feat: conversation memory verified — 159 tests passing

feat: 159 tests passing — conversation memory complete

add changes and update on claude.MD

fix: Deep mode now correctly calls summarizer agent

feat: clickable PubMed links for cited PMIDs

fix: sync app.py with streamlit_app.py — analysis mode added

fix: remove corrupted em-dash in settings.py for HuggingFace

fix: RAG prompt JSON format restored — PMIDs working

refactor: centralize system prompts in prompts.py

change the claude model from opus to sonnet again

docs: Plan 4 complete — all 4 plans done

feat(plan3): complete — 44,676 vectors, pTau217 cross-source working

feat(preprocessing): ADNI loader complete — 24,916 clinical records in Pinecone

fix: replace corrupted .gitignore with clean version

add changes notebook files

docs: Session 4 complete — HuggingFace live, ADNI files ready

docs: Plan 2 complete — HuggingFace deployment live

fix: add root app.py, correct README metadata, minimal requirements

fix: minimal requirements for HuggingFace — remove torch and heavy packages

fix: add correct HuggingFace Space metadata to README

docs: add HuggingFace Space metadata to README

modified ingestion /pubmed_diagnostics.py

test: add 94 unit tests across all pipeline stages — 100% passing

chore: exclude CLAUDE_CONTEXT.md from repo

docs: Plan 1 complete — update session context

feat(app): enterprise Streamlit chatbot — RAG pipeline working end to end

feat(generation): enterprise RAG pipeline — verified working with Claude + Pinecone

delete exposed api key

feat(vector_store): enterprise Pinecone query layer — retrieval verified working

add changes in embedded py

feat(notebooks): add Colab T4 embedding notebook — 19,771 chunks to Pinecone

feat(preprocessing): cleaner + chunker enterprise code | fix pinecone index name to alzdetect

feat(preprocessing): enterprise paper cleaner — 49 years inferred, 5120 HTML stripped, 100% retention

feat(ingestion): pubmed_fetch and query_config verified — 19637 papers 100% valid

fix(ingestion): handle old project format — coerce authors/keywords/year types

feat(ingestion): enterprise PubMed fetcher with Pydantic validation and checkpointing

feat(ingestion): Add 40 enterprise PubMed queries with Pydantic validation

Merge branch 'main' of https://github.com/tpriyadata/Alz-Bot

Initial commit

chore: initialise enterprise folder structure Alz-Bot

chore: add .gitignore before anything else