v2.3.1: Complete channel sync — all files v2.3.1, 12 cols, English-only, pretty-printed samples

MANIFEST_v2.json

@@ -6,6 +6,10 @@
   "unique_species": 2313,
   "sha256_parquet": "6f72ab8e13437bd04b5a84a22539216191486b477bc4abb009b24ad2ba933910",
   "sha256_json": "3b042206b5c65ef9b051fff7f891ca09197be447bf715313ff290008b0de2136",
+  "filenames": {
+    "json": "ethno_dataset_2026_v2.3.1.json",
+    "parquet": "ethno_dataset_2026_v2.3.1.parquet"
+  },
   "file_sizes_mb": {
     "parquet": 1.423,
     "json": 27.765

METHODOLOGY.md

@@ -1,6 +1,6 @@
-# Methodology — Ethno-API Phytochemical Dataset v2.3
+# Methodology — Ethno-API Phytochemical Dataset v2.3.1
 
-> **Schema v2.3 · 76,907 records · 24,746 compounds · 2,313 species · 10 fields**
+> **Schema v2.3.1 · 76,907 records · 24,746 compounds · 2,313 species · 12 fields**
 
 ---
 
@@ -188,6 +188,7 @@ All enrichment scripts are available in the repository:
 | v2.1 | 2026-03 | 10 (+pubchem_cid, canonical_smiles) | 76,907 | PubChem CID + SMILES enrichment (10,484 chemicals resolved, 71.8% record coverage — corrected to 42.4% in v2.2) |
 | v2.2 | 2026-03 | 10 (same schema) | 76,907 | Stereo-prefix normalization for CT matching (+2 compounds), corrected SMILES coverage reporting (42.4% of unique chemicals), local CT XML matching replaces API |
 | **v2.3** | **2026-03** | **10 (same schema)** | **76,907** | **CTS synonym enrichment: 997 compounds resolved via PubChem name variants (hyphen→space normalization), PubChem CID coverage 42.4%→46.4% unique / 71.8%→75.4% records** |
+| **v2.3.1** | **2026-04** | **12 (+compound_type, patent_count_method)** | **76,907** | **CID audit, compound classification, patent method transparency, RESIN/RESINS CID correction, inorganic/generic reclassification** |
 
 ---
 
@@ -204,4 +205,7 @@ All enrichment scripts are available in the repository:
 | `ethno_dataset_2026_v2.3.json` | 25.6 MB | `956cd7b08d279792e132629ce608ab72eb9249b09ac84ed9c60bd108acb5057e` |
 | `ethno_dataset_2026_v2.3.parquet` | 1,211 KB | `4b7351048db025cbf575b4538e66afd70729c327b8a0b48ead87d5546a39762e` |
 
-Export timestamp: `2026-03-25T12:00:00Z` (v2.3 final)
+| `ethno_dataset_2026_v2.3.1.json` | 27.8 MB | `3b042206b5c65ef9b051fff7f891ca09197be447bf715313ff290008b0de2136` |
+| `ethno_dataset_2026_v2.3.1.parquet` | 1,458 KB | `6f72ab8e13437bd04b5a84a22539216191486b477bc4abb009b24ad2ba933910` |
+
+Export timestamp: `2026-04-11T22:47:02Z` (v2.3.1 final)

README.md

@@ -126,7 +126,7 @@ Full methodology is documented in `METHODOLOGY.md`. Known limitations are listed
 | `chembl_bioactivity_count` | `int32` | 0% | ChEMBL documented bioactivity measurement count |
 | `patent_count_since_2020` | `int32` | ~0.9% | US patents since 2020-01-01 mentioning compound (USPTO PatentsView) |
 | `pubchem_cid` | `int64` | ~28.2% | PubChem Compound ID (CID) — resolved via PubChem PUG REST (March 2026) |
-| `canonical_smiles` | `string` | ~28.2% | Canonical SMILES notation — molecular structure from PubChem (75.4% of unique compounds resolved in v2.3) |
+| `canonical_smiles` | `string` | ~28.2% | Canonical SMILES notation — molecular structure from PubChem (75.4% of unique compounds resolved in v2.3/v2.3.1) |
 | `compound_type` | `string` | 0% | Classification: `discrete_phytochemical`, `substance_class`, `complex_mixture`, `inorganic_element`, `generic_ambiguous` — added in v2.3.1 |
 | `patent_count_method` | `string` | ~0.9% | Query methodology: `name_based_with_cid`, `name_based_no_cid`, `name_based_invalidated`, `NULL` — added in v2.3.1 |
 
