Skip to content

Releases: zzkws/evidence-rag-pilot

Evidence RAG Pilot v0.1.2

Choose a tag to compare

@zzkws zzkws released this 07 Aug 15:21
0631c62

Evidence RAG Pilot v0.1.2

This release focuses the public repository on the technical design and original TXC/TKD demos.

  • removes the incomplete evaluation dataset, evaluation runners, saved metrics, and generated evaluation reports
  • removes duplicate Markdown reports and unused presentation code
  • makes the root README the single technical project narrative
  • states explicitly that no formal quantitative evaluation has been completed
  • keeps the original static demo sites, pipeline implementation, MCP evidence interface, tests, licenses, and data notices

Live demos:

Full corpus archives and SHA-256 checksums remain in v0.1.0:
https://github.com/zzkws/evidence-rag-pilot/releases/tag/v0.1.0

Evidence RAG Pilot v0.1.1

Choose a tag to compare

@zzkws zzkws released this 07 Aug 15:04

Evidence RAG Pilot v0.1.1

This patch corrects the public presentation layer so the two Sites deployments preserve the original demo interfaces developed for the project.

  • Restores the original four-page structure: corpus/search overview, chunk browser, document tree, and evidence-dialog page.
  • Publishes TXC and TKD as separate static corpus snapshots.
  • Keeps the original visual language and page organization.
  • Removes any requirement for a hosted Python RAG service, API key, GPU, or cloud model.
  • Labels browser BM25 and static evidence summaries honestly; the full dense/RRF and generation pipeline remains local.

Live demos:

The four complete corpus archives and SHA-256 checksums remain available in v0.1.0 and are not duplicated in this patch release:
https://github.com/zzkws/evidence-rag-pilot/releases/tag/v0.1.0

Evidence RAG Pilot v0.1.0

Choose a tag to compare

@zzkws zzkws released this 07 Aug 14:31

Evidence RAG Pilot v0.1.0

A public technical prototype and portfolio demonstration of page-native, traceable RAG for engineering datasheets. This release focuses on the design and implementation; the project has not undergone a formal quantitative evaluation and makes no benchmark or SOTA claim.

Live demonstrations

The public sites preserve the original demo structure and read versioned static corpus snapshots without a GPU or API key. Dense retrieval, RRF, ingestion, and optional multimodal answering remain local-pipeline capabilities.

Corpus snapshot

  • 182 indexed documents
  • 338 source PDF pages
  • 1,422 page-native evidence chunks

These are corpus inventory counts, not evaluation results.

Data assets

This release contains four archives: TXC source PDFs, TXC derived processing outputs, TKD source PDFs, and TKD derived processing outputs. Verify downloads with SHA256SUMS.txt. Model weights are not duplicated; the repository documents their official sources.

Code is Apache-2.0. Dataset rights and third-party materials are governed separately as described in DATA_NOTICE.md and THIRD_PARTY_NOTICES.md.