Datasets:

Modalities:
Tabular
Text
Formats:
parquet
Languages:
code
ArXiv:
Libraries:
Datasets
Dask
License:
denisko commited on
Commit
46d20d8
1 Parent(s): 8653b08

added example, cleaned content

Browse files
Files changed (1) hide show
  1. README.md +69 -3
README.md CHANGED
@@ -39,11 +39,9 @@ extra_gated_fields:
39
  - [Changelog](#changelog)
40
  - [Dataset Summary](#dataset-summary)
41
  - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
42
- - [Languages](#languages)
43
  - [Dataset Structure](#dataset-structure)
44
- - [Data Instances](#data-instances)
45
  - [Data Fields](#data-fields)
46
- - [Data Splits](#data-splits)
47
  - [Dataset Creation](#dataset-creation)
48
  - [Considerations for Using the Data](#considerations-for-using-the-data)
49
  - [Additional Information](#additional-information)
@@ -85,6 +83,74 @@ As na example of aggregation operation on The Stack, the image above shows conce
85
 
86
  The meta data will allow you to reconstruct repository directory structures. For this, for each repository form `ri` tabele it is needed to take all its files from `fi` table, find them in The Stack by file's `hexsha` and save those files' content under its path for a repository from `fi` table. For speed it is preferable to index The Stack by hexsha first.
87
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
88
 
89
  ## Dataset Creation
90
 
 
39
  - [Changelog](#changelog)
40
  - [Dataset Summary](#dataset-summary)
41
  - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
 
42
  - [Dataset Structure](#dataset-structure)
 
43
  - [Data Fields](#data-fields)
44
+ - [Usage Example](#usage-example)
45
  - [Dataset Creation](#dataset-creation)
46
  - [Considerations for Using the Data](#considerations-for-using-the-data)
47
  - [Additional Information](#additional-information)
 
83
 
84
  The meta data will allow you to reconstruct repository directory structures. For this, for each repository form `ri` tabele it is needed to take all its files from `fi` table, find them in The Stack by file's `hexsha` and save those files' content under its path for a repository from `fi` table. For speed it is preferable to index The Stack by hexsha first.
85
 
86
+ ### Usage Example
87
+
88
+ Restore folder structure for python files in numpy repository
89
+ ```
90
+ import datasets
91
+ from pathlib import Path
92
+ from tqdm.auto import tqdm
93
+ import pandas as pd
94
+
95
+ # assuming metadata is cloned into the local folder /data/hf_repos/the-stack-metadata
96
+ # the stack is cloned into the local folder /data/hf_repos/the-stack-v1.1
97
+ # destination folder is in /repo_workdir/numpy_restored
98
+ the_stack_meta_path = Path('/data/hf_repos/the-stack-metadata')
99
+ the_stack_path = Path('/data/hf_repos/the-stack-v1.1')
100
+ repo_dst_root = Path('/repo_workdir/numpy_restored')
101
+ repo_name = 'numpy/numpy'
102
+
103
+ # Get bucket with numpy repo info
104
+ # meta_bucket_path = None
105
+ #for fn in tqdm(list((the_stack_meta_path/'data').glob('*/ri.parquet'))):
106
+ # df = pd.read_parquet(fn)
107
+ # if any(df['name'] == repo_name):
108
+ # meta_bucket_path = fn
109
+ # break
110
+ meta_bucket_path = the_stack_meta_path / 'data/255_944'
111
+
112
+
113
+ # Get repository id from repo name
114
+ ri_id = pd.read_parquet(
115
+ meta_bucket_path / 'ri.parquet'
116
+ ).query(
117
+ f'`name` == "{repo_name}"'
118
+ )['id'].to_list()[0]
119
+
120
+ # Get files information for the reopository
121
+ files_info = pd.read_parquet(
122
+ meta_bucket_path / 'fi.parquet'
123
+ ).query(
124
+ f'`ri_id` == {ri_id} and `size` != 0 and `is_deleted` == False'
125
+ )
126
+
127
+ # Convert DF with files information to a dictionary by language and then file hexsha
128
+ # there can be more than one file with the same hexsha in the repo so we gather
129
+ # all instances per unique hexsha
130
+ files_info_dict = {
131
+ k: v[['hexsha', 'path']].groupby('hexsha').apply(lambda x: list(x['path'])).to_dict()
132
+ for k, v in files_info.groupby('lang_ex')
133
+ }
134
+
135
+ # Load Python part of The Stack
136
+ ds = datasets.load_dataset(
137
+ str(the_stack_path/'data/python'),
138
+ num_proc=10, ignore_verifications=True
139
+ )
140
+
141
+ # Save file content of the python files in the numpy reposirotry in their appropriate locations
142
+ def save_file_content(example, files_info_dict, repo_dst_root):
143
+ if example['hexsha'] in files_info_dict:
144
+ for el in files_info_dict[example['hexsha']]:
145
+ path = repo_dst_root / el
146
+ path.parent.mkdir(parents=True, exist_ok=True)
147
+ path.write_text(example['content'])
148
+ ds.map(
149
+ save_file_content,
150
+ fn_kwargs={'files_info_dict': files_info_dict['Python'], 'repo_dst_root': repo_dst_root},
151
+ num_proc=10
152
+ )
153
+ ```
154
 
155
  ## Dataset Creation
156