tamnd commited on
Commit
bf2eecd
·
verified ·
1 Parent(s): f6e6012

Publish shard CC-MAIN-2026-08/00884

Browse files
Files changed (3) hide show
  1. README.md +11 -11
  2. data/CC-MAIN-2026-08/00884.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 **3,982,228 documents across 204 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 204 files of CC-MAIN-2026-08 (3,982,228 pages total), projected to the full crawl of 100,000 WARC files.
170
 
171
- | Stage | 204 files (measured) | 100,000 files (projected) | Reduction |
172
  |---|---|---|---|
173
- | Raw WARC (.warc.gz, downloaded) | ~165.4 GB | ~83 TB | — |
174
- | HTML extracted (uncompressed) | 492.9 GB | ~295 TB | — |
175
  | Packed markdown WARC (.md.warc.gz) | ~8.4 GB | ~3.7 TB | **-98.3%** vs HTML |
176
  | Final Parquet (Zstd level 19) | 5.7 GB | ~2.9 TB | **-32.0%** 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 492.9 GB of uncompressed HTML down to 17.8 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 68.1%.
179
 
180
- End to end: ~165.4 GB of raw gzipped WARCs becomes **5.7 GB of Parquet** — a **96.6% total reduction** — containing 3,982,228 clean markdown documents.
181
 
182
  ### Processing Times
183
 
184
- Pipeline timings across 204 shards of CC-MAIN-2026-08:
185
 
186
  ```
187
- Download (raw WARC) ████████████████████████ total 6h 41m 55s avg 1m 58s
188
  Convert (HTML → MD) █░░░░░░░░░░░░░░░░░░░░░░░ total 16m 2s avg 4s
189
- Export (Parquet) ████░░░░░░░░░░░░░░░░░░░░ total 1h 22m 36s avg 24s
190
- Publish (HuggingFace) ███████████░░░░░░░░░░░░░ total 3h 5m 0s avg 54s
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 **4,001,771 documents across 205 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 205 files of CC-MAIN-2026-08 (4,001,771 pages total), projected to the full crawl of 100,000 WARC files.
170
 
171
+ | Stage | 205 files (measured) | 100,000 files (projected) | Reduction |
172
  |---|---|---|---|
173
+ | Raw WARC (.warc.gz, downloaded) | ~166.2 GB | ~83 TB | — |
174
+ | HTML extracted (uncompressed) | 495.3 GB | ~295 TB | — |
175
  | Packed markdown WARC (.md.warc.gz) | ~8.4 GB | ~3.7 TB | **-98.3%** vs HTML |
176
  | Final Parquet (Zstd level 19) | 5.7 GB | ~2.9 TB | **-32.0%** 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 495.3 GB of uncompressed HTML down to 17.9 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 68.1%.
179
 
180
+ End to end: ~166.2 GB of raw gzipped WARCs becomes **5.7 GB of Parquet** — a **96.6% total reduction** — containing 4,001,771 clean markdown documents.
181
 
182
  ### Processing Times
183
 
184
+ Pipeline timings across 205 shards of CC-MAIN-2026-08:
185
 
186
  ```
187
+ Download (raw WARC) ████████████████████████ total 6h 43m 43s avg 1m 58s
188
  Convert (HTML → MD) █░░░░░░░░░░░░░░░░░░░░░░░ total 16m 2s avg 4s
189
+ Export (Parquet) ████░░░░░░░░░░░░░░░░░░░░ total 1h 22m 55s avg 24s
190
+ Publish (HuggingFace) ███████████░░░░░░░░░░░░░ total 3h 7m 42s avg 54s
191
  ```
192
 
193
  ### Dataset Charts
data/CC-MAIN-2026-08/00884.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf0f38badb20fc4d81542eff1aa8b8f20421ac6f7d1627e4dc03e9ef70484f97
3
+ size 29251284
stats.csv CHANGED
@@ -119,7 +119,7 @@ CC-MAIN-2026-08,283,19405,2575019840,93976705,30759536,2026-03-13T08:32:01Z,147,
119
  CC-MAIN-2026-08,284,19238,2531057319,96819690,32221564,2026-03-13T08:34:31Z,100,0,41,6
120
  CC-MAIN-2026-08,285,19103,2525313092,97567435,30988415,2026-03-13T08:36:32Z,76,0,23,20
121
  CC-MAIN-2026-08,286,19521,2604143328,95453072,31055879,2026-03-13T08:39:07Z,116,0,23,13
122
- CC-MAIN-2026-08,287,19554,2584064050,99218907,31127830,2026-03-13T08:41:27Z,116,0,22,0
123
  CC-MAIN-2026-08,401,19517,2605589961,95609438,30404753,2026-03-13T06:41:08Z,218,0,47,107
124
  CC-MAIN-2026-08,402,19446,2552785517,96654791,30588984,2026-03-13T06:43:58Z,98,0,26,44
125
  CC-MAIN-2026-08,403,19320,2579048144,91406292,29375992,2026-03-13T06:49:29Z,198,0,33,98
@@ -203,3 +203,4 @@ CC-MAIN-2026-08,880,19493,2562178225,98671696,30493564,2026-03-13T08:32:06Z,154,
203
  CC-MAIN-2026-08,881,19336,2555937495,93001627,29739925,2026-03-13T08:35:55Z,143,0,21,63
204
  CC-MAIN-2026-08,882,19452,2590686157,88956978,28905096,2026-03-13T08:38:35Z,118,0,31,8
205
  CC-MAIN-2026-08,883,19257,2558252415,92279950,30785208,2026-03-13T08:41:09Z,114,0,30,8
 
 
119
  CC-MAIN-2026-08,284,19238,2531057319,96819690,32221564,2026-03-13T08:34:31Z,100,0,41,6
120
  CC-MAIN-2026-08,285,19103,2525313092,97567435,30988415,2026-03-13T08:36:32Z,76,0,23,20
121
  CC-MAIN-2026-08,286,19521,2604143328,95453072,31055879,2026-03-13T08:39:07Z,116,0,23,13
122
+ CC-MAIN-2026-08,287,19554,2584064050,99218907,31127830,2026-03-13T08:43:14Z,116,0,22,106
123
  CC-MAIN-2026-08,401,19517,2605589961,95609438,30404753,2026-03-13T06:41:08Z,218,0,47,107
124
  CC-MAIN-2026-08,402,19446,2552785517,96654791,30588984,2026-03-13T06:43:58Z,98,0,26,44
125
  CC-MAIN-2026-08,403,19320,2579048144,91406292,29375992,2026-03-13T06:49:29Z,198,0,33,98
 
203
  CC-MAIN-2026-08,881,19336,2555937495,93001627,29739925,2026-03-13T08:35:55Z,143,0,21,63
204
  CC-MAIN-2026-08,882,19452,2590686157,88956978,28905096,2026-03-13T08:38:35Z,118,0,31,8
205
  CC-MAIN-2026-08,883,19257,2558252415,92279950,30785208,2026-03-13T08:41:09Z,114,0,30,8
206
+ CC-MAIN-2026-08,884,19543,2583118668,92146843,29251284,2026-03-13T08:43:18Z,108,0,19,0