Mercurial > repos > guerler > hhblits

annotate lib/python3.8/site-packages/pip/_internal/network/download.py @ 0:9e54283cc701 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
author guerler
date Mon, 27 Jul 2020 03:47:31 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
1 """Download files with progress indicators.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
2 """
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
3 import cgi
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
4 import logging
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
5 import mimetypes
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
6 import os
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
7
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
8 from pip._vendor import requests
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
9 from pip._vendor.requests.models import CONTENT_CHUNK_SIZE
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
10
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
11 from pip._internal.models.index import PyPI
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
12 from pip._internal.network.cache import is_from_cache
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
13 from pip._internal.network.utils import response_chunks
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
14 from pip._internal.utils.misc import (
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
15 format_size,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
16 redact_auth_from_url,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
17 splitext,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
18 )
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
19 from pip._internal.utils.typing import MYPY_CHECK_RUNNING
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
20 from pip._internal.utils.ui import DownloadProgressProvider
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
21
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
22 if MYPY_CHECK_RUNNING:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
23 from typing import Iterable, Optional
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
24
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
25 from pip._vendor.requests.models import Response
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
26
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
27 from pip._internal.models.link import Link
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
28 from pip._internal.network.session import PipSession
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
29
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
30 logger = logging.getLogger(__name__)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
31
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
32
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
33 def _get_http_response_size(resp):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
34 # type: (Response) -> Optional[int]
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
35 try:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
36 return int(resp.headers['content-length'])
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
37 except (ValueError, KeyError, TypeError):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
38 return None
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
39
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
40
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
41 def _prepare_download(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
42 resp, # type: Response
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
43 link, # type: Link
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
44 progress_bar # type: str
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
45 ):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
46 # type: (...) -> Iterable[bytes]
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
47 total_length = _get_http_response_size(resp)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
48
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
49 if link.netloc == PyPI.file_storage_domain:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
50 url = link.show_url
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
51 else:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
52 url = link.url_without_fragment
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
53
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
54 logged_url = redact_auth_from_url(url)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
55
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
56 if total_length:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
57 logged_url = '{} ({})'.format(logged_url, format_size(total_length))
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
58
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
59 if is_from_cache(resp):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
60 logger.info("Using cached %s", logged_url)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
61 else:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
62 logger.info("Downloading %s", logged_url)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
63
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
64 if logger.getEffectiveLevel() > logging.INFO:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
65 show_progress = False
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
66 elif is_from_cache(resp):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
67 show_progress = False
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
68 elif not total_length:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
69 show_progress = True
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
70 elif total_length > (40 * 1000):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
71 show_progress = True
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
72 else:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
73 show_progress = False
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
74
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
75 chunks = response_chunks(resp, CONTENT_CHUNK_SIZE)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
76
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
77 if not show_progress:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
78 return chunks
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
79
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
80 return DownloadProgressProvider(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
81 progress_bar, max=total_length
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
82 )(chunks)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
83
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
84
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
85 def sanitize_content_filename(filename):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
86 # type: (str) -> str
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
87 """
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
88 Sanitize the "filename" value from a Content-Disposition header.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
89 """
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
90 return os.path.basename(filename)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
91
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
92
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
93 def parse_content_disposition(content_disposition, default_filename):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
94 # type: (str, str) -> str
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
95 """
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
96 Parse the "filename" value from a Content-Disposition header, and
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
97 return the default filename if the result is empty.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
98 """
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
99 _type, params = cgi.parse_header(content_disposition)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
100 filename = params.get('filename')
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
101 if filename:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
102 # We need to sanitize the filename to prevent directory traversal
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
103 # in case the filename contains ".." path parts.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
104 filename = sanitize_content_filename(filename)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
105 return filename or default_filename
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
106
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
107
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
108 def _get_http_response_filename(resp, link):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
109 # type: (Response, Link) -> str
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
110 """Get an ideal filename from the given HTTP response, falling back to
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
111 the link filename if not provided.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
112 """
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
113 filename = link.filename # fallback
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
114 # Have a look at the Content-Disposition header for a better guess
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
115 content_disposition = resp.headers.get('content-disposition')
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
116 if content_disposition:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
117 filename = parse_content_disposition(content_disposition, filename)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
118 ext = splitext(filename)[1] # type: Optional[str]
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
119 if not ext:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
120 ext = mimetypes.guess_extension(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
121 resp.headers.get('content-type', '')
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
122 )
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
123 if ext:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
124 filename += ext
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
125 if not ext and link.url != resp.url:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
126 ext = os.path.splitext(resp.url)[1]
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
127 if ext:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
128 filename += ext
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
129 return filename
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
130
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
131
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
132 def _http_get_download(session, link):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
133 # type: (PipSession, Link) -> Response
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
134 target_url = link.url.split('#', 1)[0]
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
135 resp = session.get(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
136 target_url,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
137 # We use Accept-Encoding: identity here because requests
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
138 # defaults to accepting compressed responses. This breaks in
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
139 # a variety of ways depending on how the server is configured.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
140 # - Some servers will notice that the file isn't a compressible
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
141 # file and will leave the file alone and with an empty
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
142 # Content-Encoding
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
143 # - Some servers will notice that the file is already
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
144 # compressed and will leave the file alone and will add a
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
145 # Content-Encoding: gzip header
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
146 # - Some servers won't notice anything at all and will take
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
147 # a file that's already been compressed and compress it again
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
148 # and set the Content-Encoding: gzip header
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
149 # By setting this to request only the identity encoding We're
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
150 # hoping to eliminate the third case. Hopefully there does not
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
151 # exist a server which when given a file will notice it is
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
152 # already compressed and that you're not asking for a
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
153 # compressed file and will then decompress it before sending
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
154 # because if that's the case I don't think it'll ever be
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
155 # possible to make this work.
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
156 headers={"Accept-Encoding": "identity"},
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
157 stream=True,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
158 )
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
159 resp.raise_for_status()
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
160 return resp
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
161
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
162
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
163 class Download(object):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
164 def __init__(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
165 self,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
166 response, # type: Response
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
167 filename, # type: str
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
168 chunks, # type: Iterable[bytes]
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
169 ):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
170 # type: (...) -> None
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
171 self.response = response
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
172 self.filename = filename
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
173 self.chunks = chunks
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
174
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
175
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
176 class Downloader(object):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
177 def __init__(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
178 self,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
179 session, # type: PipSession
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
180 progress_bar, # type: str
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
181 ):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
182 # type: (...) -> None
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
183 self._session = session
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
184 self._progress_bar = progress_bar
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
185
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
186 def __call__(self, link):
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
187 # type: (Link) -> Download
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
188 try:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
189 resp = _http_get_download(self._session, link)
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
190 except requests.HTTPError as e:
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
191 logger.critical(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
192 "HTTP error %s while getting %s", e.response.status_code, link
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
193 )
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
194 raise
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
195
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
196 return Download(
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
197 resp,
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
198 _get_http_response_filename(resp, link),
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
199 _prepare_download(resp, link, self._progress_bar),
9e54283cc701 "planemo upload commit d12c32a45bcd441307e632fca6d9af7d60289d44"
guerler
parents:
diff changeset
200 )