tombryan commited on
Commit
f7ccb61
1 Parent(s): 12e1be7

updating for data streaming

Browse files
Files changed (1) hide show
  1. AmericanStories.py +17 -26
AmericanStories.py CHANGED
@@ -1,11 +1,5 @@
1
  import json
2
- import tarfile
3
- from datasets import DatasetInfo, DatasetBuilder, DownloadManager, BuilderConfig, SplitGenerator, Split, Version
4
  import datasets
5
- import os
6
- import requests
7
- import re
8
- from tqdm import tqdm
9
 
10
  SUPPORTED_YEARS = ["1774"]
11
  # Add years from 1798 to 1964 to the supported years
@@ -19,10 +13,11 @@ def make_year_file_splits():
19
  dict: A dictionary mapping each year to its corresponding file URL.
20
  list: A list of years.
21
  """
 
22
  base_url = "https://huggingface.co/datasets/dell-research-harvard/AmericanStories/resolve/main/"
23
 
24
  # Make a list of years from 1774 to 1960
25
- # TODO: Can we fix the year list to pull directly from the repo instead of needing the hard-coded list?
26
  year_list = SUPPORTED_YEARS
27
  data_files = [f"faro_{year}.tar.gz" for year in year_list]
28
  url_list = [base_url + file for file in data_files]
@@ -47,7 +42,7 @@ _FILE_DICT, _YEARS = make_year_file_splits()
47
  class CustomBuilderConfig(datasets.BuilderConfig):
48
  """BuilderConfig for AmericanStories dataset with different configurations."""
49
 
50
- def __init__(self, year_list=None,features=["article_id", "newspaper_name", "edition", "date", "page", "headline", "byline", "article"], **kwargs):
51
  """
52
  BuilderConfig for AmericanStories dataset.
53
 
@@ -57,8 +52,6 @@ class CustomBuilderConfig(datasets.BuilderConfig):
57
  """
58
  super(CustomBuilderConfig, self).__init__(**kwargs)
59
  self.year_list = year_list
60
- self.features = features
61
-
62
 
63
  class AmericanStories(datasets.GeneratorBasedBuilder):
64
  """Dataset builder class for AmericanStories dataset."""
@@ -76,7 +69,6 @@ class AmericanStories(datasets.GeneratorBasedBuilder):
76
  version=VERSION,
77
  description="Subset of years in the dataset",
78
  year_list=["1774", "1804"]
79
-
80
  ),
81
  CustomBuilderConfig(
82
  name="all_years_content_regions",
@@ -88,10 +80,11 @@ class AmericanStories(datasets.GeneratorBasedBuilder):
88
  version=VERSION,
89
  description="Subset of years in the dataset",
90
  year_list=["1774", "1804"],
91
-
92
  )
93
  ]
 
94
  DEFAULT_CONFIG_NAME = "subset_years"
 
95
 
96
  def _info(self):
97
  """
@@ -111,8 +104,6 @@ class AmericanStories(datasets.GeneratorBasedBuilder):
111
  "headline": datasets.Value("string"),
112
  "byline": datasets.Value("string"),
113
  "article": datasets.Value("string"),
114
-
115
-
116
  }
117
  )
118
  else:
@@ -150,24 +141,25 @@ class AmericanStories(datasets.GeneratorBasedBuilder):
150
 
151
  # Subset _FILE_DICT and year_list to only include years in config.year_list
152
  if self.config.year_list:
153
- urls = {year: urls[year] for year in self.config.year_list}
154
  year_list = self.config.year_list
155
 
156
- data_dir = dl_manager.download_and_extract(urls)
157
 
158
  # Return a list of splits, where each split corresponds to a year
159
  return [
160
  datasets.SplitGenerator(
161
  name=year,
162
  gen_kwargs={
163
- "year_dir": os.path.join(data_dir[year], "mnt", "122a7683-fa4b-45dd-9f13-b18cc4f4a187", "ca_rule_based_fa_clean", "faro_" + year),
 
164
  "split": year,
165
  "associated": True if not self.config.name.endswith("content_regions") else False,
166
  },
167
  ) for year in year_list
168
  ]
169
 
170
- def _generate_examples(self, year_dir,split, associated):
171
  """
172
  Generates examples for the specified year and split.
173
 
