fix: move notebooks from root into notebooks/ folder, add READMEs and requirements

- Moved 5 .ipynb files from root → notebooks/
- notebooks/README.md: Hub-specific guide with notebook index, run order, and output persistence
- scripts/README.md: Pipeline guide with Pro vs basic fetcher comparison
- requirements.txt: Full dependency list for scripts and notebooks

notebooks/README.md

@@ -1,69 +1,86 @@
-# SafeVisionAI - Data Science & ML Pipelines
+# SafeVisionAI — Research Notebooks 📓
 
-This directory contains the original Google Colab / Jupyter notebooks utilized to train the machine learning models and process the massive datasets for the SafeVisionAI platform. 
+These notebooks are the **research and training layer** of SafeVisionAI. Each one processes raw data from the Hub and produces a model, index, or processed dataset used by the live application.
 
-The outputs from these notebooks form the entire "intelligence" layer of the application, driving everything from the real-time webcam inference to the RAG chatbot.
+---
+
+## ⚡ One-Click Colab Setup
+
+Run this at the top of **any** notebook to mount the full dataset:
+
+```python
+# Step 1 — Clone the Hub (only run once per session)
+!git clone https://huggingface.co/datasets/rohith083/SafeVisionAI-Dataset-Hub /content/hub
+
+# Step 2 — Install dependencies
+!pip install -q pdfplumber chromadb sentence-transformers ultralytics onnx
+
+# Step 3 — Confirm data is accessible
+import os
+print("Data folders:", os.listdir("/content/hub"))
+```
+
+---
+
+## 📓 Notebook Index
+
+| # | Notebook | What it Produces | Input Data |
+|---|---|---|---|
+| 1 | `YOLOv8_Pothole_Detector_Training` | ONNX road damage model | `pothole_training/road_damage_2025/` |
+| 2 | `ChromaDB_RAG_Vectorstore_Build` | ChromaDB index for legal RAG | `legal/`, `medical/` PDFs |
+| 3 | `Accident_EDA_&_Hotspot_Generator` | Blackspot seed CSV + heatmap | `accidents/kaggle_india_accidents.csv` |
+| 4 | `Roads_Data_Processing` | Sampled PMGSY GeoJSON | `roads/pmgsy_roads.geojson` |
+| 5 | `Risk_Model_ONNX_Training` | Risk scoring ONNX model | `accidents/` + `roads/` |
+
+---
+
+## 📁 Data Paths Used by These Notebooks
+
+All notebooks expect data at these paths after cloning the Hub:
+
+```
+/content/hub/
+├── chatbot_service/data/
+│   ├── accidents/              ← Notebooks 3, 5
+│   ├── legal/                  ← Notebook 2
+│   ├── medical/                ← Notebook 2
+│   ├── pothole_training/       ← Notebook 1
+│   └── roads/                  ← Notebook 4
+├── backend/datasets/           ← Notebook 4, 5
+└── frontend/public/models/     ← Output for Notebooks 1, 5
+```
 
 ---
 
-## 🚀 Getting Started in Colab (Fastest Way)
-
-1.  **Clone the Data Hub**: Since we moved the data to Hugging Face, you can clone the entire hub directly into your session:
-    ```bash
-    # Clone the Hub
-    !git clone https://huggingface.co/datasets/rohith083/SafeVisionAI-Dataset-Hub /content/dataset_hub
-    
-    # Link the data folder so notebooks find it at /content/data
-    !ln -s /content/dataset_hub/data /content/data
-    ```
-2.  **Extract Dataset Archives**: Some datasets are double-zipped. Run this to extract the pothole training data:
-    ```bash
-    !unzip /content/data/chatbot_service/data/pothole_training/road_damage_2025/archive.zip -d /content/data/chatbot_service/data/pothole_training/road_damage_2025/
-    ```
-
-## 📁 Dataset Path Architecture
-
-If cloned and linked correctly, the notebooks will automatically find data in:
-
-| Feature | Colab Path | Status |
-| :--- | :--- | :--- |
-| **Accidents Data** | `/content/data/chatbot_service/data/accidents/` | Ready |
-| **Legal/Medical PDFs** | `/content/data/chatbot_service/data/legal/*.pdf` | Ready |
-| **Pothole Training** | `/content/data/chatbot_service/data/pothole_training/road_damage_2025/` | **Needs Extract** |
-| **Roads & Tolls** | `/content/data/backend/data/roads/` | Ready |
-| **AI Models (ONNX)** | `/content/data/frontend/public/models/` | Ready |
+## 🏃 Running Order
+
+For a full pipeline run, execute notebooks in this order:
+
+```
+1 → Train pothole detector → produces ONNX model
+2 → Build RAG vectorstore  → produces ChromaDB index
+3 → Accident EDA           → produces blackspot_seed.csv
+4 → Roads processing       → produces sampled GeoJSON
+5 → Risk model             → combines accidents + roads → risk ONNX
+```
 
