pylint -E will be much less verbose next time (+ update some copyrights on the way)
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 5426
diff changeset
     1
# copyright 2003-2011 LOGILAB S.A. (Paris, FRANCE), all rights reserved.
5421
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     2
# contact http://www.logilab.fr/ -- mailto:contact@logilab.fr
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     3
#
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     4
# This file is part of CubicWeb.
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     5
#
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     6
# CubicWeb is free software: you can redistribute it and/or modify it under the
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     7
# terms of the GNU Lesser General Public License as published by the Free
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     8
# Software Foundation, either version 2.1 of the License, or (at your option)
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
     9
# any later version.
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
    10
#
5424
8ecbcbff9777 replace logilab-common by CubicWeb in disclaimer
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 5421
diff changeset
    11
# CubicWeb is distributed in the hope that it will be useful, but WITHOUT
5421
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
    12
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
    13
# FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public License for more
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
    14
# details.
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
    15
#
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
    16
# You should have received a copy of the GNU Lesser General Public License along
8167de96c523 proper licensing information (LGPL-2.1). Hope I get it right this time.
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 4212
diff changeset
    17
# with CubicWeb.  If not, see <http://www.gnu.org/licenses/>.
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    18
"""WSGI request adapter for cubicweb
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    19
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    20
NOTE: each docstring tagged with ``COME FROM DJANGO`` means that
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    21
the code has been taken (or adapted) from Djanco source code :
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    22
  http://www.djangoproject.com/
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    23
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    24
"""
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    25
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    26
__docformat__ = "restructuredtext en"
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    27
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    28
from StringIO import StringIO
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    29
from urllib import quote
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    30
from urlparse import parse_qs
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    31
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    32
from cubicweb.multipart import copy_file, parse_form_data
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    33
from cubicweb.web.request import CubicWebRequestBase
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    34
from cubicweb.wsgi import pformat, normalize_header
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    35
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    36
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    37
class CubicWebWsgiRequest(CubicWebRequestBase):
8752
e19f4bba89cd Add CubicWebRequestBase.content (closes #2742453)
Julien Cristau <julien.cristau@logilab.fr>
parents: 8316
diff changeset
    38
    """most of this code COMES FROM DJANGO
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    39
    """
1802
d628defebc17 delete-trailing-whitespace + some copyright update
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents: 0
diff changeset
    40
8309
48ef505aa9f9 [request] gather all base_url logic in a single place (closes #2200756)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 7879
diff changeset
    41
    def __init__(self, environ, vreg):
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    42
        self.environ = environ
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    43
        self.path = environ['PATH_INFO']
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    44
        self.method = environ['REQUEST_METHOD'].upper()
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    45
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    46
        # content_length "may be empty or absent"
9563
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    47
        try:
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    48
            length = int(environ['CONTENT_LENGTH'])
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    49
        except (KeyError, ValueError):
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    50
            length = 0
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    51
        # wsgi.input is not seekable, so copy the request contents to a temporary file
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    52
        if length < 100000:
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    53
            self.content = StringIO()
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    54
        else:
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    55
            self.content = tempfile.TemporaryFile()
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    56
        copy_file(environ['wsgi.input'], self.content, maxread=length)
9563
48f0ff3e2a32 [wsgi] make sure request.content is available for consumption
Julien Cristau <julien.cristau@logilab.fr>
parents: 8752
diff changeset
    57
        self.content.seek(0, 0)
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    58
        environ['wsgi.input'] = self.content
8314
cfd6ab461849 [Web-Request] Use rich header (closes #2204164)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 8309
diff changeset
    59
cfd6ab461849 [Web-Request] Use rich header (closes #2204164)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 8309
diff changeset
    60
        headers_in = dict((normalize_header(k[5:]), v) for k, v in self.environ.items()
cfd6ab461849 [Web-Request] Use rich header (closes #2204164)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 8309
diff changeset
    61
                          if k.startswith('HTTP_'))
9737
c6f47e635845 [wsgi] look at wsgi.url_scheme to decide if we're on https
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    62
        https = self.is_secure()
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    63
        post, files = self.get_posted_data()
8314
cfd6ab461849 [Web-Request] Use rich header (closes #2204164)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 8309
diff changeset
    64
cfd6ab461849 [Web-Request] Use rich header (closes #2204164)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 8309
diff changeset
    65
        super(CubicWebWsgiRequest, self).__init__(vreg, https, post,
cfd6ab461849 [Web-Request] Use rich header (closes #2204164)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 8309
diff changeset
    66
                                                  headers= headers_in)
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    67
        if files is not None:
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    68
            for key, part in files.iteritems():
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    69
                name = None
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    70
                if part.filename is not None:
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    71
                    name = unicode(part.filename, self.encoding)
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
    72
                self.form[key] = (name, part.file)
1802
d628defebc17 delete-trailing-whitespace + some copyright update
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents: 0
diff changeset
    73
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    74
    def __repr__(self):
