jblitzar commited on
Commit
0d3c6b9
·
verified ·
1 Parent(s): 9f25c11

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -58,3 +58,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
58
  *.mp4 filter=lfs diff=lfs merge=lfs -text
59
  *.webm filter=lfs diff=lfs merge=lfs -text
60
  mega_licensed_corpus_redacted.txt filter=lfs diff=lfs merge=lfs -text
 
 
 
58
  *.mp4 filter=lfs diff=lfs merge=lfs -text
59
  *.webm filter=lfs diff=lfs merge=lfs -text
60
  mega_licensed_corpus_redacted.txt filter=lfs diff=lfs merge=lfs -text
61
+ python_files_elaborated.txt filter=lfs diff=lfs merge=lfs -text
62
+ python_files_elaborated_metadata.csv filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -11,32 +11,42 @@ dataset_type: code
11
  tags:
12
  - code
13
  - python
14
- - code-generation
15
  size_categories:
16
  - 100K<n⩽1M
17
  task_categories:
18
  - text-generation
19
- task_ids:
20
- - code-completion
21
  ---
22
 
23
- # GitHub-Python
24
 
25
- A **767 MB** corpus of permissively-licensed Python code drawn from public GitHub repositories.
26
- The dataset was created to support training and evaluation of **code-completion / generation** models.
 
 
 
 
 
 
 
 
 
27
 
28
  ## Dataset at a glance
29
 
30
- | | Value |
31
- | ---------------- | --------------------------------------------- |
32
- | Files | 53,017 `.py` files |
33
- | Repositories | 16,447 |
34
- | Owners | 12,515 |
35
- | Compressed size | 732 MB (`mega_licensed_corpus_redacted.txt`) |
36
- | Vocabulary | 443,431 tokens (`custom_tokens_vocab.txt`) |
37
- | Time period | Commits 2015-01-01 |
38
- | License coverage | MIT, Apache-2.0, BSD, ISC, Unlicense |
39
- | Removed secrets | all hard-coded secrets/API keys redacted |
 
 
 
 
40
 
41
  Numbers were obtained from the final redacted corpus and companion metadata.
42
 
@@ -46,8 +56,10 @@ Numbers were obtained from the final redacted corpus and companion metadata.
46
 
47
  ```
48
  huggingface_dataset/
49
- ├─ mega_licensed_corpus_redacted.txt # concatenated code corpus
50
- ├─ python_files.txt # list of raw file URLs (1-per-line)
 
 
51
  └─ custom_tokens_vocab.txt # `<token>\t<id>` vocabulary file
52
  ```
53
 
@@ -64,6 +76,27 @@ code of one file.
64
 
65
  ---
66
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
67
  ## Collection methodology
68
 
69
  1. **Repository discovery**
 
11
  tags:
12
  - code
13
  - python
 
14
  size_categories:
15
  - 100K<n⩽1M
16
  task_categories:
17
  - text-generation
 
 
18
  ---
19
 
20
+ # GitHub-Python — Licensed & Elaborated Variants
21
 
22
+ This repository ships **two complementary Python-code corpora** extracted from
23
+ public GitHub:
24
+
25
+ - **Licensed Subset** – strictly _permissive-licensed_ files suitable for
26
+ commercial redistribution / model training (main corpus used in our
27
+ experiments).
28
+ - **Elaborated Collection** – a broader crawl that additionally contains files
29
+ under _copyleft_ or unclear licenses (GPL/AGPL/LGPL, etc.). Useful for
30
+ analysis or pre-training where license mixing is acceptable.
31
+
32
+ Both variants target **code-completion / generation** research.
33
 
34
  ## Dataset at a glance
35
 
36
+ | | **Licensed Subset** | **Elaborated Collection** |
37
+ | ------------------- | ------------------- | ------------------------- |
38
+ | Files (.py) | 53,017 | 186,066 |
39
+ | Unique repositories | 16,447 | 59,852 |
40
+ | Repository owners | 12,515 | 43,517 |
41
+ | Compressed size | 732 MB | 2.4 GB \* |
42
+ | Vocabulary (tokens) | 443,431 | 443,431 |
43
+ | License coverage | Permissive only | Mixed (perm. + copyleft) |
44
+ | Secrets redacted | | ⚠️ not guaranteed |
45
+ | Time window | 2015-01-01 | 2015-01-01 |
46
+
47
+ \* estimated – elaborated corpus is distributed as raw file list, not a single
48
+ text file.
49
+ † same tokenizer file is shared by both variants.
50
 
51
  Numbers were obtained from the final redacted corpus and companion metadata.
52
 
 
56
 
57
  ```
58
  huggingface_dataset/
59
+ ├─ mega_licensed_corpus_redacted.txt # Licensed Subset – concatenated code
60
+ ├─ python_files.txt # Licensed Subset raw file URLs
61
+ ├─ python_files_elaborated.txt # Elaborated Collection – raw file URLs
62
+ ├─ python_files_elaborated_metadata.csv # Elaborated Collection metadata
63
  └─ custom_tokens_vocab.txt # `<token>\t<id>` vocabulary file
64
  ```
65
 