@@ -136,11 +136,11 @@ Full methodology is documented in `METHODOLOGY.md`. Known limitations are listed
 
 | Tier | Price | Includes | Purchase |
 |------|-------|----------|----------|
-| **Single Entity** | **€699** netto | JSON + Parquet + SHA-256 Manifest. 1 juristische Person, interne Nutzung. Perpetual license. | [**Buy Now →**](https://buy.stripe.com/00w6oGgFh58v6Toeqsebu02?utm_source=github&utm_medium=readme&utm_campaign=launch_2026_03) |
-| **Team** | **€1,349** netto | Alles aus Single + `duckdb_queries.sql` (20 Queries, 5 Kategorien) + `compound_priority_score.py` + 4 Pre-computed Views (Top-500 nach PubMed, Trials, Patent-Dichte, Anti-Inflammatory Panel). Unbegrenzte interne Nutzer einer juristischen Person. | [**Buy Now →**](https://buy.stripe.com/dRm7sK9cP1Wj0v06Y0ebu03?utm_source=github&utm_medium=readme&utm_campaign=launch_2026_03) |
-| **Enterprise** | **€1,699** netto | Alles aus Team + `snowflake_load.sql` + `chromadb_ingest.py` + `pinecone_ingest.py` + `embedding_guide.md` (ClinicalBERT, RAG-Pipelines) + Compound Opportunity Matrix + Clinical Pipeline Gaps CSV + Pre-chunked RAG JSONL. Multi-Entity / Konzernnutzung, interne Produktintegration erlaubt. | [**Contact for Enterprise →**](mailto:founder@ethno-api.com?subject=Enterprise%20License%20Inquiry) |
+| **Single Entity** | **€699** net | JSON + Parquet + SHA-256 Manifest. 1 legal entity, internal use. Perpetual license. | [**Buy Now →**](https://buy.stripe.com/00w6oGgFh58v6Toeqsebu02?utm_source=github&utm_medium=readme&utm_campaign=launch_2026_03) |
+| **Team** | **€1,349** net | Everything in Single + `duckdb_queries.sql` (20 queries, 5 categories) + `compound_priority_score.py` + 4 pre-computed views (Top-500 by PubMed, Trials, Patent Density, Anti-Inflammatory Panel). Unlimited internal users within 1 legal entity. | [**Buy Now →**](https://buy.stripe.com/dRm7sK9cP1Wj0v06Y0ebu03?utm_source=github&utm_medium=readme&utm_campaign=launch_2026_03) |
+| **Enterprise** | **€1,699** net | Everything in Team + `snowflake_load.sql` + `chromadb_ingest.py` + `pinecone_ingest.py` + `embedding_guide.md` (ClinicalBERT, RAG pipelines) + Compound Opportunity Matrix + Clinical Pipeline Gaps CSV + Pre-chunked RAG JSONL. Multi-entity / group use, internal product integration permitted. | [**Contact for Enterprise →**](mailto:founder@ethno-api.com?subject=Enterprise%20License%20Inquiry) |
 
-> Gemäß § 19 UStG wird keine Umsatzsteuer berechnet. Alle Preise netto. One-time purchase — keine Subscription, keine wiederkehrenden Kosten.
+> No VAT charged (German small business exemption, §19 UStG). All prices net. One-time purchase — no subscription, no recurring costs.
 
