LEAD evidence chassis: review ledger, kernel, and agent tools #1427

Merged

kert merged 31 commits from feat/lead-evidence-chassis-part1 into main 2026-11-02 06:09:51 +00:00

Conversation 0 Commits 31 Files Changed 113 +16913 -9

kert commented 2026-11-01 03:08:20 +00:00

Owner

Copy Link

Summary

• Part 1 adds the five LEAD review projects, the search parsers and the nightly due check, dedup, the calibration and audit policy, PRISMA 2020 flow counts, the grey-literature and Congress.gov parsers, the BLS, FRED, and IRS readers, and the append-only lead_evidence ledger exported from Zotero.
• Parts 2–4 add High Needs scoring, coding and care suspect moves kept as separate numbers, the limits calculator and five rule-tier claims, the public-manifest loader, promotion gates, and the Python tool surface. Unpublished CMS pieces stay recorded gaps. The loader records manifest rows and does not download the files.
• A release row is identified by source, release, sha256, path, and URL. A repeated Databricks start returns the existing run and does not call the transport again.

Tests

The full suite on d6590a2 passed: 16378 passed, 56 skipped.

Outside this branch

OSF registration and the live search are operational steps. They are not commits here.

## Summary - Part 1 adds the five LEAD review projects, the search parsers and the nightly due check, dedup, the calibration and audit policy, PRISMA 2020 flow counts, the grey-literature and Congress.gov parsers, the BLS, FRED, and IRS readers, and the append-only `lead_evidence` ledger exported from Zotero. - Parts 2–4 add High Needs scoring, coding and care suspect moves kept as separate numbers, the limits calculator and five rule-tier claims, the public-manifest loader, promotion gates, and the Python tool surface. Unpublished CMS pieces stay recorded gaps. The loader records manifest rows and does not download the files. - A release row is identified by source, release, sha256, path, and URL. A repeated Databricks start returns the existing run and does not call the transport again. ## Tests The full suite on `d6590a2` passed: 16378 passed, 56 skipped. ## Outside this branch OSF registration and the live search are operational steps. They are not commits here.

kert added 28 commits 2026-11-01 03:08:20 +00:00

docs: LEAD evidence chassis design and part 1 plan 6634109484

feat(lead): add the LEAD target registry 07a0f32afe

feat(lead): append-only evidence ledger in DuckDB 4be049c07b

fix(prisma): fail the screen command when the model errors fbc8fa998d

feat(prisma): per-project Sci-Hub switch, on for LEAD ae8e46f046

feat(prisma): add the five LEAD review starters 44b21f8d8d

feat(prisma): dedup on DOI, PMID, then title and year 937a2bbfd1

feat(prisma): PubMed search log and seed recall 43d8e51886

feat(prisma): parsers for Europe PMC, OpenAlex, Crossref, Semantic Scholar, arXiv aa1d31db24

feat(prisma): screening calibration gate 797f679141

feat(prisma): two-pass disagreements and a 10% audit sample 55614648f3

feat(prisma): PRISMA 2020 flow counts from tags fc4266a938

feat(lead): export Zotero review rows into the ledger aab21a0fd1

feat(lead): claims, GRADE start, and the precursor catalog ad7f78e52d

feat(prisma): grey-literature harvest parsers 9e9908dd9f

feat(prisma): Congress.gov bill parser 9a85d766d7

feat: parse PPI, JOLTS, QCEW, FRED, and IRS SOI extracts 2153094b77

feat(prisma): nightly and triggered search due check 32a1c5a0b8

fix: correct the part 1 review findings ... fe99b6bbf2

Export reads the extraction JSON instead of a tag fragment and writes the claim. A logged search stores its project so the nightly check can clear. Claims keep the effect and the subject, a later strand tags an existing DOI, the dedup command prints borderline pairs, and due dates compare the calendar day.

fix(perf): let the collector trace fixture replace an existing provider ... 9788f95ede

OpenTelemetry ignores a second set_tracer_provider. The full suite sets one from the BCDA metrics test, so the collector's first span test saw no spans.

docs(lead): design the kernel, loop, and tool surface 289649b3ff

fix: close the part 1 review minors 080df238b3

feat(lead): score High Needs criteria and separate suspect moves 9bddb3ddf0

feat(lead): add the limits calculator and rule-tier claims c1a8db91b4

feat(lead): load the public manifest and record measure gaps 31a706beb5

feat(lead): gate promotion and record discarded runs 97816d60af

feat(lead): add the agent tool surface 370f64f193

fix(lead): keep every manifest row and refuse a repeated Databricks start d6590a2357

kert added 1 commit 2026-11-02 05:23:35 +00:00

feat(lead): load the public mirror and serve the explorer ... bb747f272d

Local files land in the DuckDB mirror with the source headers kept: group labels, stacked IRS names, headerless code maps, and the HCPCS action key. The read-only explorer and the review page are wired at lead.fhirworx.io. OEWS now includes 2025.

kert added 1 commit 2026-11-02 05:48:46 +00:00

fix(lead): open the Zotero library on each labels run ... ed2e21ce8d

The cached connection was created on one script thread and reused on the next, so the labels page died in project_keys. The file is mounted read-only, so the open is immutable and does not create a -shm beside it. The explorer pages are covered for the 99% gate.

kert added 1 commit 2026-11-02 05:53:06 +00:00

merge main into feat/lead-evidence-chassis-part1 246aa2ac30

kert merged commit c071ab5feb into main 2026-11-02 06:09:51 +00:00

kert referenced this issue from a commit 2026-11-02 06:09:53 +00:00

Merge pull request 'LEAD evidence chassis: review ledger, kernel, and agent tools' (#1427) from feat/lead-evidence-chassis-part1 into main

Sign in to join this conversation.

Reviewers

No Reviewers

Labels

Clear labels

api

REST API (FastAPI)

auth

Authentication and authorization

bib

Bibliography, Zotero, citations

ci

CI/CD pipelines and Woodpecker plugins

cli

CLI commands and UX

cloud

Cloud provider abstraction (AWS, GCP, Azure, self-hosted)

config

Configuration and wiring

data

Data loading, CCLF, BCDA, seeds

databricks

Databricks jobs and deployment

docs

Documentation and generated references

image:api

image:docs

image:mc

image:notebooks

image:zotero

infra

lakehouse

Nessie, Polaris, Trino, Iceberg

llm

llm module — local RAG over comments + Zotero corpus

notebooks

marimo notebook errors (auto-filed by nb-watcher/nb-integration)

packaging

Python packaging, optional deps, build system

pfs

pipeline

Pipeline execution and validation

pipeline:ci

pipeline:deploy

pipeline:harden

pipeline:infra-ci

pipeline:rebuild-all

pipeline:release

qpp

quality

Testing, coverage, error handling

step:build

step:deploy

step:hadolint

step:lint

step:publish

step:scan

step:test

step:validate

styles

Charts, figures, visual theming

tree

Project directory structure and file organization

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

kert renovate-bot

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: homelab/stack#1427