Leyo commited on
Commit
b78a270
·
1 Parent(s): 3160afe

chop videos in chunks

Browse files
Files changed (1) hide show
  1. yttemporal180m.py +55 -33
yttemporal180m.py CHANGED
@@ -20,15 +20,16 @@ _URL_BASE = "https://rowanzellers.com/merlot/#data"
20
 
21
  url_numbers = ["00" + str(i) if i < 10 else "0" + str(i) for i in range(100)]
22
  _DL_URLS = [
23
- f"https://storage.googleapis.com/merlot/yttemporal180m/yttemporal180m_{num}of100.jsonl.gz" for num in url_numbers]
 
 
24
 
25
 
26
  def json_serializer(o):
27
  if isinstance(o, datetime):
28
  return str(o)
29
 
30
- raise TypeError(
31
- f"Object of type {o.__class__.__name__} is not JSON serializable")
32
 
33
 
34
  class yttemporal180mConfig(datasets.BuilderConfig):
@@ -36,7 +37,8 @@ class yttemporal180mConfig(datasets.BuilderConfig):
36
 
37
  def __init__(self, **kwargs):
38
  super(yttemporal180mConfig, self).__init__(
39
- version=datasets.Version("2.1.0", ""), **kwargs)
 
40
 
41
 
42
  class yttemporal180m(datasets.GeneratorBasedBuilder):
@@ -44,7 +46,8 @@ class yttemporal180m(datasets.GeneratorBasedBuilder):
44
  DEFAULT_CONFIG_NAME = "default"
45
  BUILDER_CONFIGS = [
46
  yttemporal180mConfig(
47
- name="default", description="Default full yttemporal180m dataset"),
 
48
  ]
49
 
50
  def _info(self):
@@ -54,8 +57,9 @@ class yttemporal180m(datasets.GeneratorBasedBuilder):
54
  {
55
  "video_id": datasets.Value("string"),
56
  "video_url": datasets.Value("string"),
57
- "asr": datasets.Value("string"),
58
- "title": datasets.Value("string"),
 
59
  "meta": datasets.Value("string"),
60
  }
61
  ),
@@ -65,15 +69,12 @@ class yttemporal180m(datasets.GeneratorBasedBuilder):
65
  )
66
 
67
  def _split_generators(self, dl_manager):
68
- archive_paths = [dl_manager.download_and_extract(
69
- url) for url in _DL_URLS]
70
 
71
  train_split = [
72
  datasets.SplitGenerator(
73
  name=datasets.Split.TRAIN,
74
- gen_kwargs={
75
- "jsonl_files": archive_paths
76
- },
77
  )
78
  ]
79
 
@@ -88,26 +89,47 @@ class yttemporal180m(datasets.GeneratorBasedBuilder):
88
  for json_str in json_list:
89
  infos = json.loads(json_str)
90
 
91
- id = infos['info']['display_id']
92
  url = "https://www.youtube.com/watch?v=" + id
93
- asr = ""
94
- for example in infos['denoised']:
95
- asr += example["cleanasr"]
96
-
97
- metadata_dict = {
98
- "asr_info": infos["denoised"],
99
- "info": infos["info"],
100
- "subtitles": infos["subtitles"],
101
- }
102
- yield idx, {
103
- "video_id": id,
104
- "video_url": url,
105
- "asr": asr,
106
- "title": infos['info']['title'],
107
- "meta": json.dumps(
108
- metadata_dict,
109
- default=json_serializer,
110
- indent=2
111
  )
112
- }
113
- idx += 1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
 
21
  url_numbers = ["00" + str(i) if i < 10 else "0" + str(i) for i in range(100)]
22
  _DL_URLS = [
23
+ f"https://storage.googleapis.com/merlot/yttemporal180m/yttemporal180m_{num}of100.jsonl.gz"
24
+ for num in url_numbers
25
+ ]
26
 
27
 
28
  def json_serializer(o):
29
  if isinstance(o, datetime):
30
  return str(o)
31
 
32
+ raise TypeError(f"Object of type {o.__class__.__name__} is not JSON serializable")
 
33
 
34
 
35
  class yttemporal180mConfig(datasets.BuilderConfig):
 
37
 
38
  def __init__(self, **kwargs):
39
  super(yttemporal180mConfig, self).__init__(
40
+ version=datasets.Version("2.1.0", ""), **kwargs
41
+ )
42
 
43
 
44
  class yttemporal180m(datasets.GeneratorBasedBuilder):
 
46
  DEFAULT_CONFIG_NAME = "default"
47
  BUILDER_CONFIGS = [
48
  yttemporal180mConfig(
49
+ name="default", description="Default full yttemporal180m dataset"
50
+ ),
51
  ]
52
 
53
  def _info(self):
 
57
  {
58
  "video_id": datasets.Value("string"),
59
  "video_url": datasets.Value("string"),
60
+ "caption": datasets.Value("string"),
61
+ "timestamp_start": datasets.Value("float32"),
62
+ "timestamp_stop": datasets.Value("float32"),
63
  "meta": datasets.Value("string"),
64
  }
65
  ),
 
69
  )
70
 
71
  def _split_generators(self, dl_manager):
72
+ archive_paths = [dl_manager.download_and_extract(url) for url in _DL_URLS]
 
73
 
74
  train_split = [
75
  datasets.SplitGenerator(
76
  name=datasets.Split.TRAIN,
77
+ gen_kwargs={"jsonl_files": archive_paths},
 
 
78
  )
79
  ]
80
 
 
89
  for json_str in json_list:
90
  infos = json.loads(json_str)
91
 
92
+ id = infos["info"]["display_id"]
93
  url = "https://www.youtube.com/watch?v=" + id
94
+
95
+ # Divide video by segments of 15 sec
96
+ max_sec_per_segment = 15
97
+ num_chunks = (
98
+ int(
99
+ divmod(infos["subtitles"][-1]["time"], max_sec_per_segment)[
100
+ 0
101
+ ]
 
 
 
 
 
 
 
 
 
 
102
  )
103
+ + 1
104
+ )
105
+ time_chunks = [
106
+ i * max_sec_per_segment for i in range(num_chunks + 1)
107
+ ]
108
+ time_chunk_idx = 0
109
+ caption = ""
110
+ for el in infos["subtitles"]:
111
+ if el["time"] <= time_chunks[time_chunk_idx + 1]:
112
+ caption += el["word"] + " "
113
+ else:
114
+ timestamp_start = float(time_chunks[time_chunk_idx])
115
+ timestamp_stop = float(time_chunks[time_chunk_idx + 1])
116
+ time_chunk_idx += 1
117
+
118
+ metadata_dict = {
119
+ "asr_info": infos["denoised"],
120
+ "info": infos["info"],
121
+ "subtitles": infos["subtitles"],
122
+ "title": infos["info"]["title"],
123
+ }
124
+ yield idx, {
125
+ "video_id": id,
126
+ "video_url": url,
127
+ "caption": caption,
128
+ "timestamp_start": timestamp_start,
129
+ "timestamp_stop": timestamp_stop,
130
+ "meta": json.dumps(
131
+ metadata_dict, default=json_serializer, indent=2
132
+ ),
133
+ }
134
+ idx += 1
135
+ caption = ""