README

patankar/cache.ToxTree/README.md

@@ -0,0 +1,35 @@
+# ToxTree Cache for SFPPy 🍏⏩🍎
+
+This directory is managed by `patankar.private.loadpubchem.migrantToxtree` and serves as a **local cache** for toxicological assessments performed using **Toxtree**. It helps optimize repeated evaluations by storing results locally, allowing for faster access usage (even on machines without ToxTree).
+
+## 📁 Folder Contents
+
+- `cidXXXX.default.csv` 📄: **Base Toxtree evaluation** (always generated for each compound).
+- `cidXXXX.engine.csv` 📑: Engine-specific Toxtree outputs, where `engine` corresponds to:
+  - `cramer`
+  - `cramer2`
+  - `cramer3`
+  - `kroes`
+  - `dnabinding`
+  - `skin`
+  - `eye`
+  - `Ames`
+- `cidXXXX.engine.json` 📂: Processed **JSON versions** of the Toxtree outputs for faster lookup.
+- `structure/cidXXXX.sdf` 🧱: Cached **SDF structure file** used as input for Toxtree.
+- `structure/cidXXXX.png` 🎨: **Cropped PNG image** of the molecule structure.
+
+## 🔹 How It Works
+
+1. When a compound is assessed, `migrantToxtree` first checks if **JSON results exist**.
+2. If found, data is retrieved directly from JSON (fastest lookup).
+3. If missing, `Toxtree` is executed to generate a **CSV file**.
+4. The CSV results are then **converted into JSON** and stored for future use.
+5. If structure files are not available, they are downloaded from **PubChem** and stored in `structure/`.
+
+## ⚠️ Notes
+
+- **Do not manually modify or delete files** unless necessary, as `migrantToxtree` automatically manages the cache.
+- To force an update, use the `refresh=True` flag when initializing `migrantToxtree`.
+- To fully regenerate all files, use `no_cache=True` to overwrite existing data.
+
+🚀 **Using local caching improves performance and allows for seamless offline toxicological assessments in SFPPy.**