Production deployment with LFS models

.dockerignore

@@ -0,0 +1,9 @@
+venv/
+env/
+.env
+backend/chroma_db/
+backend/__pycache__/
+middleware/__pycache__/
+frontend/node_modules/
+frontend/dist/
+.git/

.gitattributes

@@ -0,0 +1,8 @@
+*.db filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.cache filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,28 @@
+# Virtual Environment
+venv/
+env/
+.env
+
+# Python Caches
+__pycache__/
+*.py[cod]
+*$py.class
+
+# Large data & model files
+data/
+runs/
+*.cache
+*.db-journal
+*.sqlite3-journal
+
+# Keep these for production
+!middleware/wafer_control.db
+!middleware/material_model.pkl
+!middleware/best.pt
+
+
+# OS generated files
+.DS_Store
+
+# Vector DB
+backend/chroma_db/

Dockerfile

@@ -0,0 +1,21 @@
+FROM python:3.12-slim
+
+WORKDIR /app
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y build-essential && rm -rf /var/lib/apt/lists/*
+
+# Copy and install dependencies
+COPY backend/requirements.txt ./backend/
+RUN pip install --no-cache-dir -r backend/requirements.txt
+
+# Copy everything needed
+COPY backend/ ./backend/
+COPY middleware/ ./middleware/
+COPY runs/ ./runs/
+
+# Hugging Face requires port 7860
+EXPOSE 7860
+
+# Start the server
+CMD ["uvicorn", "backend.main:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -0,0 +1,87 @@
+# Semiconductor Wafer Defect Detection: End-to-End AI Pipeline
+
+## Project Overview
+This project is a complete, end-to-end Applied AI pipeline designed for the semiconductor manufacturing industry. It takes raw mathematical array data representing defective semiconductor wafers, engineers them into an AI-ready computer vision dataset, trains a custom YOLOv8 object detection model, and feeds the results into a predictive material waste model and real-time dashboard.
+
+**Final YOLOv8 Model Performance:** `0.962 mAP@50` (96.2% overall accuracy on unseen validation data).
+**Predictive Waste Model Performance:** `R² = 0.9637` (Highly accurate material waste prediction).
+
+## Business Value
+In semiconductor fabrication, identifying microscopic defects early in the manufacturing process saves millions in scrapped materials. This project automates quality control by transitioning from manual coordinate analysis to real-time, AI-driven visual defect detection, while simultaneously forecasting future material waste to optimize supply chain planning.
+
+## The Technical Pipeline
+
+### Phase 1: Data Engineering (`src/data_prep.py`)
+* **The Challenge:** The original dataset consisted of raw `.txt` files containing numeric 2D arrays (0=background, 1=good chip, 2=defect). YOLOv8 cannot read text arrays; it requires physical images and normalized bounding box coordinates.
+* **The Solution:** Built a custom Python pipeline using `NumPy` and `OpenCV` to parse over 25,000 text files. 
+* **The Math:** Programmatically identified the spatial extremes (`xmin`, `ymin`, `xmax`, `ymax`) of the `2` values, normalized them to YOLO's strict `0.0 - 1.0` format, and dynamically rendered high-contrast `.jpg` images alongside corresponding `.txt` label files.
+
+### Phase 2: Dataset Architecture (`src/split_data.py`)
+* Used `scikit-learn` to execute a mathematically rigorous 80/20 train/validation split.
+* Programmatically generated the strict directory architecture required by YOLO, migrating over 50,000 individual files into structured `train` and `val` directories.
+
+### Phase 3: Model Training (`src/model_train.py`)
+* Initialized a pre-trained **YOLOv8 Nano** (`yolov8n.pt`) model for lightweight, high-speed inference.
+* Trained on 20,415 wafer images for 10 epochs.
+* Mapped 8 specific manufacturing defect classes (Center, Donut, Edge-Loc, Edge-Ring, Loc, Random, Scratch, Near-full).
+
+### Phase 4: Batch Inference & Evaluation (`src/batch_inference.py` & `src/model_eval.py`)
+* Deployed the custom-trained `best.pt` weights to run batch inference on unseen validation images.
+* Model successfully drew accurate bounding boxes and assigned confidence scores entirely autonomously.
+
+### Phase 5: Production Middleware, Predictive Modeling & Dashboard
+* **Robotic Scanner Simulation (`middleware/robot_controller.py`):** Operates on a massive hybrid dataset of **823,953 wafers** (Mixed-type + WM-811K datasets) with a realistic 95.5% pass rate. It automatically routes passed wafers and runs YOLOv8 inference on defective ones, logging everything into a centralized SQLite database (`wafer_control.db`).
+* **Material Waste Predictor (`middleware/material_predictor.py`):** A Random Forest Regressor trained on the historical scan database. It accurately predicts the average percentage of material wasted within defective wafers, allowing fabs to estimate future material needs.
+* **Real-time Dashboard (`middleware/dashboard.py`):** A **Plotly Dash** web application that visualizes historical defect rates, defect distributions, routing actions, and integrates interactive material forecasting inputs.
+
+## Upcoming Feature: LLM Troubleshooting Assistant (Planned)
+**Goal:** Integrate an intelligent Large Language Model (LLM) bot to assist fab engineers directly on the factory floor.
+* **Functionality:** When the dashboard flags a sudden spike in a specific defect type (e.g., "Edge-Ring" defects), the engineer can consult the LLM bot.
+* **Use Case:** The bot will analyze the defect trends, cross-reference historical manufacturing guidelines, and suggest potential root causes (such as misaligned etching tools or incorrect gas pressure), drastically reducing troubleshooting and downtime.
+*(Note: This feature is currently in the design phase and not yet implemented).*
+
+## Performance Metrics
+The YOLOv8 model achieved phenomenal results on the blind validation set:
+
+| Metric | Score | Note |
+| :--- | :--- | :--- |
+| **mAP50 (All Classes)** | **96.2%** | Overall model accuracy at a 50% confidence threshold. |
+| **Recall** | **93.1%** | The model successfully located 93.1% of all physical defects. |
+| **Edge-Ring (mAP50)** | **99.4%** | Near-flawless detection of Edge-Ring anomalies. |
+
+The Random Forest Material Waste Predictor achieved:
+| Metric | Score | Note |
+| :--- | :--- | :--- |
+| **R² Score** | **0.9637** | Excellent correlation on predictive targets. |
+| **MAE** | **0.09%** | Average prediction error is less than one-tenth of a percent. |
+
+## Tech Stack
+* **Languages:** Python
+* **Computer Vision:** Ultralytics (YOLOv8), OpenCV (`cv2`)
+* **Machine Learning & Data:** Pandas, NumPy, Scikit-learn, SQLite
+* **Web UI & Visualization:** Plotly, Dash
+
+## Deployment (Docker)
+
+This application is fully containerized for easy deployment.
+
+1.  **Clone the repository:**
+    ```bash
+    git clone https://github.com/Udayan2001/Semiconductor_defect_detection.git
+    cd Semiconductor_defect_detection
+    ```
+2.  **Add API Key:**
+    Create a `.env` file in the `backend/` directory and add your Google Gemini API key:
+    ```
+    GEMINI_API_KEY=your_api_key_here
+    ```
+3.  **Start the Application:**
+    Run the following command from the root directory to build and start both the backend and frontend servers:
+    ```bash
+    docker compose up --build
+    ```
+4.  **Access the Dashboard:**
+    Open your browser and navigate to `http://localhost:5173`.
+
+---
+*Designed and engineered by Udayan Shashank Shukla.*

backend/Dockerfile

@@ -0,0 +1,19 @@
+FROM python:3.12-slim
+
+WORKDIR /app
+
+# Install essential system dependencies
+RUN apt-get update && apt-get install -y build-essential && rm -rf /var/lib/apt/lists/*
+
+# Copy and install python dependencies
+COPY backend/requirements.txt ./backend/
+RUN pip install --no-cache-dir -r backend/requirements.txt
+
+# Copy backend and middleware code
+COPY backend/ ./backend/
+COPY middleware/ ./middleware/
+
+EXPOSE 8000
+
+# Run ingestion first to ensure vector DB is seeded, then start the server
+CMD ["sh", "-c", "python backend/ingest_knowledge.py && uvicorn backend.main:app --host 0.0.0.0 --port 8000"]

backend/ingest_knowledge.py

@@ -0,0 +1,65 @@
+import os
+import chromadb
+from chromadb.config import Settings
+
+# Engineering knowledge base regarding semiconductor wafer defects
+KNOWLEDGE_BASE = [
+    {
+        "id": "defect_edge_ring",
+        "title": "Edge-Ring Defect Troubleshooting",
+        "content": "Edge-Ring defects typically appear as a continuous ring of failing dies around the outer edge of the wafer. Common Root Causes: 1. Uneven gas distribution in the etching chamber. 2. Non-uniform chuck temperature during deposition or etching. 3. Edge-bead removal issues during photolithography. Recommended Action: Inspect gas flow regulators and recalibrate chuck temperature sensors. Schedule maintenance for edge-bead removal module."
+    },
+    {
+        "id": "defect_center",
+        "title": "Center Defect Troubleshooting",
+        "content": "Center defects are concentrated in the middle of the wafer. Common Root Causes: 1. Poor spin-coating uniformity (photoresist pooling in the center). 2. Center-heavy deposition profile. 3. Excessive center heating on the electrostatic chuck. Recommended Action: Verify spin speed and acceleration in the coating track. Check gas showerhead for clogging in the center region."
+    },
+    {
+        "id": "defect_scratch",
+        "title": "Scratch Defect Troubleshooting",
+        "content": "Scratch defects manifest as linear patterns of failing dies, often crossing the wafer. Common Root Causes: 1. Mechanical handling damage by robotic arms or end-effectors. 2. Particulate contamination causing dragging during CMP (Chemical Mechanical Polishing). 3. Cassette or FOUP abrasion. Recommended Action: Check robot alignment and end-effector cleanliness. Inspect CMP pad conditioning and slurry filtration system."
+    },
+    {
+        "id": "defect_donut",
+        "title": "Donut Defect Troubleshooting",
+        "content": "Donut defects appear as a ring, but not at the very edge (like Edge-Ring), leaving the center and extreme edge relatively clean. Common Root Causes: 1. Radially dependent temperature non-uniformity during rapid thermal processing (RTP). 2. Specific gas flow dynamics creating standing waves or depletion zones in the chamber. Recommended Action: Recalibrate RTP lamp zones. Inspect gas showerhead and exhaust pumping symmetry."
+    },
+    {
+        "id": "general_forecast_strategy",
+        "title": "Material Forecast & Yield Strategy",
+        "content": "When the predicted material waste percentage rises above 5%, the factory must proactively increase raw material orders (wafers, photoresist, precursor gases) for the next quarter to compensate for the lower yield. High fail rates typically necessitate a temporary slow-down of production throughput to allow for deep tool maintenance and recalibration."
+    }
+]
+
+def ingest_data():
+    print("Initializing ChromaDB Persistent Client...")
+    db_path = os.path.join(os.path.dirname(__file__), "chroma_db")
+    client = chromadb.PersistentClient(path=db_path)
+    
+    # Create or get collection
+    collection = client.get_or_create_collection(
+        name="semiconductor_knowledge",
+        metadata={"hnsw:space": "cosine"}
+    )
+    
+    # Clear existing data if any (for idempotency)
+    existing_ids = collection.get()['ids']
+    if existing_ids:
+        collection.delete(ids=existing_ids)
+    
+    # Prepare data for insertion
+    ids = [item['id'] for item in KNOWLEDGE_BASE]
+    documents = [item['content'] for item in KNOWLEDGE_BASE]
+    metadatas = [{"title": item['title']} for item in KNOWLEDGE_BASE]
+    
+    print(f"Adding {len(documents)} documents to the knowledge base...")
+    collection.add(
+        documents=documents,
+        metadatas=metadatas,
+        ids=ids
+    )
+    
+    print("Ingestion complete. ChromaDB is ready.")
+
+if __name__ == "__main__":
+    ingest_data()

backend/main.py

@@ -0,0 +1,257 @@
+import sys
+import os
+import pickle
+import sqlite3
+import pandas as pd
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+from dotenv import load_dotenv
+from google import genai
+import chromadb
+from typing import List, Dict
+
+env_path = os.path.join(os.path.dirname(__file__), '.env')
+load_dotenv(env_path)
+
+# Add parent dir to path so we can import from middleware
+sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
+from middleware.material_predictor import predict_material_needs
+
+app = FastAPI(title="Wafer Defect API")
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# Robust paths for Docker/Hosting
+BASE_DIR = os.path.dirname(os.path.abspath(__file__))
+DB_PATH = os.path.join(BASE_DIR, '..', 'middleware', 'wafer_control.db')
+MODEL_PATH = os.path.join(BASE_DIR, '..', 'middleware', 'material_model.pkl')
+CHROMA_PATH = os.path.join(BASE_DIR, 'chroma_db')
+
+# Ensure directories exist
+os.makedirs(CHROMA_PATH, exist_ok=True)
+
+DEFECT_COLORS = {
+    'Center': '#ef4444', 'Donut': '#f59e0b', 'Edge-Loc': '#10b981',
+    'Edge-Ring': '#3b82f6', 'Loc': '#8b5cf6', 'Random': '#ec4899',
+    'Scratch': '#06b6d4', 'Near-full': '#f97316', 'None': '#6b7280',
+    'Undetected': '#374151',
+}
+
+# Globally load data so we don't block requests
+df = pd.DataFrame()
+if os.path.exists(DB_PATH):
+    print(f"Loading DB from {DB_PATH}...")
+    conn = sqlite3.connect(DB_PATH)
+    df = pd.read_sql_query("SELECT * FROM wafer_logs", conn)
+    conn.close()
+    df['scan_time'] = pd.to_datetime(df['scan_time'])
+    df['scan_date'] = df['scan_time'].dt.date
+else:
+    print(f"Warning: DB not found at {DB_PATH}. Dashboard will be empty.")
+
+# Setup Vector DB and LLM
+print(f"Connecting to ChromaDB at {CHROMA_PATH}...")
+try:
+    chroma_client = chromadb.PersistentClient(path=CHROMA_PATH)
+    collection = chroma_client.get_or_create_collection(name="semiconductor_knowledge")
+except Exception as e:
+    print(f"Warning: Could not connect to ChromaDB collection. Error: {e}")
+    collection = None
+
+print("Initializing Gemini API...")
+gemini_client = None
+if os.getenv("GEMINI_API_KEY"):
+    gemini_client = genai.Client(api_key=os.getenv("GEMINI_API_KEY"))
+else:
+    print("Warning: GEMINI_API_KEY not found in environment.")
+
+print("Loading ML model...")
+model_pkg = None
+if os.path.exists(MODEL_PATH):
+    with open(MODEL_PATH, 'rb') as f:
+        model_pkg = pickle.load(f)
+
+@app.get("/api/kpi")
+def get_kpis():
+    total_scans = len(df)
+    fail_df = df[df['status'] == 'FAIL']
+    fail_count = len(fail_df)
+    pass_count = len(df[df['status'] == 'PASS'])
+    pass_rate = round((pass_count / total_scans) * 100, 1) if total_scans else 0
+    scrap_count = len(df[df['action'] == 'ROUTE_TO_SCRAP'])
+    avg_waste = round(fail_df['material_wasted_pct'].mean(), 2) if fail_count else 0
+    avg_confidence = round(fail_df['confidence'].mean(), 2) if fail_count else 0
+    
+    return {
+        "total_scans": total_scans,
+        "pass_count": pass_count,
+        "pass_rate": pass_rate,
+        "fail_count": fail_count,
+        "fail_rate": round(100 - pass_rate, 1),
+        "scrap_count": scrap_count,
+        "avg_waste": avg_waste,
+        "avg_confidence": avg_confidence
+    }
+
+@app.get("/api/charts/defects")
+def get_defects():
+    fail_df = df[df['status'] == 'FAIL']
+    defect_counts = fail_df['defect_type'].value_counts().reset_index()
+    defect_counts.columns = ['defect_type', 'count']
+    
+    gt_counts = fail_df['ground_truth'].value_counts().reset_index()
+    gt_counts.columns = ['ground_truth', 'count']
+    
+    return {
+        "predictions": defect_counts.to_dict(orient="records"),
+        "ground_truth": gt_counts.head(15).to_dict(orient="records")
+    }
+
+@app.get("/api/charts/waste")
+def get_waste():
+    fail_df = df[df['status'] == 'FAIL']
+    
+    waste_by_type = fail_df.groupby('defect_type').agg(
+        total_waste=('material_wasted_pct', lambda x: x.sum() / 100.0)
+    ).reset_index().sort_values('total_waste', ascending=True)
+    
+    action_counts = df['action'].value_counts().reset_index()
+    action_counts.columns = ['action', 'count']
+    
+    return {
+        "waste_by_type": waste_by_type.to_dict(orient="records"),
+        "actions": action_counts.to_dict(orient="records")
+    }
+
+@app.get("/api/charts/trends")
+def get_trends():
+    daily = df.groupby('scan_date').agg(
+        scans=('id', 'count'),
+        fails=('status', lambda x: (x == 'FAIL').sum()),
+        waste=('material_wasted_pct', lambda x: x.sum() / 100.0)
+    ).reset_index()
+    daily['fail_rate'] = round((daily['fails'] / daily['scans']) * 100, 1)
+    
+    return {
+        "dates": daily['scan_date'].astype(str).tolist(),
+        "fail_rate": daily['fail_rate'].tolist(),
+        "waste": daily['waste'].tolist()
+    }
+
+@app.get("/api/model/status")
+def model_status():
+    if not model_pkg:
+        return {"loaded": False}
+    
+    m = model_pkg['metrics']
+    imp = model_pkg['metrics']['importances']
+    imp_df = pd.DataFrame({'feature': list(imp.keys()), 'importance': list(imp.values())})
+    imp_df = imp_df.sort_values('importance', ascending=True).tail(10)
+    
+    return {
+        "loaded": True,
+        "metrics": {"r2": round(m['r2'], 4), "mae": round(m['mae'], 2)},
+        "importance": imp_df.to_dict(orient="records")
+    }
+
+class PredictionRequest(BaseModel):
+    scans: int
+    fail_rate: float
+
+@app.post("/api/predict")
+def predict_waste(req: PredictionRequest):
+    if not model_pkg:
+        return {"error": "No model loaded"}
+        
+    fail_df = df[df['status'] == 'FAIL']
+    dist = fail_df['defect_type'].value_counts(normalize=True).to_dict()
+    
+    pred = predict_material_needs(model_pkg['model'], model_pkg['feature_cols'], req.scans, req.fail_rate / 100.0, dist)
+    pred['fail_rate'] = req.fail_rate
+    return pred
+
+
+class ChatMessage(BaseModel):
+    role: str
+    content: str
+
+class ChatRequest(BaseModel):
+    messages: List[ChatMessage]
+
+@app.post("/api/chat")
+def chat_with_bot(req: ChatRequest):
+    if not gemini_client:
+        return {"error": "Gemini API key not configured"}
+        
+    user_message = req.messages[-1].content if req.messages else ""
+    
+    # 1. RAG Retrieval from ChromaDB
+    context_docs = ""
+    if collection and user_message:
+        try:
+            results = collection.query(query_texts=[user_message], n_results=2)
+            if results and results['documents'] and results['documents'][0]:
+                context_docs = "\n".join(results['documents'][0])
+        except Exception as e:
+            print(f"ChromaDB Query Error: {e}")
+            
+    # 2. Get Live Dashboard Context
+    total_scans = len(df)
+    fail_df = df[df['status'] == 'FAIL']
+    fail_count = len(fail_df)
+    pass_rate = round(((total_scans - fail_count) / total_scans) * 100, 1) if total_scans else 0
+    top_defects = fail_df['defect_type'].value_counts().head(3).to_dict()
+    
+    live_kpis = f"""
+    Current Dashboard State:
+    - Total Wafers Scanned: {total_scans}
+    - Current Pass Rate: {pass_rate}%
+    - Total Defective Wafers: {fail_count}
+    - Top Defect Types Right Now: {top_defects}
+    """
+    
+    # 3. Construct System Prompt
+    system_instruction = f"""
+    You are the 'Gorilla Semiconductors Engineering Assistant', an expert semiconductor manufacturing assistant. 
+    You help engineers understand dashboard data and troubleshoot wafer defects.
+    Maintain a strictly professional, analytical, and authoritative engineering tone.
+    
+    Here is the LIVE DATA from the dashboard:
+    {live_kpis}
+    
+    Here is retrieved technical context from our engineering database based on the user's query:
+    {context_docs if context_docs else "No specific engineering docs retrieved."}
+    
+    Use the live data to answer questions about 'current status' or 'dashboard'. 
+    Use the engineering docs to answer questions about 'why' a defect happens.
+    """
+    
+    try:
+        # Convert messages to format expected by google-genai
+        contents = []
+        for msg in req.messages:
+            role = "user" if msg.role == "user" else "model"
+            contents.append(
+                genai.types.Content(role=role, parts=[genai.types.Part.from_text(text=msg.content)])
+            )
+            
+        response = gemini_client.models.generate_content(
+            model='gemini-2.5-flash-lite',
+            contents=contents,
+            config=genai.types.GenerateContentConfig(
+                system_instruction=system_instruction,
+                temperature=0.3
+            )
+        )
+        return {"response": response.text}
+    except Exception as e:
+        print(f"Gemini API Error: {e}")
+        return {"error": str(e)}
+

backend/requirements.txt

@@ -0,0 +1,8 @@
+fastapi>=0.100.0
+uvicorn>=0.22.0
+pandas>=2.0.0
+scikit-learn>=1.3.0
+pydantic>=2.0.0
+google-genai>=0.3.0
+chromadb>=0.4.24
+python-dotenv>=1.0.0

dataset.yaml

@@ -0,0 +1,19 @@
+# YOLOv8 Dataset Configuration File
+
+# The base path to your dataset folder
+path: data/yolo_dataset
+
+# The subfolders for training and validation images
+train: images/train
+val: images/val
+
+# The 8 defect classes we mapped earlier
+names:
+  0: Center
+  1: Donut
+  2: Edge-Loc
+  3: Edge-Ring
+  4: Loc
+  5: Random
+  6: Scratch
+  7: Near-full

docker-compose.yml

@@ -0,0 +1,23 @@
+services:
+  backend:
+    build:
+      context: .
+      dockerfile: backend/Dockerfile
+    ports:
+      - "8000:8000"
+    # Ensure the container has access to the environment variables
+    env_file:
+      - ./backend/.env
+    # Optional: Mount the SQLite DB so changes persist
+    volumes:
+      - ./middleware/wafer_control.db:/app/middleware/wafer_control.db
+
+  frontend:
+    build:
+      context: ./frontend
+      dockerfile: Dockerfile
+    ports:
+      # Map the Nginx internal port 80 to 5173 to match the dev environment
+      - "5173:80"
+    depends_on:
+      - backend

frontend/.dockerignore

@@ -0,0 +1,3 @@
+node_modules/
+dist/
+.env

frontend/.gitignore

@@ -0,0 +1,24 @@
+# Logs
+logs
+*.log
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+pnpm-debug.log*
+lerna-debug.log*
+
+node_modules
+dist
+dist-ssr
+*.local
+
+# Editor directories and files
+.vscode/*
+!.vscode/extensions.json
+.idea
+.DS_Store
+*.suo
+*.ntvs*
+*.njsproj
+*.sln
+*.sw?

frontend/Assets/Gorilla_Chest_Thumping_Animation_Generated.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95a1c3d740e86a5ec1c9ef8f26063000280eb1219d173ecbc2c4ca6c3d5ccbe8
+size 945018

frontend/Assets/freesound_community-monkey-30631.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4960fd96e5623ecc3f90c262e78e1f63ba7bb01e619cc52104e59a80fbc49c16
+size 429600

frontend/Dockerfile

@@ -0,0 +1,20 @@
+# Build stage
+FROM node:20-alpine AS builder
+
+WORKDIR /app
+
+# Install dependencies
+COPY package*.json ./
+RUN npm install
+
+# Copy source code and build
+COPY . .
+RUN npm run build
+
+# Production stage
+FROM nginx:alpine
+# Copy built assets to Nginx
+COPY --from=builder /app/dist /usr/share/nginx/html
+
+EXPOSE 80
+CMD ["nginx", "-g", "daemon off;"]

frontend/README.md

@@ -0,0 +1,16 @@
+# React + Vite
+
+This template provides a minimal setup to get React working in Vite with HMR and some ESLint rules.
+
+Currently, two official plugins are available:
+
+- [@vitejs/plugin-react](https://github.com/vitejs/vite-plugin-react/blob/main/packages/plugin-react) uses [Oxc](https://oxc.rs)
+- [@vitejs/plugin-react-swc](https://github.com/vitejs/vite-plugin-react/blob/main/packages/plugin-react-swc) uses [SWC](https://swc.rs/)
+
+## React Compiler
+
+The React Compiler is not enabled on this template because of its impact on dev & build performances. To add it, see [this documentation](https://react.dev/learn/react-compiler/installation).
+
+## Expanding the ESLint configuration
+
+If you are developing a production application, we recommend using TypeScript with type-aware lint rules enabled. Check out the [TS template](https://github.com/vitejs/vite/tree/main/packages/create-vite/template-react-ts) for information on how to integrate TypeScript and [`typescript-eslint`](https://typescript-eslint.io) in your project.

frontend/eslint.config.js

@@ -0,0 +1,29 @@
+import js from '@eslint/js'
+import globals from 'globals'
+import reactHooks from 'eslint-plugin-react-hooks'
+import reactRefresh from 'eslint-plugin-react-refresh'
+import { defineConfig, globalIgnores } from 'eslint/config'
+
+export default defineConfig([
+  globalIgnores(['dist']),
+  {
+    files: ['**/*.{js,jsx}'],
+    extends: [
+      js.configs.recommended,
+      reactHooks.configs.flat.recommended,
+      reactRefresh.configs.vite,
+    ],
+    languageOptions: {
+      ecmaVersion: 2020,
+      globals: globals.browser,
+      parserOptions: {
+        ecmaVersion: 'latest',
+        ecmaFeatures: { jsx: true },
+        sourceType: 'module',
+      },
+    },
+    rules: {
+      'no-unused-vars': ['error', { varsIgnorePattern: '^[A-Z_]' }],
+    },
+  },
+])

frontend/index.html

@@ -0,0 +1,16 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <link rel="icon" type="image/svg+xml" href="/favicon.svg" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <link rel="preconnect" href="https://fonts.googleapis.com">
+    <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
+    <link href="https://fonts.googleapis.com/css2?family=Fredoka:wght@400;500;600;700&display=swap" rel="stylesheet">
+    <title>Gorilla Semiconductors</title>
+  </head>
+  <body>
+    <div id="root"></div>
+    <script type="module" src="/src/main.jsx"></script>
+  </body>
+</html>

frontend/package.json

@@ -0,0 +1,31 @@
+{
+  "name": "frontend",
+  "private": true,
+  "version": "0.0.0",
+  "type": "module",
+  "scripts": {
+    "dev": "vite",
+    "build": "vite build",
+    "lint": "eslint .",
+    "preview": "vite preview"
+  },
+  "dependencies": {
+    "axios": "^1.14.0",
+    "chart.js": "^4.5.1",
+    "lucide-react": "^1.7.0",
+    "react": "^19.2.4",
+    "react-chartjs-2": "^5.3.1",
+    "react-dom": "^19.2.4"
+  },
+  "devDependencies": {
+    "@eslint/js": "^9.39.4",
+    "@types/react": "^19.2.14",
+    "@types/react-dom": "^19.2.3",
+    "@vitejs/plugin-react": "^6.0.1",
+    "eslint": "^9.39.4",
+    "eslint-plugin-react-hooks": "^7.0.1",
+    "eslint-plugin-react-refresh": "^0.5.2",
+    "globals": "^17.4.0",
+    "vite": "^8.0.1"
+  }
+}

frontend/src/App.css

@@ -0,0 +1,184 @@
+.counter {
+  font-size: 16px;
+  padding: 5px 10px;
+  border-radius: 5px;
+  color: var(--accent);
+  background: var(--accent-bg);
+  border: 2px solid transparent;
+  transition: border-color 0.3s;
+  margin-bottom: 24px;
+
+  &:hover {
+    border-color: var(--accent-border);
+  }
+  &:focus-visible {
+    outline: 2px solid var(--accent);
+    outline-offset: 2px;
+  }
+}
+
+.hero {
+  position: relative;
+
+  .base,
+  .framework,
+  .vite {
+    inset-inline: 0;
+    margin: 0 auto;
+  }
+
+  .base {
+    width: 170px;
+    position: relative;
+    z-index: 0;
+  }
+
+  .framework,
+  .vite {
+    position: absolute;
+  }
+
+  .framework {
+    z-index: 1;
+    top: 34px;
+    height: 28px;
+    transform: perspective(2000px) rotateZ(300deg) rotateX(44deg) rotateY(39deg)
+      scale(1.4);
+  }
+
+  .vite {
+    z-index: 0;
+    top: 107px;
+    height: 26px;
+    width: auto;
+    transform: perspective(2000px) rotateZ(300deg) rotateX(40deg) rotateY(39deg)
+      scale(0.8);
+  }
+}
+
+#center {
+  display: flex;
+  flex-direction: column;
+  gap: 25px;
+  place-content: center;
+  place-items: center;
+  flex-grow: 1;
+
+  @media (max-width: 1024px) {
+    padding: 32px 20px 24px;
+    gap: 18px;
+  }
+}
+
+#next-steps {
+  display: flex;
+  border-top: 1px solid var(--border);
+  text-align: left;
+
+  & > div {
+    flex: 1 1 0;
+    padding: 32px;
+    @media (max-width: 1024px) {
+      padding: 24px 20px;
+    }
+  }
+
+  .icon {
+    margin-bottom: 16px;
+    width: 22px;
+    height: 22px;
+  }
+
+  @media (max-width: 1024px) {
+    flex-direction: column;
+    text-align: center;
+  }
+}
+
+#docs {
+  border-right: 1px solid var(--border);
+
+  @media (max-width: 1024px) {
+    border-right: none;
+    border-bottom: 1px solid var(--border);
+  }
+}
+
+#next-steps ul {
+  list-style: none;
+  padding: 0;
+  display: flex;
+  gap: 8px;
+  margin: 32px 0 0;
+
+  .logo {
+    height: 18px;
+  }
+
+  a {
+    color: var(--text-h);
+    font-size: 16px;
+    border-radius: 6px;
+    background: var(--social-bg);
+    display: flex;
+    padding: 6px 12px;
+    align-items: center;
+    gap: 8px;
+    text-decoration: none;
+    transition: box-shadow 0.3s;
+
+    &:hover {
+      box-shadow: var(--shadow);
+    }
+    .button-icon {
+      height: 18px;
+      width: 18px;
+    }
+  }
+
+  @media (max-width: 1024px) {
+    margin-top: 20px;
+    flex-wrap: wrap;
+    justify-content: center;
+
+    li {
+      flex: 1 1 calc(50% - 8px);
+    }
+
+    a {
+      width: 100%;
+      justify-content: center;
+      box-sizing: border-box;
+    }
+  }
+}
+
+#spacer {
+  height: 88px;
+  border-top: 1px solid var(--border);
+  @media (max-width: 1024px) {
+    height: 48px;
+  }
+}
+
+.ticks {
+  position: relative;
+  width: 100%;
+
+  &::before,
+  &::after {
+    content: '';
+    position: absolute;
+    top: -4.5px;
+    border: 5px solid transparent;
+  }
+
+  &::before {
+    left: 0;
+    border-left-color: var(--border);
+  }
+  &::after {
+    right: 0;
+    border-right-color: var(--border);
+  }
+}

frontend/src/App.jsx

@@ -0,0 +1,68 @@
+import React, { useState } from 'react';
+import './index.css';
+import { HistoricalAnalytics } from './components/HistoricalAnalytics.jsx';
+import { MaterialPredictor } from './components/MaterialPredictor.jsx';
+import { ChatBot } from './components/ChatBot.jsx';
+import logo from '../Assets/g.png';
+import thumpVideo from '../Assets/Gorilla_Chest_Thumping_Animation_Generated.mp4';
+
+function App() {
+  const [activeTab, setActiveTab] = useState('waste');
+  const [showChat, setShowChat] = useState(false);
+  const [showVideo, setShowVideo] = useState(false);
+
+  const handleGorillaClick = () => {
+    setShowVideo(true);
+    setShowChat(false);
+  };
+
+  return (
+    <>
+      <div className="dashboard-header">
+        <div className="header-title-container">
+          <img 
+            src={logo} 
+            alt="Gorilla Semiconductors Logo" 
+            className="gorilla-logo" 
+            onClick={handleGorillaClick}
+          />
+          <h1 className="header-title">Gorilla Semiconductors</h1>
+        </div>
+      </div>
+
+      <div className="tabs-container">
+        <button 
+          className={`tab-btn ${activeTab === 'waste' ? 'active' : ''}`}
+          onClick={() => setActiveTab('waste')}
+        >
+          Historical Waste Analysis
+        </button>
+        <button 
+          className={`tab-btn ${activeTab === 'predict' ? 'active' : ''}`}
+          onClick={() => setActiveTab('predict')}
+        >
+          Material Prediction
+        </button>
+      </div>
+
+      <div style={{ flex: 1, minHeight: 0, display: 'flex', flexDirection: 'column' }}>
+        {activeTab === 'waste' ? <HistoricalAnalytics /> : <MaterialPredictor />}
+      </div>
+
+      <ChatBot isOpen={showChat} onClose={() => setShowChat(false)} />
+
+      {showVideo && (
+        <div className="thump-video-overlay" onClick={() => { setShowVideo(false); setShowChat(true); }}>
+          <video 
+            src={thumpVideo} 
+            autoPlay 
+            className="thump-video"
+            onEnded={() => { setShowVideo(false); setShowChat(true); }}
+          />
+        </div>
+      )}
+    </>
+  );
+}
+
+export default App;

frontend/src/apiConfig.js

@@ -0,0 +1,3 @@
+const API_BASE_URL = import.meta.env.VITE_API_URL || 'http://localhost:8000';
+
+export default API_BASE_URL;

frontend/src/components/ChatBot.jsx

@@ -0,0 +1,131 @@
+import React, { useState, useRef, useEffect } from 'react';
+import axios from 'axios';
+
+import API_BASE_URL from '../apiConfig';
+
+export const ChatBot = ({ isOpen, onClose }) => {
+  const [messages, setMessages] = useState([]);
+  const [input, setInput] = useState('');
+  const [isLoading, setIsLoading] = useState(false);
+  const messagesEndRef = useRef(null);
+
+  // Dragging state
+  const [position, setPosition] = useState({ 
+    x: typeof window !== 'undefined' ? window.innerWidth - 400 : 0, 
+    y: typeof window !== 'undefined' ? window.innerHeight - 650 : 0 
+  });
+  const [isDragging, setIsDragging] = useState(false);
+  const dragRef = useRef({ startX: 0, startY: 0, initialX: 0, initialY: 0 });
+
+  const scrollToBottom = () => {
+    messagesEndRef.current?.scrollIntoView({ behavior: 'smooth' });
+  };
+
+  useEffect(() => {
+    scrollToBottom();
+  }, [messages]);
+
+  // Handle Dragging
+  const handlePointerDown = (e) => {
+    // Don't drag if clicking the close button
+    if (e.target.tagName.toLowerCase() === 'button') return;
+    setIsDragging(true);
+    dragRef.current = {
+      startX: e.clientX,
+      startY: e.clientY,
+      initialX: position.x,
+      initialY: position.y
+    };
+    e.currentTarget.setPointerCapture(e.pointerId);
+  };
+
+  const handlePointerMove = (e) => {
+    if (!isDragging) return;
+    const dx = e.clientX - dragRef.current.startX;
+    const dy = e.clientY - dragRef.current.startY;
+    setPosition({
+      x: dragRef.current.initialX + dx,
+      y: dragRef.current.initialY + dy
+    });
+  };
+
+  const handlePointerUp = (e) => {
+    setIsDragging(false);
+    e.currentTarget.releasePointerCapture(e.pointerId);
+  };
+
+  const handleSend = async (e) => {
+    e.preventDefault();
+    if (!input.trim()) return;
+
+    const userMessage = { role: 'user', content: input };
+    setMessages(prev => [...prev, userMessage]);
+    setInput('');
+    setIsLoading(true);
+
+    try {
+      const response = await axios.post(`${API_BASE_URL}/api/chat`, {
+        messages: [...messages, userMessage].map(m => ({ role: m.role, content: m.content }))
+      });
+      
+      if (response.data.response) {
+        setMessages(prev => [...prev, { role: 'model', content: response.data.response }]);
+      } else if (response.data.error) {
+        setMessages(prev => [...prev, { role: 'model', content: `Error: ${response.data.error}` }]);
+      }
+    } catch (error) {
+      console.error('Chat error:', error);
+      setMessages(prev => [...prev, { role: 'model', content: "GRRR... I couldn't reach the server. Is it running?" }]);
+    } finally {
+      setIsLoading(false);
+    }
+  };
+
+  if (!isOpen) return null;
+
+  return (
+    <div 
+      className="chat-widget-container" 
+      style={{ left: position.x, top: position.y, margin: 0 }}
+    >
+      <div 
+        className="chat-header" 
+        onPointerDown={handlePointerDown}
+        onPointerMove={handlePointerMove}
+        onPointerUp={handlePointerUp}
+      >
+        <h3>🦍 Gorilla Bot</h3>
+        <button className="chat-close-btn" onClick={onClose}>×</button>
+      </div>
+      <div className="chat-messages">
+        {messages.map((msg, idx) => (
+          <div key={idx} className={`chat-message ${msg.role}`}>
+            <div className="chat-bubble">
+              {msg.content}
+            </div>
+          </div>
+        ))}
+        {isLoading && (
+          <div className="chat-message model">
+            <div className="chat-bubble loading">
+              Thinking... 🍌
+            </div>
+          </div>
+        )}
+        <div ref={messagesEndRef} />
+      </div>
+      <form className="chat-input-area" onSubmit={handleSend}>
+        <input 
+          type="text" 
+          value={input}
+          onChange={(e) => setInput(e.target.value)}
+          placeholder="Ask about defects, KPIs, or forecasts..." 
+          className="chat-input"
+        />
+        <button type="submit" className="chat-send-btn" disabled={isLoading}>
+          Send
+        </button>
+      </form>
+    </div>
+  );
+};

frontend/src/components/HistoricalAnalytics.jsx

@@ -0,0 +1,150 @@
+import React, { useState, useEffect } from 'react';
+import axios from 'axios';
+import { Chart as ChartJS, ArcElement, Tooltip, Legend, CategoryScale, LinearScale, PointElement, LineElement, BarElement } from 'chart.js';
+import { Pie, Bar, Line } from 'react-chartjs-2';
+import KPICard from './KPICard';
+
+import API_BASE_URL from '../apiConfig';
+
+ChartJS.register(ArcElement, Tooltip, Legend, CategoryScale, LinearScale, PointElement, LineElement, BarElement);
+
+const COLORS = {
+  accent: '#f472b6',
+  accent2: '#38bdf8',
+  accent3: '#4ade80',
+  danger: '#fb7185',
+  warning: '#fbbf24',
+  text: '#000000',
+  textMuted: '#3f3f46'
+};
+
+const DEFECT_COLORS = {
+  'Center': '#ef4444', 'Donut': '#f59e0b', 'Edge-Loc': '#10b981',
+  'Edge-Ring': '#3b82f6', 'Loc': '#8b5cf6', 'Random': '#ec4899',
+  'Scratch': '#06b6d4', 'Near-full': '#f97316', 'None': '#6b7280',
+  'Undetected': '#374151'
+};
+
+const chartOptions = {
+    color: COLORS.text,
+    plugins: {
+        legend: {
+            labels: { color: COLORS.textMuted }
+        }
+    },
+    scales: {
+        x: { ticks: { color: COLORS.textMuted }, grid: { color: 'rgba(0,0,0,0.1)' } },
+        y: { ticks: { color: COLORS.textMuted }, grid: { color: 'rgba(0,0,0,0.1)' } }
+    }
+};
+const pieOptions = {
+    color: COLORS.text,
+    plugins: { legend: { labels: { color: COLORS.textMuted } } }
+};
+
+export const HistoricalAnalytics = () => {
+    const [kpis, setKpis] = useState(null);
+    const [defects, setDefects] = useState(null);
+    const [waste, setWaste] = useState(null);
+    const [trends, setTrends] = useState(null);
+
+    useEffect(() => {
+        const fetchData = async () => {
+            const [kRes, dRes, wRes, tRes] = await Promise.all([
+                axios.get(`${API_BASE_URL}/api/kpi`),
+                axios.get(`${API_BASE_URL}/api/charts/defects`),
+                axios.get(`${API_BASE_URL}/api/charts/waste`),
+                axios.get(`${API_BASE_URL}/api/charts/trends`)
+            ]);
+            setKpis(kRes.data);
+            setDefects(dRes.data);
+            setWaste(wRes.data);
+            setTrends(tRes.data);
+        };
+        fetchData();
+    }, []);
+
+    if (!kpis || !defects || !waste || !trends) return <div>Loading Analytics...</div>;
+
+    const pieData = {
+        labels: defects.predictions.map(d => d.defect_type),
+        datasets: [{
+            data: defects.predictions.map(d => d.count),
+            backgroundColor: defects.predictions.map(d => DEFECT_COLORS[d.defect_type] || COLORS.textMuted),
+            borderColor: 'transparent'
+        }]
+    };
+
+    const barData = {
+        labels: waste.waste_by_type.map(w => w.defect_type),
+        datasets: [{
+            label: 'Total Material Waste (Wafers)',
+            data: waste.waste_by_type.map(w => w.total_waste),
+            backgroundColor: waste.waste_by_type.map(w => DEFECT_COLORS[w.defect_type] || COLORS.textMuted)
+        }]
+    };
+
+    const trendData = {
+        labels: trends.dates,
+        datasets: [{
+            label: 'Fail Rate %',
+            data: trends.fail_rate,
+            borderColor: COLORS.danger,
+            backgroundColor: 'rgba(251, 113, 133, 0.4)',
+            fill: true,
+            yAxisID: 'y'
+        }]
+    };
+
+    const wasteTrendData = {
+        labels: trends.dates,
+        datasets: [{
+            label: 'Total Lost Wafers',
+            data: trends.waste,
+            borderColor: COLORS.warning,
+            backgroundColor: 'rgba(251, 191, 36, 0.4)',
+            fill: true,
+            yAxisID: 'y'
+        }]
+    };
+
+    return (
+        <div style={{ display: 'flex', flexDirection: 'column', height: '100%' }}>
+            <div className="kpi-container">
+                <KPICard title="Total Scans" value={kpis.total_scans.toLocaleString()} subtitle="wafers inspected" color={COLORS.accent} />
+                <KPICard title="Pass Rate" value={`${kpis.pass_rate}%`} subtitle={`${kpis.pass_count.toLocaleString()} passed`} color={COLORS.accent3} />
+                <KPICard title="Fail Rate" value={`${kpis.fail_rate}%`} subtitle={`${kpis.fail_count.toLocaleString()} defective`} color={COLORS.danger} />
+                <KPICard title="Scrapped" value={kpis.scrap_count.toLocaleString()} subtitle="routed to scrap" color={COLORS.warning} />
+                <KPICard title="Avg Waste/Wafer" value={`${kpis.avg_waste}%`} subtitle="per defective wafer" color={COLORS.danger} />
+                <KPICard title="Avg Confidence" value={kpis.avg_confidence} subtitle="model certainty" color={COLORS.accent3} />
+            </div>
+
+            <div className="charts-master-grid">
+                <div className="glass-card chart-card">
+                    <h3 style={{marginTop:0, marginBottom: '8px', fontSize: '14px'}}>YOLOv8 Predicted Distributions</h3>
+                    <div className="canvas-container">
+                        <Pie data={pieData} options={{...pieOptions, maintainAspectRatio: false}} />
+                    </div>
+                </div>
+                <div className="glass-card chart-card">
+                    <h3 style={{marginTop:0, marginBottom: '8px', fontSize: '14px'}}>Total Material Waste by Predict Defect</h3>
+                    <div className="canvas-container">
+                        <Bar data={barData} options={{...chartOptions, maintainAspectRatio: false}} />
+                    </div>
+                </div>
+                <div className="glass-card chart-card">
+                    <h3 style={{marginTop:0, marginBottom: '8px', fontSize: '14px'}}>Daily Defect Rate Over Time</h3>
+                    <div className="canvas-container">
+                        <Line data={trendData} options={{...chartOptions, maintainAspectRatio: false}} />
+                    </div>
+                </div>
+                <div className="glass-card chart-card">
+                    <h3 style={{marginTop:0, marginBottom: '8px', fontSize: '14px'}}>Daily Material Waste Over Time</h3>
+                    <div className="canvas-container">
+                        <Line data={wasteTrendData} options={{...chartOptions, maintainAspectRatio: false}} />
+                    </div>
+                </div>
+            </div>
+        </div>
+    );
+};

frontend/src/components/KPICard.jsx

@@ -0,0 +1,14 @@
+import React from 'react';
+import '../index.css';
+
+const KPICard = ({ title, value, subtitle, color }) => {
+  return (
+    <div className="kpi-card">
+      <p className="kpi-title">{title}</p>
+      <h2 className="kpi-value" style={{ color: color }}>{value}</h2>
+      {subtitle && <p className="kpi-subtitle">{subtitle}</p>}
+    </div>
+  );
+};
+
+export default KPICard;