 ---
 
@@ -246,18 +246,18 @@ Below is a real record from the dataset — QUERCETIN, one of the most-studied p
 }
 ```
 
-All 76,907 records contain all 12 schema fields. The 4 enrichment columns are always non-null; `pubchem_cid` and `canonical_smiles` are filled for 75.4% of unique compounds (18,675 of 24,746 resolved via PubChem PUG REST in v2.3); `compound_type` and `patent_count_method` are populated for all records; `application` (~50% null) and `dosage` (~87% null) reflect USDA source gaps. Unresolved compounds are phytochemical trivial names, mixture descriptions, or non-specific ethnobotanical terms not indexed in PubChem by name.
+All 76,907 records contain all 12 schema fields. The 4 enrichment columns are always non-null; `pubchem_cid` and `canonical_smiles` are filled for 75.4% of unique compounds (18,675 of 24,746 resolved via PubChem PUG REST in v2.3/v2.3.1); `compound_type` and `patent_count_method` are populated for all records; `application` (~50% null) and `dosage` (~87% null) reflect USDA source gaps. Unresolved compounds are phytochemical trivial names, mixture descriptions, or non-specific ethnobotanical terms not indexed in PubChem by name.
 The free 400-row sample contains real, final enrichment values across all five layers.
 
 ## File Manifest
 
 | File | Size | Format | Access |
 |------|------|--------|--------|
-| `ethno_sample_400.json` | 108 KB | JSON | Free (this repo) |
+| `ethno_sample_400.json` | 179 KB | JSON | Free (this repo) |
 | `ethno_sample_400.parquet` | 20 KB | Parquet | Free (this repo) |
 | `quickstart.ipynb` | 9 KB | Notebook | Free (this repo) |
-| `ethno_dataset_2026_v2.3.json` | ~28 MB | JSON | Included in all tiers |
-| `ethno_dataset_2026_v2.3.parquet` | ~1.5 MB | Parquet | Included in all tiers |
+| `ethno_dataset_2026_v2.3.1.json` | ~28 MB | JSON | Included in all tiers |
+| `ethno_dataset_2026_v2.3.1.parquet` | ~1.5 MB | Parquet | Included in all tiers |
 | `MANIFEST_v2.json` (SHA-256) | ~1 KB | JSON | Included in all tiers |
 | `duckdb_queries.sql` (20 Queries) | ~13 KB | SQL | Team + Enterprise |
 | `compound_priority_score.py` | ~5 KB | Python | Team + Enterprise |
@@ -330,7 +330,7 @@ Ethno-API is the deterministic answer to these pipeline requirements.
 - **Team License — €1,349** one-time: [**Buy →**](https://buy.stripe.com/dRm7sK9cP1Wj0v06Y0ebu03?utm_source=github&utm_medium=readme&utm_campaign=launch_2026_03) — all employees of 1 legal entity, unlimited internal users, includes analytics toolkit.
 - **Enterprise License — €1,699** one-time: [**Contact for Enterprise →**](mailto:founder@ethno-api.com?subject=Enterprise%20License%20Inquiry) — multi-entity / group use, internal product integration rights, full RAG integration toolkit.
 
-> Gemäß § 19 UStG wird keine Umsatzsteuer berechnet.
+> No VAT charged (German small business exemption, §19 UStG).
 
 ```bibtex
 @misc{ethno_api_v23_2026,

UPDATE_POLICY.md

@@ -12,9 +12,17 @@ Next scheduled update: Q1 2027.
 - PubChem PUG REST API (CID + Canonical SMILES)
 
 ## Versioning
-Current version: v2.3 (March 2026)
+Current version: v2.3.1 (April 2026)
 All versions are tagged in this repository.
 
+## v2.3.1 Changelog
+- Added compound_type column (5 categories: discrete_phytochemical, substance_class, complex_mixture, inorganic_element, generic_ambiguous)
+- Added patent_count_method column (query methodology transparency)
+- RESIN/RESINS CID correction (CID 133110026 was alpha-Copaene, set to NULL)
+- Forensic audit corrections for inorganic/generic compounds
+- PubChem CID/SMILES coverage: 75.4% (unchanged from v2.3)
+- DOI: 10.5281/zenodo.19265853
+
 ## v2.3 Changelog
 - PubChem CID/SMILES coverage: 75.4% (up from 71.8% in v2.2)
 - CTS worker resolved 997 additional compounds via PubChem PUG-REST name-to-CID
@@ -23,7 +31,7 @@ All versions are tagged in this repository.
 
