Deploy app from main@a3254bf (no paper/ binaries)

Resolve organism names to NCBI accessions in /api/predict

Fix Space root health checks

Deploy hybrid catalog predictions

Update README for hybrid UI

Surface hybrid oxygen source in UI

Add unified strain catalog (100K rows w/ provenance) + selective weak supervision for pH

Re-train deployed phenotype heads on 46K corpus with fixed salt labels

Add MediaDive-derived features (medium pH, NaCl, n_media) — all 4 targets improve

Fix _derive_salt to pick optimum entries — salt MAE 2.52 → 2.17 (-13.7%)

Expand training corpus to 46K strains: species-name → NCBI genome + isolation features

Trim short_description to fit HF Spaces 60-char limit

Track parquet + ubj files via Git LFS

HF Spaces: Dockerfile for v2 (FastAPI + React) deployment

v2 architecture: FastAPI backend + React frontend (pixel-perfect Prototype A)

HF Spaces: add Streamlit SDK frontmatter to README + requirements.txt for pip-based deploy

UI redesign: Prototype A "Lab notebook" from Claude Design bundle

Streamlit UI — catalog-first design with NCBI name resolution

UI prep: pre-train phenotype heads + pre-score uncultured media

v2 results: ESM-2 t6 (8M, 20-protein sample) loses to v1 hand-crafted features on all 4 phenotype targets

Tier 1.5: uncertainty quantification via quantile regression in recommend.py

v2 scaffolding: ESM-2 embedding extraction + GPU runner doc

Final demo deliverable: v1 baseline + 5K uncultured predictions + recommender online

Phase E #2: scripts/recommend.py — single-genome → ranked media + phenotype CLI

Tests for media recommender (3 passing) — ROC-AUC > 0.6 on synthetic signal

Phase E modeling: per-medium classifiers + recommender training script

Phase E scaffolding: MediaDive integration + strain↔medium links

Fix GTDB column names + accession resolution for v226 metadata schema

Phase C scaffolding: GTDB candidate selection + uncultured prediction

v1 features: wire tetranucleotides + codon usage into streaming pipeline

Final cleanup: sync OVERNIGHT_SUMMARY.md + fix size display for small files

Final overnight commit: trained baseline + eval report + summary

Make OVERNIGHT_SUMMARY.md write atomic (avoid race with regen loop)

Add v1 composition features (tetranucleotides + codon usage)

Fix predictions parquet type mix + plumb feature_cols through eval

Add tests for explore module (correlations + class means)

Harden post-featurize chain: each phase runs even if previous fails

Add feature↔target correlation analysis to eval report

Update README to reflect current state

Eval report enhancements: TL;DR + per-strain predictions + per-family error

Fix classification fold bug + add end-to-end integration tests

Add eval report generator + training table persistence + group-col override

Streaming fetch+featurize pipeline + 6× pyrodigal speedup + GCA version resolution

Rewrite BacDive client for v2 public API (no auth required)

Scaffold v0: BacDive + NCBI ingestion, genome feature extractor, XGBoost baseline