frontend/src/components/MaterialPredictor.jsx

@@ -0,0 +1,73 @@
+import React, { useState, useEffect } from 'react';
+import axios from 'axios';
+import KPICard from './KPICard';
+
+import API_BASE_URL from '../apiConfig';
+
+export const MaterialPredictor = () => {
+    const [scans, setScans] = useState(1300);
+    const [failRate, setFailRate] = useState(97);
+    const [prediction, setPrediction] = useState(null);
+    const [modelStatus, setModelStatus] = useState(null);
+
+    useEffect(() => {
+        axios.get(`${API_BASE_URL}/api/model/status`).then(res => {
+            setModelStatus(res.data);
+        }).catch(() => setModelStatus({loaded: false}));
+    }, []);
+
+    const handlePredict = async () => {
+        try {
+            const res = await axios.post(`${API_BASE_URL}/api/predict`, { scans, fail_rate: failRate });
+            setPrediction(res.data);
+        } catch(e) {
+            console.error(e);
+        }
+    };
+
+    if (!modelStatus) return <div>Loading...</div>;
+
+    return (
+        <div>
+            <div className="glass-card predictor-header">
+                <div>
+                    <h3 className="predictor-title">Prediction Model</h3>
+                    <p style={{margin:0, color: modelStatus.loaded ? 'var(--accent3)' : 'var(--danger)'}}>
+                        {modelStatus.loaded ? 'Model loaded' : 'No model found'}
+                    </p>
+                </div>
+                {modelStatus.loaded && (
+                    <p style={{color: 'var(--text-muted)', fontFamily:'monospace'}}>
+                        R² = {modelStatus.metrics.r2} | MAE = {modelStatus.metrics.mae}%
+                    </p>
+                )}
+            </div>
+
+            <div className="glass-card">
+                <h3 className="predictor-title" style={{marginBottom: '20px'}}>Forecast Parameters</h3>
+                <div className="predictor-form">
+                    <div className="slider-container">
+                        <label>Expected Daily Production ({scans} wafers)</label>
+                        <input type="range" min="100" max="2000" step="50" value={scans} onChange={(e) => setScans(parseInt(e.target.value))} />
+                    </div>
+                    <div className="slider-container">
+                        <label>Expected Defect Rate ({failRate}%)</label>
+                        <input type="range" min="0" max="100" step="5" value={failRate} onChange={(e) => setFailRate(parseInt(e.target.value))} />
+                    </div>
+                </div>
+                <button className="btn-primary" onClick={handlePredict}>Predict Material Needs</button>
+            </div>
+
+            {prediction && (
+                <div className="glass-card prediction-result">
+                    <div className="kpi-container" style={{marginBottom: 0}}>
+                        <KPICard title="Daily Production" value={prediction.total_scans} subtitle="wafers" color="var(--accent)" />
+                        <KPICard title="Expected Defect Rate" value={`${prediction.fail_rate}%`} subtitle={`~${Math.round((prediction.total_scans*prediction.fail_rate)/100)} defective`} color="var(--danger)" />
+                        <KPICard title="Avg Waste/Wafer" value={`${prediction.avg_waste_per_wafer}%`} subtitle="per defective wafer loss" color="var(--warning)" />
+                        <KPICard title="Total Daily Waste" value={`${prediction.total_daily_waste} wafers`} subtitle="total estimated loss" color="var(--danger)" />
+                    </div>
+                </div>
+            )}
+        </div>
+    );
+};

frontend/src/index.css

@@ -0,0 +1,490 @@
+:root {
+  --bg: #fbbf24; /* Banana Yellow */
+  --card: #ffffff;
+  --card-border: #000000;
+  --accent: #f472b6; /* Bubblegum Pink */
+  --accent2: #38bdf8; /* Sky Blue */
+  --accent3: #4ade80; /* Jungle Green */
+  --danger: #fb7185;
+  --warning: #fbcfe8;
+  --text: #000000;
+  --text-muted: #3f3f46;
+  --font-family: 'Fredoka', sans-serif;
+}
+
+body {
+  margin: 0;
+  padding: 0;
+  background-color: var(--bg);
+  color: var(--text);
+  font-family: var(--font-family);
+  -webkit-font-smoothing: antialiased;
+  -moz-osx-font-smoothing: grayscale;
+}
+
+#root {
+  height: 100vh;
+  padding: 12px 16px;
+  box-sizing: border-box;
+  display: flex;
+  flex-direction: column;
+  overflow: hidden;
+}
+
+.dashboard-header {
+  display: flex;
+  flex-direction: column;
+  margin-bottom: 12px;
+  border-bottom: 4px solid var(--card-border);
+  padding-bottom: 8px;
+  flex-shrink: 0;
+  background-color: var(--card);
+  border-radius: 12px;
+  padding: 8px 16px;
+  box-shadow: 4px 4px 0px #000000;
+  border: 3px solid #000000;
+}
+
+.header-title-container {
+  display: flex;
+  align-items: center;
+  gap: 12px;
+}
+
+.header-dot {
+  width: 12px;
+  height: 12px;
+  border-radius: 50%;
+  background-color: var(--accent3);
+  box-shadow: 0 0 8px var(--accent3);
+  animation: pulse 2s infinite;
+}
+
+@keyframes pulse {
+  0% { box-shadow: 0 0 8px var(--accent3); }
+  50% { box-shadow: 0 0 16px var(--accent3); }
+  100% { box-shadow: 0 0 8px var(--accent3); }
+}
+
+.header-title {
+  margin: 0;
+  font-size: 26px;
+  font-weight: 700;
+  color: var(--text);
+  letter-spacing: 1px;
+}
+
+
+.gorilla-logo {
+  height: 48px;
+  width: auto;
+  border-radius: 8px;
+  object-fit: contain;
+  cursor: pointer;
+  transform-origin: center bottom;
+  transition: transform 0.2s cubic-bezier(0.34, 1.56, 0.64, 1);
+}
+
+.gorilla-logo:hover {
+  transform: scale(1.15);
+}
+
+@keyframes fadeInOverlay {
+  0% { opacity: 0; }
+  100% { opacity: 1; }
+}
+
+@keyframes fadeInScale {
+  0% { opacity: 0; transform: scale(0.5); }
+  100% { opacity: 1; transform: scale(1); }
+}
+
+.thump-video-overlay {
+  position: fixed;
+  top: 0;
+  left: 0;
+  width: 100vw;
+  height: 100vh;
+  background: rgba(0, 0, 0, 0.5);
+  display: flex;
+  justify-content: center;
+  align-items: center;
+  z-index: 9999;
+  animation: fadeInOverlay 0.4s ease-out forwards;
+}
+
+.thump-video {
+  width: 250px;
+  height: auto;
+  border-radius: 20px;
+  border: 6px solid #000000;
+  box-shadow: 8px 8px 0px #000000;
+  animation: fadeInScale 0.4s cubic-bezier(0.34, 1.56, 0.64, 1) forwards;
+}
+
+.header-subtitle {
+  color: var(--text-muted);
+  margin-top: 4px;
+  font-size: 13px;
+}
+
+/* Tabs */
+.tabs-container {
+  display: flex;
+  gap: 12px;
+  margin-bottom: 16px;
+  flex-shrink: 0;
+}
+
+.tab-btn {
+  background-color: var(--card);
+  color: var(--text);
+  border: 3px solid var(--card-border);
+  border-bottom: none;
+  border-radius: 16px 16px 0 0;
+  padding: 12px 24px;
+  font-weight: 600;
+  font-size: 16px;
+  cursor: pointer;
+  transition: all 0.2s cubic-bezier(0.34, 1.56, 0.64, 1);
+  outline: none;
+  font-family: var(--font-family);
+  box-shadow: inset 0 -4px 0 rgba(0,0,0,0.1);
+}
+
+.tab-btn:hover {
+  background-color: var(--accent);
+  transform: translateY(-2px);
+}
+
+.tab-btn.active {
+  background-color: var(--accent2);
+  color: #000000;
+  border-color: #000000;
+}
+
+/* KPI Container */
+.kpi-container {
+  display: flex;
+  gap: 8px;
+  flex-wrap: nowrap;
+  margin-bottom: 16px;
+  flex-shrink: 0;
+}
+
+/* KPI Card */
+.kpi-card {
+  background: var(--card);
+  border: 3px solid var(--card-border);
+  border-radius: 16px;
+  padding: 10px;
+  text-align: center;
+  flex: 1;
+  min-width: 0;
+  transition: all 0.2s cubic-bezier(0.34, 1.56, 0.64, 1);
+  box-shadow: 4px 4px 0px #000000;
+}
+
+.kpi-card:hover {
+  transform: scale(1.05) rotate(-2deg);
+  box-shadow: 8px 8px 0px #000000;
+  background-color: var(--warning);
+}
+
+.kpi-title {
+  color: var(--text-muted);
+  font-size: 11px;
+  margin-bottom: 2px;
+  font-weight: 500;
+  text-transform: uppercase;
+  letter-spacing: 1px;
+}
+
+.kpi-value {
+  font-size: 20px;
+  font-weight: 700;
+  margin: 4px 0;
+}
+
+.kpi-subtitle {
+  color: var(--text-muted);
+  font-size: 10px;
+  margin-top: 2px;
+}
+
+/* Base Card */
+.glass-card {
+  background: var(--card);
+  border: 3px solid var(--card-border);
+  border-radius: 16px;
+  padding: 12px;
+  margin-bottom: 0;
+  box-shadow: 6px 6px 0px #000000;
+  transition: transform 0.2s cubic-bezier(0.34, 1.56, 0.64, 1);
+}
+.glass-card h3 {
+  font-weight: 700;
+  letter-spacing: 0.5px;
+}
+
+/* Charts Grid */
+.charts-master-grid {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+  grid-template-rows: minmax(0, 1fr) minmax(0, 1fr);
+  gap: 12px;
+  flex: 1;
+  min-height: 0;
+}
+
+.chart-card {
+  display: flex;
+  flex-direction: column;
+}
+
+.canvas-container {
+  flex: 1;
+  min-height: 0;
+  position: relative;
+}
+
+@media (max-width: 900px) {
+  .charts-master-grid {
+    grid-template-columns: 1fr;
+    grid-template-rows: auto;
+    overflow-y: auto;
+  }
+  #root {
+    height: auto;
+    overflow: visible;
+  }
+}
+
+/* Material Predictor */
+.predictor-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+}
+
+.predictor-title {
+  margin: 0 0 4px 0;
+  font-size: 18px;
+}
+
+.predictor-form {
+  display: grid;
+  grid-template-columns: 1fr 1fr;
+  gap: 24px;
+  margin-bottom: 20px;
+}
+
+.slider-container label {
+  display: block;
+  color: var(--text-muted);
+  font-size: 13px;
+  margin-bottom: 8px;
+}
+
+.slider-container input[type="range"] {
+  width: 100%;
+  accent-color: var(--accent2);
+}
+
+.btn-primary {
+  background-color: var(--accent);
+  color: #000000;
+  border: 3px solid #000000;
+  border-radius: 16px;
+  padding: 12px 32px;
+  font-size: 18px;
+  font-weight: 700;
+  cursor: pointer;
+  width: 100%;
+  transition: all 0.2s cubic-bezier(0.34, 1.56, 0.64, 1);
+  box-shadow: 4px 4px 0px #000000;
+  font-family: var(--font-family);
+}
+
+.btn-primary:hover {
+  background-color: var(--accent2);
+  transform: translate(-2px, -2px);
+  box-shadow: 6px 6px 0px #000000;
+}
+.btn-primary:active {
+  transform: translate(2px, 2px);
+  box-shadow: 0px 0px 0px #000000;
+}
+
+.prediction-result {
+  background: var(--accent3);
+}
+
+/* ChatBot Styles */
+.chat-modal-overlay {
+  position: fixed;
+  top: 0;
+  left: 0;
+  width: 100vw;
+  height: 100vh;
+  background: rgba(0, 0, 0, 0.4);
+  display: flex;
+  justify-content: center;
+  align-items: center;
+  z-index: 10000;
+  animation: fadeInOverlay 0.2s ease-out forwards;
+}
+
+.chat-widget-container {
+  position: fixed;
+  width: 360px;
+  height: 600px;
+  min-width: 250px;
+  min-height: 300px;
+  background: var(--card);
+  border: 4px solid #000000;
+  border-radius: 16px;
+  box-shadow: 12px 12px 0px #000000;
+  display: flex;
+  flex-direction: column;
+  overflow: hidden;
+  z-index: 10000;
+  resize: both;
+}
+
+.chat-header {
+  background: var(--accent2);
+  padding: 16px;
+  border-bottom: 4px solid #000000;
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  cursor: move;
+  user-select: none;
+}
+
+.chat-header h3 {
+  margin: 0;
+  font-size: 20px;
+  font-weight: 800;
+}
+
+.chat-close-btn {
+  background: var(--card);
+  border: 3px solid #000000;
+  border-radius: 50%;
+  width: 32px;
+  height: 32px;
+  font-size: 20px;
+  font-weight: bold;
+  cursor: pointer;
+  display: flex;
+  justify-content: center;
+  align-items: center;
+  transition: transform 0.2s;
+}
+
+.chat-close-btn:hover {
+  background: var(--danger);
+  transform: scale(1.1) rotate(90deg);
+}
+
+.chat-messages {
+  flex: 1;
+  padding: 16px;
+  overflow-y: auto;
+  display: flex;
+  flex-direction: column;
+  gap: 12px;
+  background: #e0f2fe;
+}
+
+.chat-message {
+  display: flex;
+  width: 100%;
+}
+
+.chat-message.user {
+  justify-content: flex-end;
+}
+
+.chat-message.model {
+  justify-content: flex-start;
+}
+
+.chat-bubble {
+  max-width: 80%;
+  padding: 12px 16px;
+  border: 3px solid #000000;
+  border-radius: 12px;
+  font-size: 15px;
+  line-height: 1.4;
+  box-shadow: 4px 4px 0px #000000;
+  white-space: pre-wrap;
+}
+
+.chat-message.user .chat-bubble {
+  background: var(--accent);
+  border-bottom-right-radius: 0;
+}
+
+.chat-message.model .chat-bubble {
+  background: var(--card);
+  border-bottom-left-radius: 0;
+}
+
+.chat-bubble.loading {
+  font-style: italic;
+  color: var(--text-muted);
+}
+
+.chat-input-area {
+  display: flex;
+  padding: 12px;
+  background: var(--card);
+  border-top: 4px solid #000000;
+  gap: 8px;
+}
+
+.chat-input {
+  flex: 1;
+  padding: 12px;
+  border: 3px solid #000000;
+  border-radius: 8px;
+  font-size: 16px;
+  font-family: var(--font-family);
+  outline: none;
+}
+
+.chat-input:focus {
+  border-color: var(--accent2);
+}
+
+.chat-send-btn {
+  background: var(--accent3);
+  color: #000000;
+  border: 3px solid #000000;
+  border-radius: 8px;
+  padding: 0 24px;
+  font-weight: 700;
+  font-size: 16px;
+  cursor: pointer;
+  font-family: var(--font-family);
+  box-shadow: 4px 4px 0px #000000;
+  transition: transform 0.1s, box-shadow 0.1s;
+}
+
+.chat-send-btn:hover:not(:disabled) {
+  transform: translate(-2px, -2px);
+  box-shadow: 6px 6px 0px #000000;
+}
+
+.chat-send-btn:active:not(:disabled) {
+  transform: translate(2px, 2px);
+  box-shadow: 0px 0px 0px #000000;
+}
+
+.chat-send-btn:disabled {
+  background: #ccc;
+  cursor: not-allowed;
+}

