dbrun
/
youtube-dl
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[YoutubeDL] Sanitize more fields (#13313)

This commit is contained in:
Sergey M․ 2017-06-08 22:53:14 +07:00
parent 76e6378358
commit c996943418
No known key found for this signature in database
GPG Key ID: 2C393E0F18A9236D
1 changed files with 37 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

52
youtube_dl/YoutubeDL.py
View File

@ -58,6 +58,7 @@ from .utils import (
format_bytes, format_bytes,
formatSeconds, formatSeconds,
GeoRestrictedError, GeoRestrictedError,
int_or_none,
ISO3166Utils, ISO3166Utils,
locked_file, locked_file,
make_HTTPS_handler, make_HTTPS_handler,
@ -302,6 +303,17 @@ class YoutubeDL(object):
postprocessor. postprocessor.
""" """
_NUMERIC_FIELDS = set((
'width', 'height', 'tbr', 'abr', 'asr', 'vbr', 'fps', 'filesize', 'filesize_approx',
'timestamp', 'upload_year', 'upload_month', 'upload_day',
'duration', 'view_count', 'like_count', 'dislike_count', 'repost_count',
'average_rating', 'comment_count', 'age_limit',
'start_time', 'end_time',
'chapter_number', 'season_number', 'episode_number',
'track_number', 'disc_number', 'release_year',
'playlist_index',
))
params = None params = None
_ies = [] _ies = []
_pps = [] _pps = []
@ -639,22 +651,11 @@ class YoutubeDL(object):
r'%%(\1)0%dd' % field_size_compat_map[mobj.group('field')], r'%%(\1)0%dd' % field_size_compat_map[mobj.group('field')],
outtmpl) outtmpl)
NUMERIC_FIELDS = set((
'width', 'height', 'tbr', 'abr', 'asr', 'vbr', 'fps', 'filesize', 'filesize_approx',
'timestamp', 'upload_year', 'upload_month', 'upload_day',
'duration', 'view_count', 'like_count', 'dislike_count', 'repost_count',
'average_rating', 'comment_count', 'age_limit',
'start_time', 'end_time',
'chapter_number', 'season_number', 'episode_number',
'track_number', 'disc_number', 'release_year',
'playlist_index',
))
# Missing numeric fields used together with integer presentation types # Missing numeric fields used together with integer presentation types
# in format specification will break the argument substitution since # in format specification will break the argument substitution since
# string 'NA' is returned for missing fields. We will patch output # string 'NA' is returned for missing fields. We will patch output
# template for missing fields to meet string presentation type. # template for missing fields to meet string presentation type.
for numeric_field in NUMERIC_FIELDS: for numeric_field in self._NUMERIC_FIELDS:
if numeric_field not in template_dict: if numeric_field not in template_dict:
# As of [1] format syntax is: # As of [1] format syntax is:
# %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type # %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
@ -1345,9 +1346,28 @@ class YoutubeDL(object):
if 'title' not in info_dict: if 'title' not in info_dict:
raise ExtractorError('Missing "title" field in extractor result') raise ExtractorError('Missing "title" field in extractor result')
if not isinstance(info_dict['id'], compat_str): def report_force_conversion(field, field_not, conversion):
self.report_warning('"id" field is not a string - forcing string conversion') self.report_warning(
info_dict['id'] = compat_str(info_dict['id']) '"%s" field is not %s - forcing %s conversion, there is an error in extractor'
% (field, field_not, conversion))
def sanitize_string_field(info, string_field):
field = info.get(string_field)
if field is None or isinstance(field, compat_str):
return
report_force_conversion(string_field, 'a string', 'string')
info[string_field] = compat_str(field)
def sanitize_numeric_fields(info):
for numeric_field in self._NUMERIC_FIELDS:
field = info.get(numeric_field)
if field is None or isinstance(field, compat_numeric_types):
continue
report_force_conversion(numeric_field, 'numeric', 'int')
info[numeric_field] = int_or_none(field)
sanitize_string_field(info_dict, 'id')
sanitize_numeric_fields(info_dict)
if 'playlist' not in info_dict: if 'playlist' not in info_dict:
# It isn't part of a playlist # It isn't part of a playlist
@ -1435,6 +1455,8 @@ class YoutubeDL(object):
if 'url' not in format: if 'url' not in format:
raise ExtractorError('Missing "url" key in result (index %d)' % i) raise ExtractorError('Missing "url" key in result (index %d)' % i)
sanitize_string_field(format, 'format_id')
sanitize_numeric_fields(format)
format['url'] = sanitize_url(format['url']) format['url'] = sanitize_url(format['url'])
if format.get('format_id') is None: if format.get('format_id') is None: