Add Focus Crawl and BioCollection counts

Add exact Marin-tokenizer counts for the Common Crawl Focus Crawl and both TheBioCollection streams registered in the Marin datakit source catalog.

data/token_counts.csv

@@ -1,6 +1,9 @@
 marin_name,marin_tokens,category,synthetic,machine_translated,pdf,hf_repo,hf_subset,license,transform_tldr
+biocollection/free_text_stream,33186704843,specialized,True,False,False,trillionlabs/TheBioCollection,free_text_stream,apache-2.0,Free-text renderings of biological entities enriched with computed properties
+biocollection/instruction_stream,18123700603,specialized,True,False,False,trillionlabs/TheBioCollection,instruction_stream,apache-2.0,Instruction-response data with programmatically verifiable biological answers
 climblab-ja,371918129733,multilingual,False,False,False,KantaHayashiAI/ClimbLab-Ja,,odc-by,
 coderforge,10288088964,code,True,False,False,togethercomputer/CoderForge-Preview,,"Per-file: MIT, BSD-3, Apache-2.0, BSD, BSD-2, BSD-4, ISC, PostgreSQL, PSF, CC0-1.0, MIT-NA, MIT-CMU, HPND, Dual MIT/Apache-2.0, Dual Apache-2.0/BSD-3, Dual BSD-3/MIT",
+common-crawl-focus-2026-22,49702569456,web,False,False,False,,,Per-document: source website terms,Common Crawl science focus crawl HTML extracted with XenonMolecule/jusText@20d27c0
 common_corpus/english,1015389781525,web,False,False,False,PleIAs/common_corpus,,Per-document: public domain or various open licenses,
 cp/arxiv_abstracts,536806535,web,False,False,False,common-pile/arxiv_abstracts,,CC0 1.0,
 cp/arxiv_papers,6632151543,web,False,False,True,common-pile/arxiv_papers,,CC BY (per-paper),