Merge remote-tracking branch 'rupertbaxter2/master'

[youtube-dl] / youtube_dl / YoutubeDL.py

diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py
index 6d5c401dfa3418350d03105e3efe537cecc7d911..8f34b17b4135ff26263e52fc8e9e0badc197680d 100755 (executable)
--- a/youtube_dl/YoutubeDL.py
+++ b/youtube_dl/YoutubeDL.py
@@ -10,6 +10,7 @@ import io
 import itertools
 import json
 import locale
 import itertools
 import json
 import locale

+import operator

 import os
 import platform
 import re
 import os
 import platform
 import re


@@ -49,6 +50,7 @@ from .utils import (
     make_HTTPS_handler,
     MaxDownloadsReached,
     PagedList,
     make_HTTPS_handler,
     MaxDownloadsReached,
     PagedList,

+    parse_filesize,

     PostProcessingError,
     platform_name,
     preferredencoding,
     PostProcessingError,
     platform_name,
     preferredencoding,


@@ -58,6 +60,7 @@ from .utils import (
     takewhile_inclusive,
     UnavailableVideoError,
     url_basename,
     takewhile_inclusive,
     UnavailableVideoError,
     url_basename,

+    version_tuple,

     write_json_file,
     write_string,
     YoutubeDLHandler,
     write_json_file,
     write_string,
     YoutubeDLHandler,


@@ -70,6 +73,7 @@ from .extractor import get_info_extractor, gen_extractors
 from .downloader import get_suitable_downloader
 from .downloader.rtmp import rtmpdump_version
 from .postprocessor import (
 from .downloader import get_suitable_downloader
 from .downloader.rtmp import rtmpdump_version
 from .postprocessor import (

+    FFmpegFixupStretchedPP,

     FFmpegMergerPP,
     FFmpegPostProcessor,
     get_postprocessor,
     FFmpegMergerPP,
     FFmpegPostProcessor,
     get_postprocessor,


@@ -203,6 +207,17 @@ class YoutubeDL(object):
 
                        Progress hooks are guaranteed to be called at least once
                        (with status "finished") if the download is successful.
 
                        Progress hooks are guaranteed to be called at least once
                        (with status "finished") if the download is successful.

+    merge_output_format: Extension to use when merging formats.
+    fixup:             Automatically correct known faults of the file.
+                       One of:
+                       - "never": do nothing
+                       - "warn": only emit a warning
+                       - "detect_or_warn": check whether we can do anything
+                                           about it, warn otherwise
+    source_address:    (Experimental) Client-side IP address to bind to.
+    call_home:         Boolean, true iff we are allowed to contact the
+                       youtube-dl servers for debugging.
+    sleep_interval:    Number of seconds to sleep before each download.

 
 
     The following parameters are not used by YoutubeDL itself, they are used by
 
 
     The following parameters are not used by YoutubeDL itself, they are used by


@@ -756,7 +771,59 @@ class YoutubeDL(object):
         else:
             raise Exception('Invalid result type: %s' % result_type)
 
         else:
             raise Exception('Invalid result type: %s' % result_type)
 

+    def _apply_format_filter(self, format_spec, available_formats):
+        " Returns a tuple of the remaining format_spec and filtered formats "
+
+        OPERATORS = {
+            '<': operator.lt,
+            '<=': operator.le,
+            '>': operator.gt,
+            '>=': operator.ge,
+            '=': operator.eq,
+            '!=': operator.ne,
+        }
+        operator_rex = re.compile(r'''(?x)\s*\[
+            (?P<key>width|height|tbr|abr|vbr|filesize)
+            \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
+            (?P<value>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)
+            \]$
+            ''' % '|'.join(map(re.escape, OPERATORS.keys())))
+        m = operator_rex.search(format_spec)
+        if not m:
+            raise ValueError('Invalid format specification %r' % format_spec)
+
+        try:
+            comparison_value = int(m.group('value'))
+        except ValueError:
+            comparison_value = parse_filesize(m.group('value'))
+            if comparison_value is None:
+                comparison_value = parse_filesize(m.group('value') + 'B')
+            if comparison_value is None:
+                raise ValueError(
+                    'Invalid value %r in format specification %r' % (
+                        m.group('value'), format_spec))
+        op = OPERATORS[m.group('op')]
+
+        def _filter(f):
+            actual_value = f.get(m.group('key'))
+            if actual_value is None:
+                return m.group('none_inclusive')
+            return op(actual_value, comparison_value)
+        new_formats = [f for f in available_formats if _filter(f)]
+
+        new_format_spec = format_spec[:-len(m.group(0))]
+        if not new_format_spec:
+            new_format_spec = 'best'
+
+        return (new_format_spec, new_formats)
+

