tamnd commited on
Commit
934f32b
·
verified ·
1 Parent(s): 31108c1

Publish shard CC-MAIN-2026-08/00104

Browse files
Files changed (3) hide show
  1. README.md +14 -14
  2. data/CC-MAIN-2026-08/00104.parquet +3 -0
  3. stats.csv +3 -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 **1,185,523 documents across 60 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 60 files of CC-MAIN-2026-08 (1,185,523 pages total), projected to the full crawl of 100,000 WARC files.
170
 
171
- | Stage | 60 files (measured) | 100,000 files (projected) | Reduction |
172
  |---|---|---|---|
173
- | Raw WARC (.warc.gz, downloaded) | ~48.6 GB | ~83 TB | — |
174
- | HTML extracted (uncompressed) | 147.8 GB | ~295 TB | — |
175
- | Packed markdown WARC (.md.warc.gz) | ~2.4 GB | ~3.7 TB | **-98.4%** vs HTML |
176
- | Final Parquet (Zstd level 19) | 1.7 GB | ~2.9 TB | **-31.5%** 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 147.8 GB of uncompressed HTML down to 5.1 GB of markdown — a **98.4% reduction** — before any file-level compression is applied. Parquet with Zstd level 19 then compresses the markdown a further 67.8%.
179
 
180
- End to end: ~48.6 GB of raw gzipped WARCs becomes **1.7 GB of Parquet** — a **96.6% total reduction** — containing 1,185,523 clean markdown documents.
181
 