frontend/src/main.jsx

@@ -0,0 +1,10 @@
+import { StrictMode } from 'react'
+import { createRoot } from 'react-dom/client'
+import './index.css'
+import App from './App.jsx'
+
+createRoot(document.getElementById('root')).render(
+  <StrictMode>
+    <App />
+  </StrictMode>,
+)

frontend/vite.config.js

@@ -0,0 +1,7 @@
+import { defineConfig } from 'vite'
+import react from '@vitejs/plugin-react'
+
+// https://vite.dev/config/
+export default defineConfig({
+  plugins: [react()],
+})

middleware/EDA_wafer_control_db.ipynb

@@ -0,0 +1,604 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "2d67457f",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "   id wafer_id               batch_id            scan_time status  \\\n",
+      "0   1  wafer_0  BATCH_20260317_201406  2026-02-15 20:14:17   FAIL   \n",
+      "1   2  wafer_1  BATCH_20260317_201406  2026-02-15 20:14:57   FAIL   \n",
+      "2   3  wafer_2  BATCH_20260317_201406  2026-02-15 20:14:59   FAIL   \n",
+      "3   4  wafer_3  BATCH_20260317_201406  2026-02-15 20:14:21   FAIL   \n",
+      "4   5  wafer_4  BATCH_20260317_201406  2026-02-15 20:14:36   FAIL   \n",
+      "\n",
+      "             ground_truth defect_type               action  confidence  \\\n",
+      "0  Center+Edge-Loc+Random   Edge-Ring  MOVE_TO_MICRO_STAGE        0.90   \n",
+      "1  Center+Edge-Loc+Random      Center       ROUTE_TO_SCRAP        0.96   \n",
+      "2  Center+Edge-Loc+Random      Center       ROUTE_TO_SCRAP        0.88   \n",
+      "3  Center+Edge-Loc+Random   Edge-Ring  MOVE_TO_MICRO_STAGE        0.91   \n",
+      "4  Center+Edge-Loc+Random      Center       ROUTE_TO_SCRAP        0.85   \n",
+      "\n",
+      "  roi_coordinates  defect_area_px  material_wasted_pct  \n",
+      "0  [1, 0, 51, 50]            2500                92.46  \n",
+      "1  [1, 0, 51, 50]            2500                92.46  \n",
+      "2  [1, 0, 51, 50]            2500                92.46  \n",
+      "3  [1, 0, 51, 50]            2500                92.46  \n",
+      "4  [1, 0, 52, 51]            2601                96.19  \n"
+     ]
+    }
+   ],
+   "source": [
+    "import pandas as pd\n",
+    "import sqlite3\n",
+    "\n",
+    "# 1. Connect to the database file\n",
+    "conn = sqlite3.connect('/Users/udayan/CHITS_PR_1/middleware/wafer_control.db')\n",
+    "\n",
+    "# 2. Write a query to select the data you want\n",
+    "query = \"SELECT * FROM wafer_logs\"\n",
+    "\n",
+    "# 3. Load the data into a DataFrame\n",
+    "df = pd.read_sql_query(query, conn)\n",
+    "\n",
+    "# 4. Close the connection\n",
+    "conn.close()\n",
+    "\n",
+    "# View the first few rows\n",
+    "print(df.head())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "b287cad1",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>wafer_id</th>\n",
+       "      <th>batch_id</th>\n",
+       "      <th>scan_time</th>\n",
+       "      <th>status</th>\n",
+       "      <th>defect_type</th>\n",
+       "      <th>action</th>\n",
+       "      <th>confidence</th>\n",
+       "      <th>roi_coordinates</th>\n",
+       "      <th>defect_area_px</th>\n",
+       "      <th>material_wasted_pct</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>1</td>\n",
+       "      <td>wafer_100574</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-02-15 02:15:48</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Edge-Ring</td>\n",
+       "      <td>MOVE_TO_MICRO_STAGE</td>\n",
+       "      <td>0.95</td>\n",
+       "      <td>[1, 0, 115, 136]</td>\n",
+       "      <td>15504</td>\n",
+       "      <td>97.56</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>2</td>\n",
+       "      <td>wafer_101787</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-02-15 02:15:09</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Center</td>\n",
+       "      <td>ROUTE_TO_SCRAP</td>\n",
+       "      <td>0.96</td>\n",
+       "      <td>[0, 0, 43, 43]</td>\n",
+       "      <td>1849</td>\n",
+       "      <td>95.51</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>3</td>\n",
+       "      <td>wafer_103333</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-02-15 02:14:59</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Edge-Ring</td>\n",
+       "      <td>MOVE_TO_MICRO_STAGE</td>\n",
+       "      <td>0.98</td>\n",
+       "      <td>[0, 0, 43, 43]</td>\n",
+       "      <td>1849</td>\n",
+       "      <td>95.51</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>4</td>\n",
+       "      <td>wafer_106281</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-02-15 02:15:19</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Loc</td>\n",
+       "      <td>MOVE_TO_MICRO_STAGE</td>\n",
+       "      <td>0.60</td>\n",
+       "      <td>[0, 0, 30, 34]</td>\n",
+       "      <td>1020</td>\n",
+       "      <td>94.01</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>5</td>\n",
+       "      <td>wafer_106301</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-02-15 02:15:19</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Loc</td>\n",
+       "      <td>MOVE_TO_MICRO_STAGE</td>\n",
+       "      <td>0.96</td>\n",
+       "      <td>[0, 0, 29, 33]</td>\n",
+       "      <td>957</td>\n",
+       "      <td>88.20</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5099</th>\n",
+       "      <td>5100</td>\n",
+       "      <td>wafer_95994</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-03-16 02:15:15</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Center</td>\n",
+       "      <td>ROUTE_TO_SCRAP</td>\n",
+       "      <td>0.96</td>\n",
+       "      <td>[0, 0, 60, 40]</td>\n",
+       "      <td>2400</td>\n",
+       "      <td>93.68</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5100</th>\n",
+       "      <td>5101</td>\n",
+       "      <td>wafer_96083</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-03-17 02:15:17</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Center</td>\n",
+       "      <td>ROUTE_TO_SCRAP</td>\n",
+       "      <td>0.96</td>\n",
+       "      <td>[0, 0, 60, 40]</td>\n",
+       "      <td>2400</td>\n",
+       "      <td>93.68</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5101</th>\n",
+       "      <td>5102</td>\n",
+       "      <td>wafer_9637</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-03-17 02:14:52</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Edge-Loc</td>\n",
+       "      <td>MOVE_TO_MICRO_STAGE</td>\n",
+       "      <td>0.98</td>\n",
+       "      <td>[0, 0, 28, 31]</td>\n",
+       "      <td>868</td>\n",
+       "      <td>90.70</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5102</th>\n",
+       "      <td>5103</td>\n",
+       "      <td>wafer_96594</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-03-17 02:15:22</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Loc</td>\n",
+       "      <td>MOVE_TO_MICRO_STAGE</td>\n",
+       "      <td>0.81</td>\n",
+       "      <td>[0, 0, 30, 29]</td>\n",
+       "      <td>870</td>\n",
+       "      <td>90.53</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5103</th>\n",
+       "      <td>5104</td>\n",
+       "      <td>wafer_983</td>\n",
+       "      <td>BATCH_20260317_021451</td>\n",
+       "      <td>2026-03-17 02:15:00</td>\n",
+       "      <td>FAIL</td>\n",
+       "      <td>Edge-Loc</td>\n",
+       "      <td>MOVE_TO_MICRO_STAGE</td>\n",
+       "      <td>0.69</td>\n",
+       "      <td>[0, 0, 25, 25]</td>\n",
+       "      <td>625</td>\n",
+       "      <td>92.46</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>5104 rows × 11 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "        id      wafer_id               batch_id            scan_time status  \\\n",
+       "0        1  wafer_100574  BATCH_20260317_021451  2026-02-15 02:15:48   FAIL   \n",
+       "1        2  wafer_101787  BATCH_20260317_021451  2026-02-15 02:15:09   FAIL   \n",
+       "2        3  wafer_103333  BATCH_20260317_021451  2026-02-15 02:14:59   FAIL   \n",
+       "3        4  wafer_106281  BATCH_20260317_021451  2026-02-15 02:15:19   FAIL   \n",
+       "4        5  wafer_106301  BATCH_20260317_021451  2026-02-15 02:15:19   FAIL   \n",
+       "...    ...           ...                    ...                  ...    ...   \n",
+       "5099  5100   wafer_95994  BATCH_20260317_021451  2026-03-16 02:15:15   FAIL   \n",
+       "5100  5101   wafer_96083  BATCH_20260317_021451  2026-03-17 02:15:17   FAIL   \n",
+       "5101  5102    wafer_9637  BATCH_20260317_021451  2026-03-17 02:14:52   FAIL   \n",
+       "5102  5103   wafer_96594  BATCH_20260317_021451  2026-03-17 02:15:22   FAIL   \n",
+       "5103  5104     wafer_983  BATCH_20260317_021451  2026-03-17 02:15:00   FAIL   \n",
+       "\n",
+       "     defect_type               action  confidence   roi_coordinates  \\\n",
+       "0      Edge-Ring  MOVE_TO_MICRO_STAGE        0.95  [1, 0, 115, 136]   \n",
+       "1         Center       ROUTE_TO_SCRAP        0.96    [0, 0, 43, 43]   \n",
+       "2      Edge-Ring  MOVE_TO_MICRO_STAGE        0.98    [0, 0, 43, 43]   \n",
+       "3            Loc  MOVE_TO_MICRO_STAGE        0.60    [0, 0, 30, 34]   \n",
+       "4            Loc  MOVE_TO_MICRO_STAGE        0.96    [0, 0, 29, 33]   \n",
+       "...          ...                  ...         ...               ...   \n",
+       "5099      Center       ROUTE_TO_SCRAP        0.96    [0, 0, 60, 40]   \n",
+       "5100      Center       ROUTE_TO_SCRAP        0.96    [0, 0, 60, 40]   \n",
+       "5101    Edge-Loc  MOVE_TO_MICRO_STAGE        0.98    [0, 0, 28, 31]   \n",
+       "5102         Loc  MOVE_TO_MICRO_STAGE        0.81    [0, 0, 30, 29]   \n",
+       "5103    Edge-Loc  MOVE_TO_MICRO_STAGE        0.69    [0, 0, 25, 25]   \n",
+       "\n",
+       "      defect_area_px  material_wasted_pct  \n",
+       "0              15504                97.56  \n",
+       "1               1849                95.51  \n",
+       "2               1849                95.51  \n",
+       "3               1020                94.01  \n",
+       "4                957                88.20  \n",
+       "...              ...                  ...  \n",
+       "5099            2400                93.68  \n",
+       "5100            2400                93.68  \n",
+       "5101             868                90.70  \n",
+       "5102             870                90.53  \n",
+       "5103             625                92.46  \n",
+       "\n",
+       "[5104 rows x 11 columns]"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "299c2be8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pass_wafers = df[df['status'] == 'PASS']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "46fbb2d3",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>id</th>\n",
+       "      <th>wafer_id</th>\n",
+       "      <th>batch_id</th>\n",
+       "      <th>scan_time</th>\n",
+       "      <th>status</th>\n",
+       "      <th>ground_truth</th>\n",
+       "      <th>defect_type</th>\n",
+       "      <th>action</th>\n",
+       "      <th>confidence</th>\n",
+       "      <th>roi_coordinates</th>\n",
+       "      <th>defect_area_px</th>\n",
+       "      <th>material_wasted_pct</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>33866</th>\n",
+       "      <td>33867</td>\n",
+       "      <td>wafer_33866</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-13 20:14:29</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>33867</th>\n",
+       "      <td>33868</td>\n",
+       "      <td>wafer_33867</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-13 20:14:08</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>33868</th>\n",
+       "      <td>33869</td>\n",
+       "      <td>wafer_33868</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-13 20:14:08</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>33869</th>\n",
+       "      <td>33870</td>\n",
+       "      <td>wafer_33869</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-13 20:14:29</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>33870</th>\n",
+       "      <td>33871</td>\n",
+       "      <td>wafer_33870</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-13 20:14:31</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>823948</th>\n",
+       "      <td>823949</td>\n",
+       "      <td>wm811k_811442</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-16 20:14:06</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>823949</th>\n",
+       "      <td>823950</td>\n",
+       "      <td>wm811k_811445</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-16 20:14:07</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>823950</th>\n",
+       "      <td>823951</td>\n",
+       "      <td>wm811k_811449</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-16 20:14:07</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>823951</th>\n",
+       "      <td>823952</td>\n",
+       "      <td>wm811k_811455</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-16 20:14:07</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>823952</th>\n",
+       "      <td>823953</td>\n",
+       "      <td>wm811k_811456</td>\n",
+       "      <td>BATCH_20260317_201406</td>\n",
+       "      <td>2026-03-16 20:14:09</td>\n",
+       "      <td>PASS</td>\n",
+       "      <td>Normal</td>\n",
+       "      <td>None</td>\n",
+       "      <td>ROUTE_TO_ASSEMBLY</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>[]</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>786938 rows × 12 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "            id       wafer_id               batch_id            scan_time  \\\n",
+       "33866    33867    wafer_33866  BATCH_20260317_201406  2026-03-13 20:14:29   \n",
+       "33867    33868    wafer_33867  BATCH_20260317_201406  2026-03-13 20:14:08   \n",
+       "33868    33869    wafer_33868  BATCH_20260317_201406  2026-03-13 20:14:08   \n",
+       "33869    33870    wafer_33869  BATCH_20260317_201406  2026-03-13 20:14:29   \n",
+       "33870    33871    wafer_33870  BATCH_20260317_201406  2026-03-13 20:14:31   \n",
+       "...        ...            ...                    ...                  ...   \n",
+       "823948  823949  wm811k_811442  BATCH_20260317_201406  2026-03-16 20:14:06   \n",
+       "823949  823950  wm811k_811445  BATCH_20260317_201406  2026-03-16 20:14:07   \n",
+       "823950  823951  wm811k_811449  BATCH_20260317_201406  2026-03-16 20:14:07   \n",
+       "823951  823952  wm811k_811455  BATCH_20260317_201406  2026-03-16 20:14:07   \n",
+       "823952  823953  wm811k_811456  BATCH_20260317_201406  2026-03-16 20:14:09   \n",
+       "\n",
+       "       status ground_truth defect_type             action  confidence  \\\n",
+       "33866    PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "33867    PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "33868    PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "33869    PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "33870    PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "...       ...          ...         ...                ...         ...   \n",
+       "823948   PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "823949   PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "823950   PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "823951   PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "823952   PASS       Normal        None  ROUTE_TO_ASSEMBLY         1.0   \n",
+       "\n",
+       "       roi_coordinates  defect_area_px  material_wasted_pct  \n",
+       "33866               []               0                  0.0  \n",
+       "33867               []               0                  0.0  \n",
+       "33868               []               0                  0.0  \n",
+       "33869               []               0                  0.0  \n",
+       "33870               []               0                  0.0  \n",
+       "...                ...             ...                  ...  \n",
+       "823948              []               0                  0.0  \n",
+       "823949              []               0                  0.0  \n",
+       "823950              []               0                  0.0  \n",
+       "823951              []               0                  0.0  \n",
+       "823952              []               0                  0.0  \n",
+       "\n",
+       "[786938 rows x 12 columns]"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "pass_wafers"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "venv",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.13.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