 ---
 
+## 💾 Saving Outputs Back
+
+After training, outputs land in `/content/hub/`. To persist them:
+
+```python
+# Commit outputs back to Hub (requires HF token)
+import os
+os.environ["HUGGING_FACE_TOKEN"] = "your_token_here"
+
+!cd /content/hub && git config user.email "you@example.com"
+!cd /content/hub && git config user.name "SafeVisionAI"
+!cd /content/hub && git add . && git commit -m "chore: update trained model outputs"
+!cd /content/hub && git push https://your_username:$HUGGING_FACE_TOKEN@huggingface.co/datasets/rohith083/SafeVisionAI-Dataset-Hub main
+```
+
+---
 
-## 📓 Notebook 1: YOLOv8 Pothole Detector Training
-* **Purpose**: Trains a custom YOLOv8 Nano object detection model on the Indian Road Pothole dataset to identify multiple classes: `pothole`, `crack`, and `manhole`. Employs data augmentation and runs for 50 epochs to maximize mAP.
-* **Why we use it**: It acts as the core computer vision engine to warn drivers of upcoming road deformities.
-* **Where the outputs live**:
-  * ➔ `frontend/public/models/pothole.onnx` (12MB ONNX artifact used by the browser/app to perform real-time tracking entirely offline).
-  * ➔ `backend/models/pothole.pt` (PyTorch artifact utilized by the backend server for heavier image verification/reporting).
-
-## 📓 Notebook 2: ChromaDB RAG Vectorstore Build
-* **Purpose**: Ingests thousands of pages of Indian legal Pdfs, MVA (Motor Vehicles Act) penalty charts, and traffic regulations. It utilizes `langchain` and SentenceTransformers to chunk, embed, and store this data in a persistent local SQLite vector store.
-* **Why we use it**: It gives the AI Chatbot absolute knowledge over Indian road laws, ensuring it never hallucinates legal advice.
-* **Where the outputs live**:
-  * ➔ `chatbot_service/data/chroma_db/` (The actual `chroma.sqlite3` vectors used by the LangChain retrieval chain).
-
-## 📓 Notebook 3: Accident EDA & Hotspot Generator
-* **Purpose**: Analyzes the multi-million row India Accidents GPS dataset. It calculates national statistics, aggregates state-wise fatality rates, and clusters raw GPS tags into definitive geographic "blackspots" (areas with historical fatality density).
-* **Why we use it**: Displays compelling statistics to the user and warns them if their current route is historically dangerous.
-* **Where the outputs live**:
-  * ➔ `chatbot_service/data/accidents/blackspot_seed.csv` (Parsed by the SOS and routing engine).
-  * ➔ `chatbot_service/data/accidents/accidents_summary.json` & `frontend/public/offline-data/accidents_summary.json` (Used by the UI panels and chatbot context).
-
-## 📓 Notebook 4: Roads Data Processing
-* **Purpose**: Cleans and strips down massive tabular highway/toll data. Maps unformatted legacy headers (`name, lat, lon, id`) into clean schemas (`Plaza Name, NH Number`) to reduce the file payload size for mobile clients.
-* **Why we use it**: Prevents the frontend from downloading megabytes of unused CSV columns just to render map markers.
-* **Where the outputs live**:
-  * ➔ `backend/data/toll_plazas.json` (Served directly to the Mapbox/Leaflet UI components).
-
-## 📓 Notebook 5: Risk Model ONNX Training
-* **Purpose**: Synthesizes environmental data (e.g., weather condition, speed limits, time of day, road type) to train a lightweight `GradientBoostingClassifier` evaluating the current risk probability (Safe vs. Danger) of a drive.
-* **Why we use it**: Generates an ultra-lightweight (15KB) classification model capable of running edge inferences on the user's phone, even in poor-connectivity standard Indian rural areas.
-* **Where the outputs live**:
-  * ➔ `frontend/public/models/risk_model.onnx` (Triggered via ONNX.js in the browser frontend).
+*Part of the [SafeVisionAI](https://github.com/SafeVision-AI/SafeVision-AI) — IIT Madras Road Safety Hackathon 2026*

