tamnd commited on
Commit
fb948f3
·
verified ·
1 Parent(s): 8469a8b

Publish 2 shards CC-MAIN-2026-08/00522–00771

Browse files
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 **17,750,631 documents across 917 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
 
@@ -164,27 +164,27 @@ Empty conversions (pages where trafilatura could not extract meaningful content)
164
 
165
  ### Compression Ratios
166
 
167
- Numbers below are actual measurements summed across all 917 files of CC-MAIN-2026-08 (17,750,631 pages total), projected to the full crawl of 100,000 WARC files.
168
 
169
- | Stage | 917 files (measured) | 100,000 files (projected) | Reduction |
170
  |---|---|---|---|
171
- | Raw WARC (.warc.gz, downloaded) | ~743.3 GB | ~83 TB | — |
172
  | HTML extracted (uncompressed) | 2.1 TB | ~295 TB | — |
173
- | Packed markdown WARC (.md.warc.gz) | ~37.6 GB | ~3.7 TB | **-98.3%** vs HTML |
174
- | Final Parquet (Zstd level 19) | 25.4 GB | ~2.9 TB | **-32.4%** vs packed WARC |
175
 
176
- The big win is the HTML → Markdown step: trafilatura strips all tags, scripts, styles, navigation, and ads, keeping only the main content. This cuts 2.1 TB of uncompressed HTML down to 80.0 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 68.2%.
177
 
178
- End to end: ~743.3 GB of raw gzipped WARCs becomes **25.4 GB of Parquet** — a **96.6% total reduction** — containing 17,750,631 clean markdown documents.
179
 
180
  ### Processing Times
181
 
182
- Pipeline timings across 917 shards of CC-MAIN-2026-08:
183
 
184
  ```
185
- Download (raw WARC) ████████████████████████ total 58h 32m 57s avg 3m 49s
186
  Convert (HTML → MD) █░░░░░░░░░░░░░░░░░░░░░░░ total 4h 40m 31s avg 18s
187
- Export (Parquet) █████░░░░░░░░░░░░░░░░░░░ total 13h 20m 25s avg 52s
188
  Publish (HuggingFace) ██░░░░░░░░░░░░░░░░░░░░░░ total 6h 37m 14s avg 25s
189
  ```
190
 
 
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 **17,789,084 documents across 919 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
 
 
164
 
165
  ### Compression Ratios
166
 
167
+ Numbers below are actual measurements summed across all 919 files of CC-MAIN-2026-08 (17,789,084 pages total), projected to the full crawl of 100,000 WARC files.
168
 
169
+ | Stage | 919 files (measured) | 100,000 files (projected) | Reduction |
170
  |---|---|---|---|
171
+ | Raw WARC (.warc.gz, downloaded) | ~744.9 GB | ~83 TB | — |
172
  | HTML extracted (uncompressed) | 2.1 TB | ~295 TB | — |
173
+ | Packed markdown WARC (.md.warc.gz) | ~37.7 GB | ~3.7 TB | **-98.3%** vs HTML |
174
+ | Final Parquet (Zstd level 19) | 25.5 GB | ~2.9 TB | **-32.4%** vs packed WARC |
175
 
176
+ The big win is the HTML → Markdown step: trafilatura strips all tags, scripts, styles, navigation, and ads, keeping only the main content. This cuts 2.1 TB of uncompressed HTML down to 80.2 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 68.2%.
177
 
178
+ End to end: ~744.9 GB of raw gzipped WARCs becomes **25.5 GB of Parquet** — a **96.6% total reduction** — containing 17,789,084 clean markdown documents.
179
 
180
  ### Processing Times
181
 
182
+ Pipeline timings across 919 shards of CC-MAIN-2026-08:
183
 
184
  ```
185
+ Download (raw WARC) ████████████████████████ total 58h 52m 21s avg 3m 50s
186
  Convert (HTML → MD) █░░░░░░░░░░░░░░░░░░░░░░░ total 4h 40m 31s avg 18s
187
+ Export (Parquet) █████░░░░░░░░░░░░░░░░░░░ total 13h 23m 23s avg 52s
188
  Publish (HuggingFace) ██░░░░░░░░░░░░░░░░░░░░░░ total 6h 37m 14s avg 25s
189
  ```
190
 