     def select_format(self, format_spec, available_formats):
     def select_format(self, format_spec, available_formats):

+        while format_spec.endswith(']'):
+            format_spec, available_formats = self._apply_format_filter(
+                format_spec, available_formats)
+        if not available_formats:
+            return None
+

         if format_spec == 'best' or format_spec is None:
             return available_formats[-1]
         elif format_spec == 'worst':
         if format_spec == 'best' or format_spec is None:
             return available_formats[-1]
         elif format_spec == 'worst':


@@ -909,10 +976,24 @@ class YoutubeDL(object):
                                                   'contain the video, try using '
                                                   '"-f %s+%s"' % (format_2, format_1))
                                 return
                                                   'contain the video, try using '
                                                   '"-f %s+%s"' % (format_2, format_1))
                                 return

+                            output_ext = (
+                                formats_info[0]['ext']
+                                if self.params.get('merge_output_format') is None
+                                else self.params['merge_output_format'])

                             selected_format = {
                                 'requested_formats': formats_info,
                                 'format': rf,
                             selected_format = {
                                 'requested_formats': formats_info,
                                 'format': rf,

-                                'ext': self.params['merge_output_format'] if self.params['merge_output_format'] is not None else formats_info[0]['ext'],
+                                'ext': formats_info[0]['ext'],
+                                'width': formats_info[0].get('width'),
+                                'height': formats_info[0].get('height'),
+                                'resolution': formats_info[0].get('resolution'),
+                                'fps': formats_info[0].get('fps'),
+                                'vcodec': formats_info[0].get('vcodec'),
+                                'vbr': formats_info[0].get('vbr'),
+                                'stretched_ratio': formats_info[0].get('stretched_ratio'),
+                                'acodec': formats_info[1].get('acodec'),
+                                'abr': formats_info[1].get('abr'),
+                                'ext': output_ext,

                             }
                         else:
                             selected_format = None
                             }
                         else:
                             selected_format = None


@@ -1095,51 +1176,69 @@ class YoutubeDL(object):
                                             (info_dict['thumbnail'], compat_str(err)))
 
         if not self.params.get('skip_download', False):
                                             (info_dict['thumbnail'], compat_str(err)))
 
         if not self.params.get('skip_download', False):

-            if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(filename)):
-                success = True
-            else:
-                try:
-                    def dl(name, info):
-                        fd = get_suitable_downloader(info)(self, self.params)
-                        for ph in self._progress_hooks:
-                            fd.add_progress_hook(ph)
-                        if self.params.get('verbose'):
-                            self.to_stdout('[debug] Invoking downloader on %r' % info.get('url'))
-                        return fd.download(name, info)
-                    if info_dict.get('requested_formats') is not None:
-                        downloaded = []
-                        success = True
-                        merger = FFmpegMergerPP(self, not self.params.get('keepvideo'))
-                        if not merger._executable:
-                            postprocessors = []
-                            self.report_warning('You have requested multiple '
-                                                'formats but ffmpeg or avconv are not installed.'
-                                                ' The formats won\'t be merged')
-                        else:
-                            postprocessors = [merger]
-                        for f in info_dict['requested_formats']:
-                            new_info = dict(info_dict)
-                            new_info.update(f)
-                            fname = self.prepare_filename(new_info)
-                            fname = prepend_extension(fname, 'f%s' % f['format_id'])
-                            downloaded.append(fname)
-                            partial_success = dl(fname, new_info)
-                            success = success and partial_success
-                        info_dict['__postprocessors'] = postprocessors
-                        info_dict['__files_to_merge'] = downloaded
+            try:
+                def dl(name, info):
+                    fd = get_suitable_downloader(info)(self, self.params)
+                    for ph in self._progress_hooks:
+                        fd.add_progress_hook(ph)
+                    if self.params.get('verbose'):
+                        self.to_stdout('[debug] Invoking downloader on %r' % info.get('url'))
+                    return fd.download(name, info)
+                if info_dict.get('requested_formats') is not None:
+                    downloaded = []
+                    success = True
+                    merger = FFmpegMergerPP(self, not self.params.get('keepvideo'))
+                    if not merger._executable:
+                        postprocessors = []
+                        self.report_warning('You have requested multiple '
+                                            'formats but ffmpeg or avconv are not installed.'
+                                            ' The formats won\'t be merged')