middleware/best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3df8caca758c1959cc0f4dfc3190c4887a8d9ad053052a685b0bd2d5fb8f502
+size 6195306

middleware/dashboard.py

@@ -0,0 +1,369 @@
+"""
+Wafer Defect Analytics Dashboard — Phase 5
+Plotly Dash web application for semiconductor material waste analysis
+and predictive material requirement estimation.
+Now using the Mixed-type Wafer Defect Dataset (38,015 wafers).
+"""
+
+import os
+import pickle
+import sqlite3
+import numpy as np
+import pandas as pd
+import plotly.express as px
+import plotly.graph_objects as go
+from dash import Dash, html, dcc, callback, Input, Output, State
+
+# --- CONFIGURATION ---
+DB_PATH = os.path.join(os.path.dirname(__file__), 'wafer_control.db')
+MODEL_PATH = os.path.join(os.path.dirname(__file__), 'material_model.pkl')
+
+# Semiconductor-themed color palette
+COLORS = {
+    'bg': '#0a0e17',
+    'card': '#131a2e',
+    'card_border': '#1e2d4a',
+    'accent': '#00d4ff',
+    'accent2': '#7c3aed',
+    'accent3': '#10b981',
+    'danger': '#ef4444',
+    'warning': '#f59e0b',
+    'text': '#e2e8f0',
+    'text_muted': '#94a3b8',
+}
+
+DEFECT_COLORS = {
+    'Center': '#ef4444', 'Donut': '#f59e0b', 'Edge-Loc': '#10b981',
+    'Edge-Ring': '#3b82f6', 'Loc': '#8b5cf6', 'Random': '#ec4899',
+    'Scratch': '#06b6d4', 'Near-full': '#f97316', 'None': '#6b7280',
+    'Undetected': '#374151',
+}
+
+
+def load_data():
+    conn = sqlite3.connect(DB_PATH)
+    df = pd.read_sql_query("SELECT * FROM wafer_logs", conn)
+    conn.close()
+    df['scan_time'] = pd.to_datetime(df['scan_time'])
+    df['scan_date'] = df['scan_time'].dt.date
+    return df
+
+
+def load_model():
+    if os.path.exists(MODEL_PATH):
+        with open(MODEL_PATH, 'rb') as f:
+            return pickle.load(f)
+    return None
+
+
+# --- LOAD DATA ---
+df = load_data()
+model_pkg = load_model()
+
+# --- PRE-COMPUTE STATS ---
+total_scans = len(df)
+fail_count = len(df[df['status'] == 'FAIL'])
+pass_count = len(df[df['status'] == 'PASS'])
+pass_rate = round((pass_count / total_scans) * 100, 1)
+scrap_count = len(df[df['action'] == 'ROUTE_TO_SCRAP'])
+total_waste = round(df['material_wasted_pct'].sum(), 1)
+avg_waste = round(df[df['status'] == 'FAIL']['material_wasted_pct'].mean(), 2)
+avg_confidence = round(df[df['status'] == 'FAIL']['confidence'].mean(), 2)
+
+# Daily aggregations
+daily = df.groupby('scan_date').agg(
+    scans=('id', 'count'),
+    fails=('status', lambda x: (x == 'FAIL').sum()),
+    waste=('material_wasted_pct', lambda x: x.sum() / 100.0),
+    avg_waste=('material_wasted_pct', 'mean'),
+).reset_index()
+daily['fail_rate'] = round((daily['fails'] / daily['scans']) * 100, 1)
+daily['scan_date'] = pd.to_datetime(daily['scan_date'])
+
+# ============================================================
+# DASH APP
+# ============================================================
+app = Dash(__name__, suppress_callback_exceptions=True)
+app.title = "Wafer Defect Analytics — Semiconductor QC Dashboard"
+
+card_style = {
+    'backgroundColor': COLORS['card'], 'border': f"1px solid {COLORS['card_border']}",
+    'borderRadius': '12px', 'padding': '24px', 'marginBottom': '16px',
+}
+
+kpi_style = {
+    'backgroundColor': COLORS['card'], 'border': f"1px solid {COLORS['card_border']}",
+    'borderRadius': '12px', 'padding': '20px', 'textAlign': 'center', 'flex': '1', 'minWidth': '160px',
+}
+
+
+def make_kpi(title, value, subtitle="", color=COLORS['accent']):
+    return html.Div(style=kpi_style, children=[
+        html.P(title, style={'color': COLORS['text_muted'], 'fontSize': '12px', 'marginBottom': '4px', 'fontWeight': '500', 'textTransform': 'uppercase', 'letterSpacing': '1px'}),
+        html.H2(str(value), style={'color': color, 'fontSize': '28px', 'fontWeight': '700', 'margin': '4px 0'}),
+        html.P(subtitle, style={'color': COLORS['text_muted'], 'fontSize': '11px', 'marginTop': '4px'}),
+    ])
+
+
+def chart_layout(title):
+    return dict(
+        template='plotly_dark', paper_bgcolor='rgba(0,0,0,0)', plot_bgcolor='rgba(0,0,0,0)',
+        title=dict(text=title, font=dict(size=16, color=COLORS['text'])),
+        font=dict(color=COLORS['text_muted'], size=12),
+        margin=dict(l=40, r=20, t=50, b=40),
+        legend=dict(bgcolor='rgba(0,0,0,0)'),
+    )
+
+
+# ============================================================
+# FIGURES
+# ============================================================
+
+# 1. Defect type distribution (pie) — YOLO predictions
+defect_counts = df[df['status'] == 'FAIL']['defect_type'].value_counts().reset_index()
+defect_counts.columns = ['defect_type', 'count']
+fig_pie = px.pie(defect_counts, names='defect_type', values='count', color='defect_type',
+                 color_discrete_map=DEFECT_COLORS, hole=0.45)
+fig_pie.update_layout(**chart_layout('YOLOv8 Predicted Defect Distribution'))
+fig_pie.update_traces(textinfo='label+percent', textfont_size=11)
+
+# 2. Ground truth distribution (pie) — actual labels
+gt_counts = df[df['status'] == 'FAIL']['ground_truth'].value_counts().reset_index()
+gt_counts.columns = ['ground_truth', 'count']
+fig_gt_pie = px.pie(gt_counts.head(15), names='ground_truth', values='count', hole=0.45)
+fig_gt_pie.update_layout(**chart_layout('Ground Truth Label Distribution (Top 15)'))
+fig_gt_pie.update_traces(textinfo='label+percent', textfont_size=10)
+
+# 3. Material waste by defect type (bar)
+waste_by_type = df[df['status'] == 'FAIL'].groupby('defect_type').agg(
+    total_waste=('material_wasted_pct', lambda x: x.sum() / 100.0), count=('id', 'count'),
+).reset_index().sort_values('total_waste', ascending=True)
+
+fig_waste_bar = go.Figure()
+fig_waste_bar.add_trace(go.Bar(
+    y=waste_by_type['defect_type'], x=waste_by_type['total_waste'], orientation='h',
+    marker_color=[DEFECT_COLORS.get(d, '#6b7280') for d in waste_by_type['defect_type']],
+    text=[f"{v:.1f}" for v in waste_by_type['total_waste']], textposition='outside',
+))
+fig_waste_bar.update_layout(**chart_layout('Total Material Waste by Predicted Defect'))
+fig_waste_bar.update_xaxes(title_text='Equivalent Lost Wafers')
+
+# 4. Daily fail rate trend
+fig_trend = go.Figure()
+fig_trend.add_trace(go.Scatter(
+    x=daily['scan_date'], y=daily['fail_rate'], mode='lines+markers',
+    line=dict(color=COLORS['danger'], width=2), marker=dict(size=5),
+    name='Fail Rate %', fill='tozeroy', fillcolor='rgba(239, 68, 68, 0.1)',
+))
+fig_trend.update_layout(**chart_layout('Daily Defect Rate Over Time'))
+fig_trend.update_yaxes(title_text='Fail Rate %', range=[0, 105])
+
+# 5. Daily waste trend
+fig_waste_trend = go.Figure()
+fig_waste_trend.add_trace(go.Scatter(
+    x=daily['scan_date'], y=daily['waste'], mode='lines+markers',
+    line=dict(color=COLORS['warning'], width=2), marker=dict(size=5),
+    name='Lost Wafers', fill='tozeroy', fillcolor='rgba(245, 158, 11, 0.1)',
+))
+fig_waste_trend.update_layout(**chart_layout('Daily Material Waste Over Time'))
+fig_waste_trend.update_yaxes(title_text='Total Lost Wafers')
+
+# 6. Action breakdown
+action_counts = df['action'].value_counts().reset_index()
+action_counts.columns = ['action', 'count']
+action_colors = {'ROUTE_TO_SCRAP': COLORS['danger'], 'MOVE_TO_MICRO_STAGE': COLORS['warning'], 'ROUTE_TO_ASSEMBLY': COLORS['accent3']}
+fig_action = px.bar(action_counts, x='action', y='count', color='action',
+                    color_discrete_map=action_colors, text='count')
+fig_action.update_layout(**chart_layout('Wafer Routing Actions'))
+fig_action.update_traces(textposition='outside')
+
+# 7. Feature importance
+fig_importance = go.Figure()
+if model_pkg:
+    imp = model_pkg['metrics']['importances']
+    imp_df = pd.DataFrame({'feature': list(imp.keys()), 'importance': list(imp.values())})
+    imp_df = imp_df.sort_values('importance', ascending=True).tail(10)
+    fig_importance.add_trace(go.Bar(
+        y=imp_df['feature'], x=imp_df['importance'], orientation='h',
