]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - test/test_download.py
Update changelog.
[youtubedl] / test / test_download.py
index 067bde4bb9159f4d56ae34bbc4690a8425a11550..ebe820dfc1990e4df6758795345375375402900b 100644 (file)
 #!/usr/bin/env python
 
-import errno
+from __future__ import unicode_literals
+
+# Allow direct execution
+import os
+import sys
+import unittest
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from test.helper import (
+    assertGreaterEqual,
+    expect_warnings,
+    get_params,
+    gettestcases,
+    expect_info_dict,
+    try_rm,
+    report_warning,
+)
+
+
 import hashlib
 import io
-import os
 import json
-import unittest
-import sys
 import socket
-import binascii
-
-# Allow direct execution
-sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
 import youtube_dl.YoutubeDL
-import youtube_dl.extractor
-from youtube_dl.utils import *
-
-DEF_FILE = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'tests.json')
-PARAMETERS_FILE = os.path.join(os.path.dirname(os.path.abspath(__file__)), "parameters.json")
+from youtube_dl.compat import (
+    compat_http_client,
+    compat_urllib_error,
+    compat_HTTPError,
+)
+from youtube_dl.utils import (
+    DownloadError,
+    ExtractorError,
+    format_bytes,
+    UnavailableVideoError,
+)
+from youtube_dl.extractor import get_info_extractor
 
 RETRIES = 3
 
-# General configuration (from __init__, not very elegant...)
-jar = compat_cookiejar.CookieJar()
-cookie_processor = compat_urllib_request.HTTPCookieProcessor(jar)
-proxy_handler = compat_urllib_request.ProxyHandler()
-opener = compat_urllib_request.build_opener(proxy_handler, cookie_processor, YoutubeDLHandler())
-compat_urllib_request.install_opener(opener)
-socket.setdefaulttimeout(10)
-
-def _try_rm(filename):
-    """ Remove a file if it exists """
-    try:
-        os.remove(filename)
-    except OSError as ose:
-        if ose.errno != errno.ENOENT:
-            raise
-
-md5 = lambda s: hashlib.md5(s.encode('utf-8')).hexdigest()
 
 class YoutubeDL(youtube_dl.YoutubeDL):
     def __init__(self, *args, **kwargs):
         self.to_stderr = self.to_screen
         self.processed_info_dicts = []
         super(YoutubeDL, self).__init__(*args, **kwargs)
+
     def report_warning(self, message):
         # Don't accept warnings during tests
         raise ExtractorError(message)
+
     def process_info(self, info_dict):
         self.processed_info_dicts.append(info_dict)
         return super(YoutubeDL, self).process_info(info_dict)
 
+
 def _file_md5(fn):
     with open(fn, 'rb') as f:
         return hashlib.md5(f.read()).hexdigest()
 
-with io.open(DEF_FILE, encoding='utf-8') as deff:
-    defs = json.load(deff)
-with io.open(PARAMETERS_FILE, encoding='utf-8') as pf:
-    parameters = json.load(pf)
+
+defs = gettestcases()
 
 
 class TestDownload(unittest.TestCase):
+    # Parallel testing in nosetests. See
+    # http://nose.readthedocs.org/en/latest/doc_tests/test_multiprocess/multiprocess.html
+    _multiprocess_shared_ = True
+
     maxDiff = None
+
+    def __str__(self):
+        """Identify each test with the `add_ie` attribute, if available."""
+
+        def strclass(cls):
+            """From 2.7's unittest; 2.6 had _strclass so we can't import it."""
+            return '%s.%s' % (cls.__module__, cls.__name__)
+
+        add_ie = getattr(self, self._testMethodName).add_ie
+        return '%s (%s)%s:' % (self._testMethodName,
+                               strclass(self.__class__),
+                               ' [%s]' % add_ie if add_ie else '')
+
     def setUp(self):
-        self.parameters = parameters
         self.defs = defs
 
-### Dynamically generate tests
-def generator(test_case):
+# Dynamically generate tests
+
+
+def generator(test_case, tname):
 
     def test_template(self):
-        ie = youtube_dl.extractor.get_info_extractor(test_case['name'])
-        if not ie._WORKING:
-            print('Skipping: IE marked as not _WORKING')
-            return
-        if 'playlist' not in test_case and not test_case['file']:
-            print('Skipping: No output file specified')
+        ie = youtube_dl.extractor.get_info_extractor(test_case['name'])()
+        other_ies = [get_info_extractor(ie_key)() for ie_key in test_case.get('add_ie', [])]
+        is_playlist = any(k.startswith('playlist') for k in test_case)
+        test_cases = test_case.get(
+            'playlist', [] if is_playlist else [test_case])
+
+        def print_skipping(reason):
+            print('Skipping %s: %s' % (test_case['name'], reason))
+        if not ie.working():
+            print_skipping('IE marked as not _WORKING')
             return
+
+        for tc in test_cases:
+            info_dict = tc.get('info_dict', {})
+            if not (info_dict.get('id') and info_dict.get('ext')):
+                raise Exception('Test definition incorrect. The output file cannot be known. Are both \'id\' and \'ext\' keys present?')
+
         if 'skip' in test_case:
-            print('Skipping: {0}'.format(test_case['skip']))
+            print_skipping(test_case['skip'])
             return
+        for other_ie in other_ies:
+            if not other_ie.working():
+                print_skipping('test depends on %sIE, marked as not WORKING' % other_ie.ie_key())
+                return
 