                     else:
                     else:

-                        # Just a single file
-                        success = dl(filename, info_dict)
-                except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
-                    self.report_error('unable to download video data: %s' % str(err))
-                    return
-                except (OSError, IOError) as err:
-                    raise UnavailableVideoError(err)
-                except (ContentTooShortError, ) as err:
-                    self.report_error('content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
-                    return
+                        postprocessors = [merger]
+                    for f in info_dict['requested_formats']:
+                        new_info = dict(info_dict)
+                        new_info.update(f)
+                        fname = self.prepare_filename(new_info)
+                        fname = prepend_extension(fname, 'f%s' % f['format_id'])
+                        downloaded.append(fname)
+                        partial_success = dl(fname, new_info)
+                        success = success and partial_success
+                    info_dict['__postprocessors'] = postprocessors
+                    info_dict['__files_to_merge'] = downloaded
+                else:
+                    # Just a single file
+                    success = dl(filename, info_dict)
+            except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
+                self.report_error('unable to download video data: %s' % str(err))
+                return
+            except (OSError, IOError) as err:
+                raise UnavailableVideoError(err)
+            except (ContentTooShortError, ) as err:
+                self.report_error('content too short (expected %s bytes and served %s)' % (err.expected, err.downloaded))
+                return

 
             if success:
 
             if success:

+                # Fixup content
+                stretched_ratio = info_dict.get('stretched_ratio')
+                if stretched_ratio is not None and stretched_ratio != 1:
+                    fixup_policy = self.params.get('fixup')
+                    if fixup_policy is None:
+                        fixup_policy = 'detect_or_warn'
+                    if fixup_policy == 'warn':
+                        self.report_warning('%s: Non-uniform pixel ratio (%s)' % (
+                            info_dict['id'], stretched_ratio))
+                    elif fixup_policy == 'detect_or_warn':
+                        stretched_pp = FFmpegFixupStretchedPP(self)
+                        if stretched_pp.available:
+                            info_dict.setdefault('__postprocessors', [])
+                            info_dict['__postprocessors'].append(stretched_pp)
+                        else:
+                            self.report_warning(
+                                '%s: Non-uniform pixel ratio (%s). Install ffmpeg or avconv to fix this automatically.' % (
+                                    info_dict['id'], stretched_ratio))
+                    else:
+                        assert fixup_policy == 'ignore'
+

                 try:
                     self.post_process(filename, info_dict)
                 except (PostProcessingError) as err:
                 try:
                     self.post_process(filename, info_dict)
                 except (PostProcessingError) as err:


@@ -1188,14 +1287,15 @@ class YoutubeDL(object):
         """Run all the postprocessors on the given file."""
         info = dict(ie_info)
         info['filepath'] = filename
         """Run all the postprocessors on the given file."""
         info = dict(ie_info)
         info['filepath'] = filename

-        keep_video = None

         pps_chain = []
         if ie_info.get('__postprocessors') is not None:
             pps_chain.extend(ie_info['__postprocessors'])
         pps_chain.extend(self._pps)
         for pp in pps_chain:
         pps_chain = []
         if ie_info.get('__postprocessors') is not None:
             pps_chain.extend(ie_info['__postprocessors'])
         pps_chain.extend(self._pps)
         for pp in pps_chain:

+            keep_video = None
+            old_filename = info['filepath']

             try:
             try:

-                keep_video_wish, new_info = pp.run(info)
+                keep_video_wish, info = pp.run(info)

                 if keep_video_wish is not None:
                     if keep_video_wish:
                         keep_video = keep_video_wish
                 if keep_video_wish is not None:
                     if keep_video_wish:
                         keep_video = keep_video_wish


@@ -1204,12 +1304,12 @@ class YoutubeDL(object):
                         keep_video = keep_video_wish
             except PostProcessingError as e:
                 self.report_error(e.msg)
                         keep_video = keep_video_wish
             except PostProcessingError as e:
                 self.report_error(e.msg)

-        if keep_video is False and not self.params.get('keepvideo', False):
-            try:
-                self.to_screen('Deleting original file %s (pass -k to keep)' % filename)
-                os.remove(encodeFilename(filename))
-            except (IOError, OSError):
-                self.report_warning('Unable to remove downloaded video file')
+            if keep_video is False and not self.params.get('keepvideo', False):
+                try:
+                    self.to_screen('Deleting original file %s (pass -k to keep)' % old_filename)
+                    os.remove(encodeFilename(old_filename))
+                except (IOError, OSError):
+                    self.report_warning('Unable to remove downloaded video file')

 
     def _make_archive_id(self, info_dict):
         # Future-proof against any change in case
 
     def _make_archive_id(self, info_dict):
         # Future-proof against any change in case


@@ -1420,6 +1520,17 @@ class YoutubeDL(object):
                 proxy_map.update(handler.proxies)
         self._write_string('[debug] Proxy map: ' + compat_str(proxy_map) + '\n')
 
                 proxy_map.update(handler.proxies)
         self._write_string('[debug] Proxy map: ' + compat_str(proxy_map) + '\n')
 

+        if self.params.get('call_home', False):
+            ipaddr = self.urlopen('https://yt-dl.org/ip').read().decode('utf-8')
+            self._write_string('[debug] Public IP address: %s\n' % ipaddr)
+            latest_version = self.urlopen(
+                'https://yt-dl.org/latest/version').read().decode('utf-8')
+            if version_tuple(latest_version) > version_tuple(__version__):
+                self.report_warning(
+                    'You are using an outdated version (newest version: %s)! '
+                    'See https://yt-dl.org/update if you need help updating.' %
+                    latest_version)
+

     def _setup_opener(self):
         timeout_val = self.params.get('socket_timeout')
         self._socket_timeout = 600 if timeout_val is None else float(timeout_val)
     def _setup_opener(self):
         timeout_val = self.params.get('socket_timeout')
         self._socket_timeout = 600 if timeout_val is None else float(timeout_val)


@@ -1450,9 +1561,8 @@ class YoutubeDL(object):
         proxy_handler = compat_urllib_request.ProxyHandler(proxies)
 
         debuglevel = 1 if self.params.get('debug_printtraffic') else 0
         proxy_handler = compat_urllib_request.ProxyHandler(proxies)
 
         debuglevel = 1 if self.params.get('debug_printtraffic') else 0

-        https_handler = make_HTTPS_handler(
-            self.params.get('nocheckcertificate', False), debuglevel=debuglevel)
-        ydlh = YoutubeDLHandler(debuglevel=debuglevel)
+        https_handler = make_HTTPS_handler(self.params, debuglevel=debuglevel)
+        ydlh = YoutubeDLHandler(self.params, debuglevel=debuglevel)

         opener = compat_urllib_request.build_opener(
             https_handler, proxy_handler, cookie_processor, ydlh)
         # Delete the default user-agent header, which would otherwise apply in
         opener = compat_urllib_request.build_opener(
             https_handler, proxy_handler, cookie_processor, ydlh)
         # Delete the default user-agent header, which would otherwise apply in