182
  ### Processing Times
183
 
184
- Pipeline timings across 60 shards of CC-MAIN-2026-08:
185
 
186
  ```
187
- Download (raw WARC) ████████████████████████ total 1h 3m 57s avg 1m 3s
188
- Convert (HTML → MD) █████░░░░░░░░░░░░░░░░░░ total 16m 2s avg 16s
189
- Export (Parquet) ██████░░░░░░░░░░░░░░░░░░ total 16m 13s avg 16s
190
- Publish (HuggingFace) ██████████░░░░░░░░░░░░░░ total 29m 3s avg 29s
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 **1,224,824 documents across 62 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 62 files of CC-MAIN-2026-08 (1,224,824 pages total), projected to the full crawl of 100,000 WARC files.
170
 
171
+ | Stage | 62 files (measured) | 100,000 files (projected) | Reduction |
172
  |---|---|---|---|
173
+ | Raw WARC (.warc.gz, downloaded) | ~50.3 GB | ~83 TB | — |
174
+ | HTML extracted (uncompressed) | 152.7 GB | ~295 TB | — |
175
+ | Packed markdown WARC (.md.warc.gz) | ~2.5 GB | ~3.7 TB | **-98.4%** vs HTML |
176
+ | Final Parquet (Zstd level 19) | 1.7 GB | ~2.9 TB | **-31.6%** 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 152.7 GB of uncompressed HTML down to 5.3 GB of markdown — a **98.4% reduction** — before any file-level compression is applied. Parquet with Zstd level 19 then compresses the markdown a further 67.8%.
179
 
180
+ End to end: ~50.3 GB of raw gzipped WARCs becomes **1.7 GB of Parquet** — a **96.6% total reduction** — containing 1,224,824 clean markdown documents.
181
 
182
  ### Processing Times
183
 
184
+ Pipeline timings across 62 shards of CC-MAIN-2026-08:
185
 
186
  ```
187
+ Download (raw WARC) ████████████████████████ total 1h 9m 49s avg 1m 7s
188
+ Convert (HTML → MD) █████░░░░░░░░░░░░░░░░░░ total 16m 2s avg 15s
189
+ Export (Parquet) ██████░░░░░░░░░░░░░░░░░░ total 17m 40s avg 17s
190
+ Publish (HuggingFace) ████████████░░░░░░░░░░░░ total 35m 4s avg 33s
191
  ```
192
 
193
  ### Dataset Charts
data/CC-MAIN-2026-08/00104.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae7c6c210acbf1f94b10c3384f7c39e586424362a130d0d4052efe05d3fc7453
3
+ size 28774002
stats.csv CHANGED
@@ -50,12 +50,14 @@ CC-MAIN-2026-08,253,19312,2557230855,96017199,30399966,2026-03-13T06:51:42Z,130,
50
  CC-MAIN-2026-08,401,19517,2605589961,95609438,30404753,2026-03-13T06:41:08Z,218,0,47,107
51
  CC-MAIN-2026-08,402,19446,2552785517,96654791,30588984,2026-03-13T06:43:58Z,98,0,26,44
52
  CC-MAIN-2026-08,403,19320,2579048144,91406292,29375992,2026-03-13T06:49:29Z,198,0,33,98
 
53
  CC-MAIN-2026-08,551,19551,2609034951,89980913,29167227,2026-03-13T06:42:51Z,280,0,20,176
54
  CC-MAIN-2026-08,552,19402,2594817630,94092761,30309254,2026-03-13T06:48:30Z,211,0,46,80
55
- CC-MAIN-2026-08,553,19856,2617133708,99106486,30365588,2026-03-13T06:50:50Z,116,0,23,0
56
  CC-MAIN-2026-08,701,19412,2563976599,92296196,29384755,2026-03-13T06:41:28Z,272,0,22,99
57
  CC-MAIN-2026-08,702,19449,2572639558,92293569,29721653,2026-03-13T06:44:08Z,113,0,20,24
58
  CC-MAIN-2026-08,703,19592,2602952509,95185029,30720858,2026-03-13T06:49:34Z,187,0,36,100
 
59
  CC-MAIN-2026-08,851,19575,2612873485,97000364,30911294,2026-03-13T06:41:39Z,163,0,37,204
60
  CC-MAIN-2026-08,852,19334,2546015208,94959847,30205044,2026-03-13T06:45:52Z,183,0,57,11
61
  CC-MAIN-2026-08,853,19694,2567973975,95901932,30303494,2026-03-13T06:50:12Z,150,0,28,80
 
50
  CC-MAIN-2026-08,401,19517,2605589961,95609438,30404753,2026-03-13T06:41:08Z,218,0,47,107
51
  CC-MAIN-2026-08,402,19446,2552785517,96654791,30588984,2026-03-13T06:43:58Z,98,0,26,44
52
  CC-MAIN-2026-08,403,19320,2579048144,91406292,29375992,2026-03-13T06:49:29Z,198,0,33,98
53
+ CC-MAIN-2026-08,404,19609,2626946886,98313227,30726076,2026-03-13T06:53:10Z,177,0,43,0
54
  CC-MAIN-2026-08,551,19551,2609034951,89980913,29167227,2026-03-13T06:42:51Z,280,0,20,176
55
  CC-MAIN-2026-08,552,19402,2594817630,94092761,30309254,2026-03-13T06:48:30Z,211,0,46,80
56
+ CC-MAIN-2026-08,553,19856,2617133708,99106486,30365588,2026-03-13T06:52:51Z,116,0,23,120
57
  CC-MAIN-2026-08,701,19412,2563976599,92296196,29384755,2026-03-13T06:41:28Z,272,0,22,99
58
  CC-MAIN-2026-08,702,19449,2572639558,92293569,29721653,2026-03-13T06:44:08Z,113,0,20,24
59
  CC-MAIN-2026-08,703,19592,2602952509,95185029,30720858,2026-03-13T06:49:34Z,187,0,36,100
60
+ CC-MAIN-2026-08,704,19692,2578096724,95466414,30221084,2026-03-13T06:53:15Z,175,0,44,0
61
  CC-MAIN-2026-08,851,19575,2612873485,97000364,30911294,2026-03-13T06:41:39Z,163,0,37,204
62
  CC-MAIN-2026-08,852,19334,2546015208,94959847,30205044,2026-03-13T06:45:52Z,183,0,57,11
63
  CC-MAIN-2026-08,853,19694,2567973975,95901932,30303494,2026-03-13T06:50:12Z,150,0,28,80