tamnd commited on
Commit
a74f739
·
verified ·
1 Parent(s): b5d3680

Publish shard CC-MAIN-2026-08/00107

Browse files
Files changed (3) hide show
  1. README.md +12 -12
  2. data/CC-MAIN-2026-08/00107.parquet +3 -0
  3. stats.csv +2 -1
README.md CHANGED
@@ -32,7 +32,7 @@ configs:
32
 
33
  Open Index is a large-scale web text dataset built from [Common Crawl](https://commoncrawl.org). Every page goes through a pipeline that extracts the main content from raw HTML, converts it to clean Markdown using [trafilatura](https://github.com/adbar/trafilatura), and packages the result into Parquet files with full WARC metadata preserved.
34
 
35
- The dataset currently includes crawl **CC-MAIN-2026-08** with **1,671,214 documents across 85 shards**. We plan to add more snapshots over time.
36
 
37
  Open Index is released under the **Open Data Commons Attribution License (ODC-By) v1.0**, the same license used by Common Crawl.
38
 
@@ -166,28 +166,28 @@ Empty conversions (pages where trafilatura could not extract meaningful content)
166
 
167
  ### Compression Ratios
168
 
169
- Numbers below are actual measurements summed across all 85 files of CC-MAIN-2026-08 (1,671,214 pages total), projected to the full crawl of 100,000 WARC files.
170
 
171
- | Stage | 85 files (measured) | 100,000 files (projected) | Reduction |
172
  |---|---|---|---|
173
- | Raw WARC (.warc.gz, downloaded) | ~68.9 GB | ~83 TB | — |
174
- | HTML extracted (uncompressed) | 207.7 GB | ~295 TB | — |
175
- | Packed markdown WARC (.md.warc.gz) | ~3.4 GB | ~3.7 TB | **-98.3%** vs HTML |
176
  | Final Parquet (Zstd level 19) | 2.4 GB | ~2.9 TB | **-31.8%** vs packed WARC |
177
 
178
- The big win is the HTML → Markdown step: trafilatura strips all tags, scripts, styles, navigation, and ads, keeping only the main content. This cuts 207.7 GB of uncompressed HTML down to 7.3 GB of markdown — a **98.3% reduction** — before any file-level compression is applied. Parquet with Zstd level 19 then compresses the markdown a further 67.9%.
179
 
180
- End to end: ~68.9 GB of raw gzipped WARCs becomes **2.4 GB of Parquet** — a **96.6% total reduction** — containing 1,671,214 clean markdown documents.
181
 
182
  ### Processing Times
183
 
184
- Pipeline timings across 85 shards of CC-MAIN-2026-08:
185
 
186
  ```
187
- Download (raw WARC) ████████████████████████ total 2h 23m 57s avg 1m 41s
188
  Convert (HTML → MD) ██░░░░░░░░░░░░░░░░░░░░░░ total 16m 2s avg 11s
189
- Export (Parquet) █████░░░░░░░░░░░░░░░░░░░ total 32m 54s avg 23s
190
- Publish (HuggingFace) █████████░░░░░░░░░░░░░░░ total 59m 21s avg 41s
191
  ```
192
 
193
  ### Dataset Charts
 
32
 
33
  Open Index is a large-scale web text dataset built from [Common Crawl](https://commoncrawl.org). Every page goes through a pipeline that extracts the main content from raw HTML, converts it to clean Markdown using [trafilatura](https://github.com/adbar/trafilatura), and packages the result into Parquet files with full WARC metadata preserved.
34
 
35
+ The dataset currently includes crawl **CC-MAIN-2026-08** with **1,690,691 documents across 86 shards**. We plan to add more snapshots over time.
36
 
37
  Open Index is released under the **Open Data Commons Attribution License (ODC-By) v1.0**, the same license used by Common Crawl.
38
 
 
166
 
167
  ### Compression Ratios
168
 
169
+ Numbers below are actual measurements summed across all 86 files of CC-MAIN-2026-08 (1,690,691 pages total), projected to the full crawl of 100,000 WARC files.
170
 
171
+ | Stage | 86 files (measured) | 100,000 files (projected) | Reduction |
172
  |---|---|---|---|
173
+ | Raw WARC (.warc.gz, downloaded) | ~69.7 GB | ~83 TB | — |
174
+ | HTML extracted (uncompressed) | 210.1 GB | ~295 TB | — |
175
+ | Packed markdown WARC (.md.warc.gz) | ~3.5 GB | ~3.7 TB | **-98.3%** vs HTML |
176
  | Final Parquet (Zstd level 19) | 2.4 GB | ~2.9 TB | **-31.8%** vs packed WARC |
177
 
178
+ The big win is the HTML → Markdown step: trafilatura strips all tags, scripts, styles, navigation, and ads, keeping only the main content. This cuts 210.1 GB of uncompressed HTML down to 7.4 GB of markdown — a **98.3% reduction** — before any file-level compression is applied. Parquet with Zstd level 19 then compresses the markdown a further 67.9%.
179
 