requirements.txt

@@ -0,0 +1,25 @@
+# SafeVisionAI Dataset Hub — Script Dependencies
+# Install with: pip install -r requirements.txt
+
+# ── Core (all scripts) ────────────────────────────────────────
+requests>=2.31.0          # HTTP fetching for legal PDF downloader
+
+# ── GIS / Overpass fetchers ───────────────────────────────────
+# No extra deps — fetch_*.py uses only stdlib (urllib, csv, json)
+
+# ── PDF Processing ────────────────────────────────────────────
+pdfplumber>=0.10.0        # extract_morth2022_tables.py
+
+# ── Kaggle Dataset Download ───────────────────────────────────
+kaggle>=1.6.0             # setup_kaggle.ps1 + accident CSV download
+
+# ── Notebook dependencies ─────────────────────────────────────
+ultralytics>=8.0.0        # YOLOv8 pothole detector training
+onnx>=1.15.0              # ONNX model export
+chromadb>=0.4.0           # RAG vectorstore build
+sentence-transformers>=2.2.0  # ChromaDB embeddings
+scikit-learn>=1.3.0       # Risk model training
+pandas>=2.0.0             # Data processing
+geopandas>=0.14.0         # Roads GeoJSON processing
+matplotlib>=3.7.0         # Accident EDA visualizations
+seaborn>=0.12.0           # Heatmap generation

scripts/README.md

@@ -0,0 +1,93 @@
+# SafeVisionAI — Data Acquisition Scripts 🔬
+
+These scripts are the **raw data pipeline** that built the 3.6GB SafeVisionAI dataset.
+All scripts here are **pure Python** — they require no database, no Redis, no PostGIS.
+
+> Scripts are mirrored from the [SafeVisionAI main repo](https://github.com/SafeVision-AI/SafeVision-AI) and organized by their origin folder.
+
+---
+
+## 📁 Structure
+
+```
+scripts/
+├── scripts/data/             ← from SafeVisionAI/scripts/data/
+│   ├── fetch_*.py            ← Overpass GIS fetchers (basic version)
+│   ├── _overpass_utils.py    ← Core GIS utility
+│   ├── download_legal_pdfs.py
+│   ├── extract_morth2022_tables.py
+│   ├── seed_blackspots.py
+│   ├── bootstrap_local_data.py
+│   ├── verify_data.py
+│   ├── inspect_zips.py
+│   ├── audit_env.py
+│   ├── check_all_scripts.py
+│   └── setup_kaggle.ps1
+│
+├── backend/data/             ← from SafeVisionAI/backend/scripts/data/
+│   ├── seed_violations.py    ← MVA 2019 traffic fine normalizer
+│   ├── prepare_road_sources.py
+│   ├── sample_pmgsy.py
+│   └── road_sources.example.json
+│
+└── chatbot_service/data/     ← from SafeVisionAI/chatbot_service/scripts/data/
+    ├── _overpass_utils.py    ← ⭐ Pro version (retries + backoff)
+    └── fetch_*.py            ← ⭐ Pro fetchers (use these over scripts/data/)
+```
+
+---
+
+## 🚀 Recommended Run Order
+
+```bash
+# Install dependencies first
+pip install requests pdfplumber
+
+# 1. Setup Kaggle API (one-time)
+pwsh scripts/scripts/data/setup_kaggle.ps1
+
+# 2. Fetch all emergency services (use Pro versions)
+python scripts/chatbot_service/data/fetch_hospitals.py
+python scripts/chatbot_service/data/fetch_police.py
+python scripts/chatbot_service/data/fetch_ambulance.py
+python scripts/chatbot_service/data/fetch_blood_banks.py
+python scripts/chatbot_service/data/fetch_fire.py
+
+# 3. Download legal documents
+python scripts/scripts/data/download_legal_pdfs.py
+
+# 4. Extract MoRTH accident tables from PDFs
+python scripts/scripts/data/extract_morth2022_tables.py
+
+# 5. Normalize accident blackspot data
+python scripts/scripts/data/seed_blackspots.py
+
+# 6. Normalize traffic violations (MVA 2019)
+python scripts/backend/data/seed_violations.py
+
+# 7. Verify everything is correct
+python scripts/scripts/data/verify_data.py
+```
+
+---
+
+## ⭐ Pro vs Basic Fetchers
+
+| Feature | `scripts/scripts/data/fetch_*.py` | `scripts/chatbot_service/data/fetch_*.py` |
+|---|---|---|
+| Retry logic | ❌ | ✅ |
+| Exponential backoff | ❌ | ✅ |
+| Indian GIS precision | Basic | Enhanced |
+| Extra fields (email, postcode) | ❌ | ✅ |
+
+**Always prefer `chatbot_service/data/` versions** for data acquisition.
+
+---
+
+## 📦 Dependencies
+
+```
+pip install -r requirements.txt
+```
+
+See `requirements.txt` in the Hub root.