@@ -180,10 +172,10 @@ class AmericanStories(datasets.GeneratorBasedBuilder):
180
  """
181
  print("Associated: " + str(associated))
182
  if associated:
183
- for filepath in os.listdir(year_dir):
184
- with open(os.path.join(year_dir, filepath), encoding="utf-8") as f:
185
  try :
186
- data = json.load(f)
187
  except:
188
  print("Error loading file: " + filepath)
189
  continue
@@ -208,20 +200,19 @@ class AmericanStories(datasets.GeneratorBasedBuilder):
208
  }
209
  else:
210
  print("Returning a json as a string, feel free to parse it yourself!")
211
- for filepath in os.listdir(year_dir):
212
- with open(os.path.join(year_dir, filepath), encoding="utf-8") as f:
213
  try :
214
- data = json.load(f)
215
  except:
216
  # print("Error loading file: " + filepath)
217
  continue
218
  ###Convert json to strng
219
  data=json.dumps(data)
220
- print((data))
221
- print(type(data))
222
  scan_id=filepath.split('.')[0]
223
  ##Yield the scan id and the raw data string
224
  yield scan_id, {
225
  "raw_data_string": str(data)
226
  }
227
 
 
 
1
  import json
 
 
2
  import datasets
 
 
 
 
3
 
4
  SUPPORTED_YEARS = ["1774"]
5
  # Add years from 1798 to 1964 to the supported years
 
13
  dict: A dictionary mapping each year to its corresponding file URL.
14
  list: A list of years.
15
  """
16
+
17
  base_url = "https://huggingface.co/datasets/dell-research-harvard/AmericanStories/resolve/main/"
18
 
19
  # Make a list of years from 1774 to 1960
20
+ # TODO: can we set up to check the actual files in the repo instead of relying on the offline list?
21
  year_list = SUPPORTED_YEARS
22
  data_files = [f"faro_{year}.tar.gz" for year in year_list]
23
  url_list = [base_url + file for file in data_files]
 
42
  class CustomBuilderConfig(datasets.BuilderConfig):
43
  """BuilderConfig for AmericanStories dataset with different configurations."""
44
 
45
+ def __init__(self, year_list=None, **kwargs):
46
  """
47
  BuilderConfig for AmericanStories dataset.
48
 
 
52
  """
53
  super(CustomBuilderConfig, self).__init__(**kwargs)
54
  self.year_list = year_list
 
 
55
 
56
  class AmericanStories(datasets.GeneratorBasedBuilder):
57
  """Dataset builder class for AmericanStories dataset."""
 
69
  version=VERSION,
70
  description="Subset of years in the dataset",
71
  year_list=["1774", "1804"]
 
72
  ),
73
  CustomBuilderConfig(
74
  name="all_years_content_regions",
 
80
  version=VERSION,
81
  description="Subset of years in the dataset",
82
  year_list=["1774", "1804"],
 
83
  )
84
  ]
85
+
86
  DEFAULT_CONFIG_NAME = "subset_years"
87
+ BUILDER_CONFIG_CLASS = CustomBuilderConfig
88
 
89
  def _info(self):
90
  """
 
104
  "headline": datasets.Value("string"),
105
  "byline": datasets.Value("string"),
106
  "article": datasets.Value("string"),
 
 
107
  }
108
  )
109
  else:
 
141
 
142
  # Subset _FILE_DICT and year_list to only include years in config.year_list
143
  if self.config.year_list:
144
+ urls = {year: urls[year] for year in self.config.year_list if year in SUPPORTED_YEARS}
145
  year_list = self.config.year_list
146
 
147
+ archive = dl_manager.download(urls)
148
 
149
  # Return a list of splits, where each split corresponds to a year
150
  return [
151
  datasets.SplitGenerator(
152
  name=year,
153
  gen_kwargs={
154
+ "files": dl_manager.iter_archive(archive[year]),
155
+ "year_dir": "/".join(["mnt", "122a7683-fa4b-45dd-9f13-b18cc4f4a187", "ca_rule_based_fa_clean", "faro_" + year]),
156
  "split": year,
157
  "associated": True if not self.config.name.endswith("content_regions") else False,
158
  },
159
  ) for year in year_list
160
  ]
161
 
162
+ def _generate_examples(self, files, year_dir, split, associated):
163
  """
164
  Generates examples for the specified year and split.
165
 
 
172
  """
173
  print("Associated: " + str(associated))
174
  if associated:
175
+ for filepath, f in files:
176
+ if filepath.startswith(year_dir):
177
  try :
178
+ data = json.loads(f.read().decode('utf-8'))
179
  except:
180
  print("Error loading file: " + filepath)
181
  continue
 
200
  }
201
  else:
202
  print("Returning a json as a string, feel free to parse it yourself!")
203
+ for filepath, f in files:
204
+ if filepath.startswith(year_dir):
205
  try :
206
+ data = json.loads(f.read().decode('utf-8'))
207
  except:
208
  # print("Error loading file: " + filepath)
209
  continue
210
  ###Convert json to strng
211
  data=json.dumps(data)
 
 
212
  scan_id=filepath.split('.')[0]
213
  ##Yield the scan id and the raw data string
214
  yield scan_id, {
215
  "raw_data_string": str(data)
216
  }
217
 
218
+