Skip to content

cllm-claims-plan.md

Latest commit

 

History

History
32 lines (26 loc) · 1.6 KB

cllm-claims-plan.md

File metadata and controls

32 lines (26 loc) · 1.6 KB
Raw

CLLM Claims Plan

Goal

Add a resumable cllm-based claim extraction pass for every abstract and expose the results in the static UI detail panel without making the right-hand reading experience overwhelming.

Decisions

  • Keep claim extraction as its own cache file: data/claim_analyses_cllm.json
  • Default the cllm integration to LLM_PROVIDER=openai so the existing repo .env works without adding Anthropic credentials
  • Fall back to gpt-4o-2024-08-06 as the default OpenAI claim-extraction model while GPT-5-family behavior is investigated
  • Build the cllm manuscript from core abstract sections only, excluding references and acknowledgements
  • Include cached figure-analysis text in the cllm manuscript when local assets have already been analyzed
  • Merge cached claim output into data/primary/abstracts_enriched.json
  • Flow claim output into abstracts.detail.json
  • Render abstract sections, claims, figure notes, and references as collapsible blocks in the UI
  • Tighten reference spacing and font sizing for denser reading

Execution

  1. Run aacli analyze-figures first if you want cached figure notes included in the claim manuscript.
  2. Run aacli extract-claims to populate or resume the claim cache.
  3. Run aacli enrich to merge claim extraction into the enriched abstract data.
  4. Run scripts/build_ui_data.py (Stage 6) so the SvelteKit site's data package picks up the new fields. (The legacy aacli export-ui / build-ui commands have been removed; this plan predates Stage 2.1 + Stage 6 and is kept for historical reference only.)