 
76
 
77
  ---
78
 
79
+ ## Dataset variants
80
+
81
+ ### 1. Licensed Subset (`mega_licensed_corpus_redacted.txt`)
82
+
83
+ • 53 K permissively-licensed files (MIT/BSD/Apache/ISC/Unlicense).
84
+ • All API keys & credentials removed.
85
+ • Ready for redistribution & commercial use (respect upstream NOTICE files).
86
+
87
+ ### 2. Elaborated Collection (`python_files_elaborated.txt`)
88
+
89
+ • 186 K files from a much larger crawl.
90
+ • Contains **GPL / LGPL / AGPL and other copyleft** licenses.
91
+ • Shipped _as URL list_ + metadata CSV; you must download the files yourself
92
+ (`datasets.load_dataset` streaming, `wget`, etc.).
93
+ • **No license filtering or secret-redaction performed** – use with caution.
94
+
95
+ When first loading the dataset, decide which variant aligns with your use case
96
+ (e.g. proprietary model training → Licensed Subset only).
97
+
98
+ ---
99
+
100
  ## Collection methodology
101
 
102
  1. **Repository discovery**
create_elaborated_metadata_table.py ADDED
@@ -0,0 +1,197 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ """
3
+ Create metadata table for the elaborated GitHub Python dataset.
4
+
5
+ This script parses the python_files_elaborated.txt file containing GitHub URLs
6
+ and extracts repository metadata (owner, repo name, file path).
7
+ It generates a CSV file with this information and prints statistics.
8
+
9
+ The elaborated dataset contains more files than the licensed subset and
10
+ may include repositories with various licenses (not just permissive ones).
11
+ """
12
+
13
+ import csv
14
+ import os
15
+ import re
16
+ import pandas as pd
17
+ from collections import Counter, defaultdict
18
+ from tqdm import tqdm
19
+ from urllib.parse import urlparse
20
+
21
+ # Input and output files
22
+ ELABORATED_FILES_LIST = "python_files_elaborated.txt"
23
+ LICENSED_FILES_LIST = "python_files.txt"
24
+ OUTPUT_CSV = "python_files_elaborated_metadata.csv"
25
+
26
+ # Regular expression to parse GitHub raw URLs
27
+ # Format: https://raw.githubusercontent.com/OWNER/REPO/BRANCH/PATH
28
+ GITHUB_RAW_PATTERN = r"https://raw\.githubusercontent\.com/([^/]+)/([^/]+)/[^/]+/(.*)"
29
+
30
+
31
+ def parse_github_url(url):
32
+ """
33
+ Parse a GitHub raw URL to extract owner, repo name, and file path.
34
+
35
+ Args:
36
+ url (str): GitHub raw URL
37
+
38
+ Returns:
39
+ tuple: (owner, repo_name, file_path) or None if URL doesn't match pattern
40
+ """
41
+ match = re.match(GITHUB_RAW_PATTERN, url)
42
+ if match:
43
+ owner, repo_name, file_path = match.groups()
44
+ return owner, repo_name, file_path
45
+ return None
46
+
47
+
48
+ def create_metadata_table(file_list_path):
49
+ """
50
+ Create a metadata table from a list of GitHub URLs.
51
+
52
+ Args:
53
+ file_list_path (str): Path to file containing GitHub URLs
54
+
55
+ Returns:
56
+ list: List of dictionaries with metadata
57
+ """
58
+ metadata = []
59
+
60
+ # Read URLs from file
61
+ with open(file_list_path, "r") as f:
62
+ urls = [line.strip() for line in f if line.strip()]
63
+
64
+ print(f"Processing URLs from {file_list_path}...")
65
+
66
+ # Parse each URL and extract metadata
67
+ for url in tqdm(urls, desc="Parsing URLs"):
68
+ parsed = parse_github_url(url)
69
+ if parsed:
70
+ owner, repo_name, file_path = parsed
71
+ metadata.append({
72
+ "owner": owner,
73
+ "repo_name": repo_name,
74
+ "file_path": file_path,
75
+ "url": url
76
+ })
77
+
78
+ return metadata
79
+
80
+
81
+ def generate_statistics(metadata, dataset_name):
82
+ """
83
+ Generate and print statistics for the dataset.
84
+
85
+ Args:
86
+ metadata (list): List of dictionaries with metadata
87
+ dataset_name (str): Name of the dataset for display
88
+ """
89
+ # Count unique repositories and owners
90
+ repos = set((item["owner"], item["repo_name"]) for item in metadata)
91
+ owners = set(item["owner"] for item in metadata)
92
+
93
+ # Count files by repository
94
+ repo_counts = Counter((item["owner"], item["repo_name"]) for item in metadata)
95
+ top_repos = repo_counts.most_common(10)
96
+
97
+ # Count files by owner
98
+ owner_counts = Counter(item["owner"] for item in metadata)
99
+ top_owners = owner_counts.most_common(5)
100
+