-        params = self.parameters.copy()
-        params.update(test_case.get('params', {}))
+        params = get_params(test_case.get('params', {}))
+        params['outtmpl'] = tname + '_' + params['outtmpl']
+        if is_playlist and 'playlist' not in test_case:
+            params.setdefault('extract_flat', 'in_playlist')
+            params.setdefault('skip_download', True)
 
-        ydl = YoutubeDL(params)
-        for ie in youtube_dl.extractor.gen_extractors():
-            ydl.add_info_extractor(ie)
+        ydl = YoutubeDL(params, auto_init=False)
+        ydl.add_default_info_extractors()
         finished_hook_called = set()
+
         def _hook(status):
             if status['status'] == 'finished':
                 finished_hook_called.add(status['filename'])
-        ydl.fd.add_progress_hook(_hook)
+        ydl.add_progress_hook(_hook)
+        expect_warnings(ydl, test_case.get('expected_warnings', []))
 
-        test_cases = test_case.get('playlist', [test_case])
-        for tc in test_cases:
-            _try_rm(tc['file'])
-            _try_rm(tc['file'] + '.part')
-            _try_rm(tc['file'] + '.info.json')
+        def get_tc_filename(tc):
+            return ydl.prepare_filename(tc.get('info_dict', {}))
+
+        res_dict = None
+
+        def try_rm_tcs_files(tcs=None):
+            if tcs is None:
+                tcs = test_cases
+            for tc in tcs:
+                tc_filename = get_tc_filename(tc)
+                try_rm(tc_filename)
+                try_rm(tc_filename + '.part')
+                try_rm(os.path.splitext(tc_filename)[0] + '.info.json')
+        try_rm_tcs_files()
         try:
-            for retry in range(1, RETRIES + 1):
+            try_num = 1
+            while True:
                 try:
-                    ydl.download([test_case['url']])
+                    # We're not using .download here since that is just a shim
+                    # for outside error handling, and returns the exit code
+                    # instead of the result dict.
+                    res_dict = ydl.extract_info(
+                        test_case['url'],
+                        force_generic_extractor=params.get('force_generic_extractor', False))
                 except (DownloadError, ExtractorError) as err:
-                    if retry == RETRIES: raise
-
                     # Check if the exception is not a network related one
-                    if not err.exc_info[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
+                    if not err.exc_info[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError, compat_http_client.BadStatusLine) or (err.exc_info[0] == compat_HTTPError and err.exc_info[1].code == 503):
                         raise
 
-                    print('Retrying: {0} failed tries\n\n##########\n\n'.format(retry))
+                    if try_num == RETRIES:
+                        report_warning('%s failed due to network errors, skipping...' % tname)
+                        return
+
+                    print('Retrying: {0} failed tries\n\n##########\n\n'.format(try_num))
+
+                    try_num += 1
                 else:
                     break
 