 ## Buyer Notifications
 All dataset buyers receive email notification when a new version is available.
-Buyers of v2.0, v2.1, or v2.2 receive the v2.3 update free of charge.
+Buyers of v2.0, v2.1, v2.2, or v2.3 receive the v2.3.1 update free of charge.
 
 ## Breaking Changes
 Schema changes are announced minimum 30 days in advance via GitHub Discussions.

generate_samples_v23.py

@@ -0,0 +1,90 @@
+#!/usr/bin/env python3
+"""
+generate_samples_v23.py
+Generates ethno_sample_400.json and ethno_sample_400.parquet
+from the v2.3 master dataset. Guarantees 100% schema integrity.
+"""
+import json
+import pathlib
+
+MASTER = pathlib.Path("/opt/ethno-enrichment/delivery/ethno_dataset_2026_v2.3.json")
+OUT_DIR = pathlib.Path(__file__).parent
+SAMPLE_SIZE = 400
+
+# Canonical field order (matches MANIFEST_v2.json)
+FIELD_ORDER = [
+    "chemical",
+    "plant_species",
+    "application",
+    "dosage",
+    "pubmed_mentions_2026",
+    "clinical_trials_count_2026",
+    "chembl_bioactivity_count",
+    "patent_count_since_2020",
+    "pubchem_cid",
+    "canonical_smiles",
+]
+
+print(f"Loading master: {MASTER}")
+with open(MASTER) as f:
+    master = json.load(f)
+
+print(f"Master records: {len(master):,}")
+assert len(master) == 76_907, f"Expected 76,907, got {len(master)}"
+
+# Extract first 400, normalize field order and types
+sample = []
+for rec in master[:SAMPLE_SIZE]:
+    row = {}
+    for field in FIELD_ORDER:
+        val = rec.get(field)
+        # pubchem_cid: convert float → int (e.g. 119250.0 → 119250)
+        if field == "pubchem_cid" and isinstance(val, float):
+            val = int(val)
+        row[field] = val
+    sample.append(row)
+
+assert len(sample) == SAMPLE_SIZE, f"Expected {SAMPLE_SIZE}, got {len(sample)}"
+assert list(sample[0].keys()) == FIELD_ORDER, "Field order mismatch"
+
+# Write JSON
+json_path = OUT_DIR / "ethno_sample_400.json"
+with open(json_path, "w") as f:
+    json.dump(sample, f, indent=2, ensure_ascii=False)
+print(f"Wrote {json_path} ({json_path.stat().st_size:,} bytes)")
+
+# Write Parquet
+try:
+    import pandas as pd
+    df = pd.DataFrame(sample)
+    parquet_path = OUT_DIR / "ethno_sample_400.parquet"
+    df.to_parquet(parquet_path, index=False, engine="pyarrow")
+    print(f"Wrote {parquet_path} ({parquet_path.stat().st_size:,} bytes)")
+except ImportError:
+    print("WARNING: pandas/pyarrow not available, skipping parquet")
+
+# Validation
+print("\n--- Validation ---")
+with open(json_path) as f:
+    check = json.load(f)
+print(f"Records: {len(check)}")
+print(f"Fields: {list(check[0].keys())}")
+print(f"Field count: {len(check[0])}")
+
+# Null counts in sample
+from collections import Counter
+nulls = Counter()
+for r in check:
+    for k, v in r.items():
+        if v is None:
+            nulls[k] += 1
+print(f"Null counts: {dict(nulls)}")
+
+# Type check pubchem_cid
+cid_types = set()
+for r in check:
+    if r["pubchem_cid"] is not None:
+        cid_types.add(type(r["pubchem_cid"]).__name__)
+print(f"pubchem_cid types: {cid_types}")
+
+print("\n✓ Sample generation complete")

llms.txt

@@ -1,10 +1,10 @@
 # Ethno-API — LLM Context File
-# Updated: 2026-04-01
+# Updated: 2026-04-12
 # Format: llms.txt (llmstxt.org standard)
 