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    75
        # Since this is called as part of error handling, we need to be very
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    76
        # robust against potentially malformed input.
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    77
        form = pformat(self.form)
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    78
        meta = pformat(self.environ)
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    79
        return '<CubicWebWsgiRequest\FORM:%s,\nMETA:%s>' % \
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    80
            (form, meta)
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    81
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    82
    ## cubicweb request interface ################################################
1802
d628defebc17 delete-trailing-whitespace + some copyright update
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents: 0
diff changeset
    83
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    84
    def http_method(self):
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    85
        """returns 'POST', 'GET', 'HEAD', etc."""
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    86
        return self.method
1802
d628defebc17 delete-trailing-whitespace + some copyright update
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents: 0
diff changeset
    87
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    88
    def relative_path(self, includeparams=True):
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    89
        """return the normalized path of the request (ie at least relative
2476
1294a6bdf3bf application -> instance where it makes sense
Sylvain Thénault <sylvain.thenault@logilab.fr>
parents: 1977
diff changeset
    90
        to the instance's root, but some other normalization may be needed
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    91
        so that the returned path may be used to compare to generated urls
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    92
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    93
        :param includeparams:
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    94
           boolean indicating if GET form parameters should be kept in the path
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    95
        """
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    96
        path = self.environ['PATH_INFO']
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    97
        path = path[1:] # remove leading '/'
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    98
        if includeparams:
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
    99
            qs = self.environ.get('QUERY_STRING')
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   100
            if qs:
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   101
                return '%s?%s' % (path, qs)
1802
d628defebc17 delete-trailing-whitespace + some copyright update
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents: 0
diff changeset
   102
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   103
        return path
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   104
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   105
    ## wsgi request helpers ###################################################
1802
d628defebc17 delete-trailing-whitespace + some copyright update
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents: 0
diff changeset
   106
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   107
    def is_secure(self):
9741
b6cc9da206df [wsgi] the spec says wsgi.url_scheme *must* be provided
Julien Cristau <julien.cristau@logilab.fr>
parents: 9737
diff changeset
   108
        return self.environ['wsgi.url_scheme'] == 'https'
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   109
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   110
    def get_posted_data(self):
8314
cfd6ab461849 [Web-Request] Use rich header (closes #2204164)
Pierre-Yves David <pierre-yves.david@logilab.fr>
parents: 8309
diff changeset
   111
        # The WSGI spec says 'QUERY_STRING' may be absent.
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
   112
        post = parse_qs(self.environ.get('QUERY_STRING', ''))
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   113
        files = None
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   114
        if self.method == 'POST':
9735
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
   115
            forms, files = parse_form_data(self.environ, strict=True,
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
   116
                                           mem_limit=self.vreg.config['max-post-length'])
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
   117
            post.update(forms)
b71158815bc8 [wsgi] avoid reading the entire request body in memory
Julien Cristau <julien.cristau@logilab.fr>
parents: 9563
diff changeset
   118
        self.content.seek(0, 0)
0
b97547f5f1fa Showtime !
Adrien Di Mascio <Adrien.DiMascio@logilab.fr>
parents:
diff changeset
   119
        return post, files