data/CC-MAIN-2026-08/00522.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19add84da683ea60ca08d7c9e0d55d53e06c0dee8347053a059d8daeffb49c00
3
+ size 29394990
data/CC-MAIN-2026-08/00771.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d0986f1d27f6917202401e4dd064319cac6a78fe313b9c0188730a3596af45a9
3
+ size 29348198
stats.csv CHANGED
@@ -358,6 +358,7 @@ CC-MAIN-2026-08,518,19610,2602806405,95330083,30705729,2026-03-13T20:18:36Z,482,
358
  CC-MAIN-2026-08,519,19300,2516139542,91335057,28963596,2026-03-13T20:33:22Z,769,0,109,20
359
  CC-MAIN-2026-08,520,19384,2565437765,94738700,30497100,2026-03-13T20:45:44Z,637,0,107,16
360
  CC-MAIN-2026-08,521,19197,2517513528,92647585,29270097,2026-03-13T20:56:38Z,585,0,78,45
 
361
  CC-MAIN-2026-08,551,19551,2609034951,89980913,29167227,2026-03-13T06:42:51Z,280,0,20,176
362
  CC-MAIN-2026-08,552,19402,2594817630,94092761,30309254,2026-03-13T06:48:30Z,211,0,46,80
363
  CC-MAIN-2026-08,553,19856,2617133708,99106486,30365588,2026-03-13T06:52:51Z,116,0,23,120
@@ -471,6 +472,7 @@ CC-MAIN-2026-08,767,19624,2621303163,97571575,30598906,2026-03-13T20:19:09Z,498,
471
  CC-MAIN-2026-08,768,19501,2615201592,97722261,30560353,2026-03-13T20:33:22Z,739,0,131,20
472
  CC-MAIN-2026-08,769,19436,2575617917,92436824,29676466,2026-03-13T20:45:45Z,616,0,111,16
473
  CC-MAIN-2026-08,770,19550,2591679193,98878610,30913397,2026-03-13T20:59:15Z,582,0,89,28
 
474
  CC-MAIN-2026-08,851,19575,2612873485,97000364,30911294,2026-03-13T06:41:39Z,163,0,37,204
475
  CC-MAIN-2026-08,852,19334,2546015208,94959847,30205044,2026-03-13T06:45:52Z,183,0,57,11
476
  CC-MAIN-2026-08,853,19694,2567973975,95901932,30303494,2026-03-13T06:50:12Z,150,0,28,80
 
358
  CC-MAIN-2026-08,519,19300,2516139542,91335057,28963596,2026-03-13T20:33:22Z,769,0,109,20
359
  CC-MAIN-2026-08,520,19384,2565437765,94738700,30497100,2026-03-13T20:45:44Z,637,0,107,16
360
  CC-MAIN-2026-08,521,19197,2517513528,92647585,29270097,2026-03-13T20:56:38Z,585,0,78,45
361
+ CC-MAIN-2026-08,522,19304,2550225762,89991129,29394990,2026-03-13T21:07:54Z,589,0,88,0
362
  CC-MAIN-2026-08,551,19551,2609034951,89980913,29167227,2026-03-13T06:42:51Z,280,0,20,176
363
  CC-MAIN-2026-08,552,19402,2594817630,94092761,30309254,2026-03-13T06:48:30Z,211,0,46,80
364
  CC-MAIN-2026-08,553,19856,2617133708,99106486,30365588,2026-03-13T06:52:51Z,116,0,23,120
 
472
  CC-MAIN-2026-08,768,19501,2615201592,97722261,30560353,2026-03-13T20:33:22Z,739,0,131,20
473
  CC-MAIN-2026-08,769,19436,2575617917,92436824,29676466,2026-03-13T20:45:45Z,616,0,111,16
474
  CC-MAIN-2026-08,770,19550,2591679193,98878610,30913397,2026-03-13T20:59:15Z,582,0,89,28
475
+ CC-MAIN-2026-08,771,19149,2546381668,92769272,29348198,2026-03-13T21:07:54Z,575,0,90,0
476
  CC-MAIN-2026-08,851,19575,2612873485,97000364,30911294,2026-03-13T06:41:39Z,163,0,37,204
477
  CC-MAIN-2026-08,852,19334,2546015208,94959847,30205044,2026-03-13T06:45:52Z,183,0,57,11
478
  CC-MAIN-2026-08,853,19694,2567973975,95901932,30303494,2026-03-13T06:50:12Z,150,0,28,80