 # PRODUCT
 > Ethno-API: USDA Phytochemical & Ethnobotanical 
-> Database — Enriched v2.3
+> Database — Enriched v2.3.1
 
 The only production-ready dataset combining USDA 
 Dr. Duke's phytochemical records with PubMed 
@@ -12,10 +12,10 @@ citation counts, ClinicalTrials.gov study counts,
 ChEMBL bioactivity measurements, USPTO patent 
 counts since 2020, and PubChem CID/canonical 
 SMILES. 76,907 records, 24,746 unique compounds, 
-2,313 plant species. JSON (25.6 MB) + Parquet 
-(1.2 MB). DOI: 10.5281/zenodo.19265853.
+2,313 plant species. JSON (28 MB) + Parquet 
+(1.5 MB). DOI: 10.5281/zenodo.19265853.
 
-# SCHEMA (v2.3, 10 columns)
+# SCHEMA (v2.3.1, 12 columns)
 - chemical: compound name (USDA Duke nomenclature)
 - plant_species: binomial Latin species name
 - application: traditional medicinal application
@@ -26,6 +26,8 @@ SMILES. 76,907 records, 24,746 unique compounds,
 - patent_count_since_2020: USPTO patents post-2020
 - pubchem_cid: PubChem Compound ID (75.4% coverage)
 - canonical_smiles: molecular structure (75.4%)
+- compound_type: compound classification (5 categories)
+- patent_count_method: query methodology transparency
 
 # ACCESS
 - Landing page: https://ethno-api.com
@@ -78,5 +80,5 @@ Full methodology: https://github.com/wirthal1990-tech/USDA-Phytochemical-Databas
 
 # CITATION
 Wirth, A. (2026). USDA Phytochemical & 
-Ethnobotanical Database — Enriched v2.3. Zenodo.
+Ethnobotanical Database — Enriched v2.3.1. Zenodo.
 https://doi.org/10.5281/zenodo.19265853

noise_exclusion_list.txt

@@ -1,10 +1,10 @@
-# Noise Exclusion List — Ethno-API v2.3
+# Noise Exclusion List — Ethno-API v2.3.1
 # 116 generic/non-phytochemical entries excluded from analytical queries.
 # These are elements, solvents, macronutrients, and generic categories
 # present in the USDA Dr. Duke source but not meaningful for phytochemical analysis.
 #
-# Dataset: v2.3 (76,907 records, 24,746 unique chemicals, 2,313 species)
-# Matching records in v2.3: 5,287 (93 of 116 entries found in dataset)
+# Dataset: v2.3.1 (76,907 records, 24,746 unique chemicals, 2,313 species)
+# Matching records in v2.3.1: 5,287 (93 of 116 entries found in dataset)
 # DOI: 10.5281/zenodo.19265853
 # Last updated: 2026-03-31
 ACETALDEHYDE

quickstart.ipynb

@@ -29,7 +29,7 @@
     "import duckdb\n",
     "\n",
     "# Load the free 400-row sample\n",
-    "# To use full dataset: replace path with 'ethno_dataset_2026_v2.3.json'\n",
+    "# To use full dataset: replace path with 'ethno_dataset_2026_v2.3.1.json'\n",
     "SAMPLE = 'ethno_sample_400.json'\n",
     "\n",
     "df = pd.read_json(SAMPLE)\n",
@@ -230,7 +230,7 @@
     "| Team | €1,349 netto | [**Buy →**](https://buy.stripe.com/dRm7sK9cP1Wj0v06Y0ebu03?utm_source=github&utm_medium=notebook&utm_campaign=launch_2026_03) |\n",
     "| Enterprise | €1,699 netto | [**Contact →**](mailto:founder@ethno-api.com) |\n",
     "\n",
-    "> Gemäß § 19 UStG keine Umsatzsteuer."
+    "> No VAT charged (German small business exemption, §19 UStG)."
    ]
   }
  ],