Publish shard CC-MAIN-2026-08/00513
Browse files- README.md +10 -10
- data/CC-MAIN-2026-08/00513.parquet +3 -0
- stats.csv +1 -0
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 **16,
|
| 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
|
| 168 |
|
| 169 |
-
| Stage |
|
| 170 |
|---|---|---|---|
|
| 171 |
-
| Raw WARC (.warc.gz, downloaded) | ~
|
| 172 |
| HTML extracted (uncompressed) | 2.0 TB | ~295 TB | — |
|
| 173 |
-
| Packed markdown WARC (.md.warc.gz) | ~35.
|
| 174 |
| Final Parquet (Zstd level 19) | 23.9 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 2.0 TB of uncompressed HTML down to 75.
|
| 177 |
|
| 178 |
-
End to end: ~
|
| 179 |
|
| 180 |
### Processing Times
|
| 181 |
|
| 182 |
-
Pipeline timings across
|
| 183 |
|
| 184 |
```
|
| 185 |
-
Download (raw WARC) ████████████████████████ total 51h
|
| 186 |
Convert (HTML → MD) ██░░░░░░░░░░░░░░░░░░░░░░ total 4h 24m 57s avg 18s
|
| 187 |
-
Export (Parquet) █████░░░░░░░░░░░░░░░░░░░ total 11h 33m
|
| 188 |
Publish (HuggingFace) ██░░░░░░░░░░░░░░░░░░░░░░ total 6h 17m 4s avg 26s
|
| 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 **16,699,145 documents across 863 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 863 files of CC-MAIN-2026-08 (16,699,145 pages total), projected to the full crawl of 100,000 WARC files.
|
| 168 |
|
| 169 |
+
| Stage | 863 files (measured) | 100,000 files (projected) | Reduction |
|
| 170 |
|---|---|---|---|
|
| 171 |
+
| Raw WARC (.warc.gz, downloaded) | ~699.5 GB | ~83 TB | — |
|
| 172 |
| HTML extracted (uncompressed) | 2.0 TB | ~295 TB | — |
|
| 173 |
+
| Packed markdown WARC (.md.warc.gz) | ~35.4 GB | ~3.7 TB | **-98.3%** vs HTML |
|
| 174 |
| Final Parquet (Zstd level 19) | 23.9 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 2.0 TB of uncompressed HTML down to 75.3 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 68.2%.
|
| 177 |
|
| 178 |
+
End to end: ~699.5 GB of raw gzipped WARCs becomes **23.9 GB of Parquet** — a **96.6% total reduction** — containing 16,699,145 clean markdown documents.
|
| 179 |
|
| 180 |
### Processing Times
|
| 181 |
|
| 182 |
+
Pipeline timings across 863 shards of CC-MAIN-2026-08:
|
| 183 |
|
| 184 |
```
|
| 185 |
+
Download (raw WARC) ████████████████████████ total 51h 7m 42s avg 3m 33s
|
| 186 |
Convert (HTML → MD) ██░░░░░░░░░░░░░░░░░░░░░░ total 4h 24m 57s avg 18s
|
| 187 |
+
Export (Parquet) █████░░░░░░░░░░░░░░░░░░░ total 11h 33m 36s avg 48s
|
| 188 |
Publish (HuggingFace) ██░░░░░░░░░░░░░░░░░░░░░░ total 6h 17m 4s avg 26s
|
| 189 |
```
|
| 190 |
|
data/CC-MAIN-2026-08/00513.parquet
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2df8ef5d1ca024ad66ab36fbce7263708ab08dc351833e1882aefa5244562222
|
| 3 |
+
size 29438528
|
stats.csv
CHANGED
|
@@ -330,6 +330,7 @@ CC-MAIN-2026-08,509,19399,2543947909,92615835,29590316,2026-03-13T18:58:26Z,343,
|
|
| 330 |
CC-MAIN-2026-08,510,19416,2552896362,87689107,27880175,2026-03-13T19:06:41Z,361,0,136,23
|
| 331 |
CC-MAIN-2026-08,511,19553,2603640773,95830892,29638105,2026-03-13T19:12:41Z,241,0,114,26
|
| 332 |
CC-MAIN-2026-08,512,19409,2588385924,92008152,29561232,2026-03-13T19:21:41Z,365,0,176,16
|
|
|
|
| 333 |
CC-MAIN-2026-08,551,19551,2609034951,89980913,29167227,2026-03-13T06:42:51Z,280,0,20,176
|
| 334 |
CC-MAIN-2026-08,552,19402,2594817630,94092761,30309254,2026-03-13T06:48:30Z,211,0,46,80
|
| 335 |
CC-MAIN-2026-08,553,19856,2617133708,99106486,30365588,2026-03-13T06:52:51Z,116,0,23,120
|
|
|
|
| 330 |
CC-MAIN-2026-08,510,19416,2552896362,87689107,27880175,2026-03-13T19:06:41Z,361,0,136,23
|
| 331 |
CC-MAIN-2026-08,511,19553,2603640773,95830892,29638105,2026-03-13T19:12:41Z,241,0,114,26
|
| 332 |
CC-MAIN-2026-08,512,19409,2588385924,92008152,29561232,2026-03-13T19:21:41Z,365,0,176,16
|
| 333 |
+
CC-MAIN-2026-08,513,19243,2532969689,94386245,29438528,2026-03-13T19:27:51Z,337,0,31,0
|
| 334 |
CC-MAIN-2026-08,551,19551,2609034951,89980913,29167227,2026-03-13T06:42:51Z,280,0,20,176
|
| 335 |
CC-MAIN-2026-08,552,19402,2594817630,94092761,30309254,2026-03-13T06:48:30Z,211,0,46,80
|
| 336 |
CC-MAIN-2026-08,553,19856,2617133708,99106486,30365588,2026-03-13T06:52:51Z,116,0,23,120
|