101
+ # Count file extensions
102
+ extensions = Counter(os.path.splitext(item["file_path"])[1] for item in metadata)
103
+
104
+ # Print statistics
105
+ print(f"\n=== {dataset_name} Statistics ===")
106
+ print(f"Total files: {len(metadata)}")
107
+ print(f"Unique repositories: {len(repos)}")
108
+ print(f"Unique repository owners: {len(owners)}")
109
+
110
+ print("\nTop 10 repositories by file count:")
111
+ for (owner, repo), count in top_repos:
112
+ print(f" {owner}/{repo}: {count} files")
113
+
114
+ print("\nFile extensions:")
115
+ for ext, count in extensions.most_common():
116
+ if ext: # Skip empty extensions
117
+ print(f" {ext}: {count} files")
118
+
119
+ print("\nTop 5 repository owners:")
120
+ for owner, count in top_owners:
121
+ print(f" {owner}: {count} files")
122
+
123
+ return {
124
+ "total_files": len(metadata),
125
+ "unique_repos": len(repos),
126
+ "unique_owners": len(owners),
127
+ "top_repos": top_repos,
128
+ "top_owners": top_owners,
129
+ "extensions": extensions
130
+ }
131
+
132
+
133
+ def compare_datasets(elaborated_stats, licensed_stats):
134
+ """
135
+ Compare statistics between elaborated and licensed datasets.
136
+
137
+ Args:
138
+ elaborated_stats (dict): Statistics for elaborated dataset
139
+ licensed_stats (dict): Statistics for licensed dataset
140
+ """
141
+ print("\n=== Dataset Comparison ===")
142
+ print(f"Elaborated dataset: {elaborated_stats['total_files']} files")
143
+ print(f"Licensed dataset: {licensed_stats['total_files']} files")
144
+ print(f"Additional files in elaborated dataset: {elaborated_stats['total_files'] - licensed_stats['total_files']} files")
145
+
146
+ # Calculate percentage increase
147
+ pct_increase = ((elaborated_stats['total_files'] / licensed_stats['total_files']) - 1) * 100
148
+ print(f"Percentage increase: {pct_increase:.1f}%")
149
+
150
+ # Compare repositories
151
+ print(f"\nElaborated dataset: {elaborated_stats['unique_repos']} repositories")
152
+ print(f"Licensed dataset: {licensed_stats['unique_repos']} repositories")
153
+
154
+ # Compare owners
155
+ print(f"\nElaborated dataset: {elaborated_stats['unique_owners']} repository owners")
156
+ print(f"Licensed dataset: {licensed_stats['unique_owners']} repository owners")
157
+
158
+ # Find repositories unique to elaborated dataset
159
+ elaborated_repos = set((owner, repo) for (owner, repo), _ in elaborated_stats['top_repos'])
160
+ licensed_repos = set((owner, repo) for (owner, repo), _ in licensed_stats['top_repos'])
161
+ unique_to_elaborated = elaborated_repos - licensed_repos
162
+
163
+ if unique_to_elaborated:
164
+ print("\nTop repositories unique to elaborated dataset:")
165
+ for owner, repo in list(unique_to_elaborated)[:5]:
166
+ print(f" {owner}/{repo}")
167
+
168
+
169
+ def main():
170
+ # Process elaborated dataset
171
+ elaborated_metadata = create_metadata_table(ELABORATED_FILES_LIST)
172
+
173
+ # Save to CSV
174
+ with open(OUTPUT_CSV, "w", newline="") as f:
175
+ writer = csv.DictWriter(f, fieldnames=["owner", "repo_name", "file_path", "url"],
176
+ quoting=csv.QUOTE_MINIMAL)
177
+ writer.writeheader()
178
+ writer.writerows(elaborated_metadata)
179
+
180
+ print(f"Metadata saved to {OUTPUT_CSV}")
181
+
182
+ # Generate statistics for elaborated dataset
183
+ elaborated_stats = generate_statistics(elaborated_metadata, "Elaborated Dataset")
184
+
185
+ # Process licensed dataset for comparison
186
+ if os.path.exists(LICENSED_FILES_LIST):
187
+ licensed_metadata = create_metadata_table(LICENSED_FILES_LIST)
188
+ licensed_stats = generate_statistics(licensed_metadata, "Licensed Dataset")
189
+
190
+ # Compare datasets
191
+ compare_datasets(elaborated_stats, licensed_stats)
192
+ else:
193
+ print(f"Warning: {LICENSED_FILES_LIST} not found. Cannot compare datasets.")
194
+
195
+
196
+ if __name__ == "__main__":
197
+ main()
python_files_elaborated.txt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b20e7f10e0a17de893f1a172abce460b41b937e3ea83e2c81f8773982db38578
3
+ size 15102167
python_files_elaborated_metadata.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46359841e7cf01ebd2232ff24f310644481798ba85138db310dee8fb29417bbd
3
+ size 22895119