-            for tc in test_cases:
+            if is_playlist:
+                self.assertTrue(res_dict['_type'] in ['playlist', 'multi_video'])
+                self.assertTrue('entries' in res_dict)
+                expect_info_dict(self, res_dict, test_case.get('info_dict', {}))
+
+            if 'playlist_mincount' in test_case:
+                assertGreaterEqual(
+                    self,
+                    len(res_dict['entries']),
+                    test_case['playlist_mincount'],
+                    'Expected at least %d in playlist %s, but got only %d' % (
+                        test_case['playlist_mincount'], test_case['url'],
+                        len(res_dict['entries'])))
+            if 'playlist_count' in test_case:
+                self.assertEqual(
+                    len(res_dict['entries']),
+                    test_case['playlist_count'],
+                    'Expected %d entries in playlist %s, but got %d.' % (
+                        test_case['playlist_count'],
+                        test_case['url'],
+                        len(res_dict['entries']),
+                    ))
+            if 'playlist_duration_sum' in test_case:
+                got_duration = sum(e['duration'] for e in res_dict['entries'])
+                self.assertEqual(
+                    test_case['playlist_duration_sum'], got_duration)
+
+            # Generalize both playlists and single videos to unified format for
+            # simplicity
+            if 'entries' not in res_dict:
+                res_dict['entries'] = [res_dict]
+
+            for tc_num, tc in enumerate(test_cases):
+                tc_res_dict = res_dict['entries'][tc_num]
+                # First, check test cases' data against extracted data alone
+                expect_info_dict(self, tc_res_dict, tc.get('info_dict', {}))
+                # Now, check downloaded file consistency
+                tc_filename = get_tc_filename(tc)
                 if not test_case.get('params', {}).get('skip_download', False):
-                    self.assertTrue(os.path.exists(tc['file']), msg='Missing file ' + tc['file'])
-                    self.assertTrue(tc['file'] in finished_hook_called)
-                self.assertTrue(os.path.exists(tc['file'] + '.info.json'))
-                if 'md5' in tc:
-                    md5_for_file = _file_md5(tc['file'])
-                    self.assertEqual(md5_for_file, tc['md5'])
-                with io.open(tc['file'] + '.info.json', encoding='utf-8') as infof:
+                    self.assertTrue(os.path.exists(tc_filename), msg='Missing file ' + tc_filename)
+                    self.assertTrue(tc_filename in finished_hook_called)
+                    expected_minsize = tc.get('file_minsize', 10000)
+                    if expected_minsize is not None:
+                        if params.get('test'):
+                            expected_minsize = max(expected_minsize, 10000)
+                        got_fsize = os.path.getsize(tc_filename)
+                        assertGreaterEqual(
+                            self, got_fsize, expected_minsize,
+                            'Expected %s to be at least %s, but it\'s only %s ' %
+                            (tc_filename, format_bytes(expected_minsize),
+                                format_bytes(got_fsize)))
+                    if 'md5' in tc:
+                        md5_for_file = _file_md5(tc_filename)
+                        self.assertEqual(tc['md5'], md5_for_file)
+                # Finally, check test cases' data again but this time against
+                # extracted data from info JSON file written during processing
+                info_json_fn = os.path.splitext(tc_filename)[0] + '.info.json'
+                self.assertTrue(
+                    os.path.exists(info_json_fn),
+                    'Missing info file %s' % info_json_fn)
+                with io.open(info_json_fn, encoding='utf-8') as infof:
                     info_dict = json.load(infof)
-                for (info_field, expected) in tc.get('info_dict', {}).items():
-                    if isinstance(expected, compat_str) and expected.startswith('md5:'):
-                        self.assertEqual(expected, 'md5:' + md5(info_dict.get(info_field)))
-                    else:
-                        got = info_dict.get(info_field)
-                        self.assertEqual(
-                            expected, got,
-                            u'invalid value for field %s, expected %r, got %r' % (info_field, expected, got))
-
-                # If checkable fields are missing from the test case, print the info_dict
-                test_info_dict = dict((key, value if not isinstance(value, compat_str) or len(value) < 250 else 'md5:' + md5(value))
-                    for key, value in info_dict.items()
-                    if value and key in ('title', 'description', 'uploader', 'upload_date', 'uploader_id', 'location'))
-                if not all(key in tc.get('info_dict', {}).keys() for key in test_info_dict.keys()):
-                    sys.stderr.write(u'\n"info_dict": ' + json.dumps(test_info_dict, ensure_ascii=False, indent=2) + u'\n')
-
-                # Check for the presence of mandatory fields
-                for key in ('id', 'url', 'title', 'ext'):
-                    self.assertTrue(key in info_dict.keys() and info_dict[key])
+                expect_info_dict(self, info_dict, tc.get('info_dict', {}))
         finally:
-            for tc in test_cases:
-                _try_rm(tc['file'])
-                _try_rm(tc['file'] + '.part')
-                _try_rm(tc['file'] + '.info.json')
+            try_rm_tcs_files()
+            if is_playlist and res_dict is not None and res_dict.get('entries'):
+                # Remove all other files that may have been extracted if the
+                # extractor returns full results even with extract_flat
+                res_tcs = [{'info_dict': e} for e in res_dict['entries']]
+                try_rm_tcs_files(res_tcs)
 
     return test_template
 
-### And add them to TestDownload
+
+# And add them to TestDownload
 for n, test_case in enumerate(defs):
-    test_method = generator(test_case)
-    test_method.__name__ = "test_{0}".format(test_case["name"])
-    if getattr(TestDownload, test_method.__name__, False):
-        test_method.__name__ = "test_{0}_{1}".format(test_case["name"], n)
+    tname = 'test_' + str(test_case['name'])
+    i = 1
+    while hasattr(TestDownload, tname):
+        tname = 'test_%s_%d' % (test_case['name'], i)
+        i += 1
+    test_method = generator(test_case, tname)
+    test_method.__name__ = str(tname)
+    ie_list = test_case.get('add_ie')
+    test_method.add_ie = ie_list and ','.join(ie_list)
     setattr(TestDownload, test_method.__name__, test_method)
     del test_method