180
+ End to end: ~69.7 GB of raw gzipped WARCs becomes **2.4 GB of Parquet** — a **96.6% total reduction** — containing 1,690,691 clean markdown documents.
181
 
182
  ### Processing Times
183
 
184
+ Pipeline timings across 86 shards of CC-MAIN-2026-08:
185
 
186
  ```
187
+ Download (raw WARC) ████████████████████████ total 2h 27m 20s avg 1m 42s
188
  Convert (HTML → MD) ██░░░░░░░░░░░░░░░░░░░░░░ total 16m 2s avg 11s
189
+ Export (Parquet) █████░░░░░░░░░░░░░░░░░░░ total 33m 28s avg 23s
190
+ Publish (HuggingFace) █████████░░░░░░░░░░░░░░░ total 1h 1m 22s avg 42s
191
  ```
192
 
193
  ### Dataset Charts
data/CC-MAIN-2026-08/00107.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:451c8d59718ffc9c7f76522f632c6fccd1b69492b42b8cf84353a5c0f9a65d5f
3
+ size 30876795
stats.csv CHANGED
@@ -50,13 +50,14 @@ CC-MAIN-2026-08,103,19511,2593312947,96164480,30296223,2026-03-13T06:47:48Z,187,
50
  CC-MAIN-2026-08,104,19464,2548377519,91323603,28774002,2026-03-13T06:54:05Z,107,0,29,239
51
  CC-MAIN-2026-08,105,19486,2561400814,94725131,30979352,2026-03-13T07:00:59Z,295,0,42,76
52
  CC-MAIN-2026-08,106,19330,2553664208,96434475,31127446,2026-03-13T07:06:40Z,245,0,63,30
 
53
  CC-MAIN-2026-08,251,19610,2594898345,97558492,31917087,2026-03-13T06:42:12Z,277,0,23,136
54
  CC-MAIN-2026-08,252,19240,2525928701,98163046,31104034,2026-03-13T06:47:10Z,238,0,49,10
55
  CC-MAIN-2026-08,253,19312,2557230855,96017199,30399966,2026-03-13T06:51:42Z,130,0,23,117
56
  CC-MAIN-2026-08,254,19506,2588922978,95342460,30328355,2026-03-13T06:56:10Z,136,0,21,109
57
  CC-MAIN-2026-08,255,19631,2593650927,99153070,31590444,2026-03-13T07:01:30Z,218,0,26,74
58
  CC-MAIN-2026-08,256,19178,2558160132,95394551,29553555,2026-03-13T07:06:35Z,236,0,56,10
59
- CC-MAIN-2026-08,257,19389,2538319588,91853092,28846793,2026-03-13T07:08:57Z,122,0,19,0
60
  CC-MAIN-2026-08,401,19517,2605589961,95609438,30404753,2026-03-13T06:41:08Z,218,0,47,107
61
  CC-MAIN-2026-08,402,19446,2552785517,96654791,30588984,2026-03-13T06:43:58Z,98,0,26,44
62
  CC-MAIN-2026-08,403,19320,2579048144,91406292,29375992,2026-03-13T06:49:29Z,198,0,33,98
 
50
  CC-MAIN-2026-08,104,19464,2548377519,91323603,28774002,2026-03-13T06:54:05Z,107,0,29,239
51
  CC-MAIN-2026-08,105,19486,2561400814,94725131,30979352,2026-03-13T07:00:59Z,295,0,42,76
52
  CC-MAIN-2026-08,106,19330,2553664208,96434475,31127446,2026-03-13T07:06:40Z,245,0,63,30
53
+ CC-MAIN-2026-08,107,19477,2610750800,97990183,30876795,2026-03-13T07:10:39Z,203,0,34,0
54
  CC-MAIN-2026-08,251,19610,2594898345,97558492,31917087,2026-03-13T06:42:12Z,277,0,23,136
55
  CC-MAIN-2026-08,252,19240,2525928701,98163046,31104034,2026-03-13T06:47:10Z,238,0,49,10
56
  CC-MAIN-2026-08,253,19312,2557230855,96017199,30399966,2026-03-13T06:51:42Z,130,0,23,117
57
  CC-MAIN-2026-08,254,19506,2588922978,95342460,30328355,2026-03-13T06:56:10Z,136,0,21,109
58
  CC-MAIN-2026-08,255,19631,2593650927,99153070,31590444,2026-03-13T07:01:30Z,218,0,26,74
59
  CC-MAIN-2026-08,256,19178,2558160132,95394551,29553555,2026-03-13T07:06:35Z,236,0,56,10
60
+ CC-MAIN-2026-08,257,19389,2538319588,91853092,28846793,2026-03-13T07:09:49Z,122,0,19,51
61
  CC-MAIN-2026-08,401,19517,2605589961,95609438,30404753,2026-03-13T06:41:08Z,218,0,47,107
62
  CC-MAIN-2026-08,402,19446,2552785517,96654791,30588984,2026-03-13T06:43:58Z,98,0,26,44
63
  CC-MAIN-2026-08,403,19320,2579048144,91406292,29375992,2026-03-13T06:49:29Z,198,0,33,98