+        marker_color=COLORS['accent2'],
+        text=[f"{v:.3f}" for v in imp_df['importance']], textposition='outside',
+    ))
+fig_importance.update_layout(**chart_layout('Top 10 Prediction Features'))
+
+# ============================================================
+# LAYOUT
+# ============================================================
+app.layout = html.Div(style={
+    'backgroundColor': COLORS['bg'], 'minHeight': '100vh',
+    'fontFamily': "'Inter', -apple-system, BlinkMacSystemFont, sans-serif",
+    'color': COLORS['text'], 'padding': '24px 32px',
+}, children=[
+
+    # HEADER
+    html.Div(style={'marginBottom': '32px', 'borderBottom': f"1px solid {COLORS['card_border']}", 'paddingBottom': '20px'}, children=[
+        html.Div(style={'display': 'flex', 'alignItems': 'center', 'gap': '12px'}, children=[
+            html.Div(style={'width': '12px', 'height': '12px', 'borderRadius': '50%',
+                            'backgroundColor': COLORS['accent3'], 'boxShadow': f"0 0 8px {COLORS['accent3']}"}),
+            html.H1("Wafer Defect Analytics", style={
+                'margin': '0', 'fontSize': '28px', 'fontWeight': '700',
+                'background': f"linear-gradient(135deg, {COLORS['accent']}, {COLORS['accent2']})",
+                'WebkitBackgroundClip': 'text', 'WebkitTextFillColor': 'transparent',
+            }),
+        ]),
+        html.P("Mixed-type Wafer Defect Dataset — Material Waste Dashboard",
+               style={'color': COLORS['text_muted'], 'marginTop': '4px', 'fontSize': '14px'}),
+    ]),
+
+    # TABS
+    dcc.Tabs(id='tabs', value='tab-waste', style={'marginBottom': '24px'}, children=[
+        dcc.Tab(label='📊 Historical Waste Analysis', value='tab-waste', style={
+            'backgroundColor': COLORS['card'], 'color': COLORS['text_muted'],
+            'border': f"1px solid {COLORS['card_border']}", 'borderRadius': '8px 8px 0 0',
+            'padding': '12px 24px', 'fontWeight': '600',
+        }, selected_style={
+            'backgroundColor': COLORS['accent2'], 'color': '#fff',
+            'border': f"1px solid {COLORS['accent2']}", 'borderRadius': '8px 8px 0 0',
+            'padding': '12px 24px', 'fontWeight': '600',
+        }),
+        dcc.Tab(label='🔮 Material Prediction', value='tab-predict', style={
+            'backgroundColor': COLORS['card'], 'color': COLORS['text_muted'],
+            'border': f"1px solid {COLORS['card_border']}", 'borderRadius': '8px 8px 0 0',
+            'padding': '12px 24px', 'fontWeight': '600',
+        }, selected_style={
+            'backgroundColor': COLORS['accent2'], 'color': '#fff',
+            'border': f"1px solid {COLORS['accent2']}", 'borderRadius': '8px 8px 0 0',
+            'padding': '12px 24px', 'fontWeight': '600',
+        }),
+    ]),
+
+    html.Div(id='tab-content'),
+])
+
+
+# ============================================================
+# CALLBACKS
+# ============================================================
+@callback(Output('tab-content', 'children'), Input('tabs', 'value'))
+def render_tab(tab):
+    if tab == 'tab-waste':
+        return html.Div([
+            # KPI Row
+            html.Div(style={'display': 'flex', 'gap': '12px', 'flexWrap': 'wrap', 'marginBottom': '24px'}, children=[
+                make_kpi("Total Scans", f"{total_scans:,}", "wafers inspected", COLORS['accent']),
+                make_kpi("Pass Rate", f"{pass_rate}%", f"{pass_count:,} passed", COLORS['accent3']),
+                make_kpi("Fail Rate", f"{100-pass_rate}%", f"{fail_count:,} defective", COLORS['danger']),
+                make_kpi("Scrapped", f"{scrap_count:,}", "routed to scrap", COLORS['warning']),
+                make_kpi("Avg Waste/Wafer", f"{avg_waste}%", "per defective wafer", COLORS['danger']),
+                make_kpi("Avg Confidence", f"{avg_confidence}", "model certainty", COLORS['accent3']),
+            ]),
+
+            # Charts Row 1: YOLO predictions vs Ground Truth
+            html.Div(style={'display': 'grid', 'gridTemplateColumns': '1fr 1fr', 'gap': '16px', 'marginBottom': '16px'}, children=[
+                html.Div(style=card_style, children=[dcc.Graph(figure=fig_pie, config={'displayModeBar': False})]),
+                html.Div(style=card_style, children=[dcc.Graph(figure=fig_gt_pie, config={'displayModeBar': False})]),
+            ]),
+
+            # Charts Row 2
+            html.Div(style={'display': 'grid', 'gridTemplateColumns': '1fr 1fr', 'gap': '16px', 'marginBottom': '16px'}, children=[
+                html.Div(style=card_style, children=[dcc.Graph(figure=fig_waste_bar, config={'displayModeBar': False})]),
+                html.Div(style=card_style, children=[dcc.Graph(figure=fig_action, config={'displayModeBar': False})]),
+            ]),
+
+            # Trend charts
+            html.Div(style={'display': 'grid', 'gridTemplateColumns': '1fr 1fr', 'gap': '16px'}, children=[
+                html.Div(style=card_style, children=[dcc.Graph(figure=fig_trend, config={'displayModeBar': False})]),
+                html.Div(style=card_style, children=[dcc.Graph(figure=fig_waste_trend, config={'displayModeBar': False})]),
+            ]),
+        ])
+
+    elif tab == 'tab-predict':
+        model_status = "✅ Model loaded" if model_pkg else "❌ No model found"
+        model_metrics = ""
+        if model_pkg:
+            m = model_pkg['metrics']
+            model_metrics = f"R² = {m['r2']:.4f}  |  MAE = {m['mae']:.2f}%"
+
+        return html.Div([
+            html.Div(style={**card_style, 'display': 'flex', 'justifyContent': 'space-between', 'alignItems': 'center'}, children=[
+                html.Div([
+                    html.H3("Prediction Model", style={'margin': '0 0 4px 0', 'fontSize': '18px'}),
+                    html.P(model_status, style={'margin': '0', 'color': COLORS['accent3'] if model_pkg else COLORS['danger']}),
+                ]),
+                html.P(model_metrics, style={'color': COLORS['text_muted'], 'fontFamily': 'monospace'}),
+            ]),
+
+            html.Div(style=card_style, children=[
+                html.H3("Forecast Parameters", style={'marginTop': '0', 'fontSize': '18px', 'marginBottom': '20px'}),
+                html.Div(style={'display': 'grid', 'gridTemplateColumns': '1fr 1fr', 'gap': '24px'}, children=[
+                    html.Div([
+                        html.Label("Expected Daily Production (wafers)", style={'color': COLORS['text_muted'], 'fontSize': '13px'}),
+                        dcc.Slider(id='slider-scans', min=100, max=2000, step=50, value=1300,
+                                   marks={100: '100', 500: '500', 1000: '1000', 1500: '1500', 2000: '2000'},
+                                   tooltip={"placement": "bottom", "always_visible": True}),
+                    ]),
+                    html.Div([
+                        html.Label("Expected Defect Rate (%)", style={'color': COLORS['text_muted'], 'fontSize': '13px'}),
+                        dcc.Slider(id='slider-fail-rate', min=0, max=100, step=5, value=97,
+                                   marks={0: '0%', 25: '25%', 50: '50%', 75: '75%', 100: '100%'},
+                                   tooltip={"placement": "bottom", "always_visible": True}),
+                    ]),
+                ]),
+                html.Br(),
+                html.Button("🔮 Predict Material Needs", id='btn-predict', n_clicks=0, style={
+                    'backgroundColor': COLORS['accent2'], 'color': '#fff', 'border': 'none',
+                    'borderRadius': '8px', 'padding': '12px 32px', 'fontSize': '15px',
+                    'fontWeight': '600', 'cursor': 'pointer', 'width': '100%',
+                }),
+            ]),
+
+            html.Div(id='prediction-result'),
+
+            html.Div(style=card_style, children=[
+                dcc.Graph(figure=fig_importance, config={'displayModeBar': False}),
+            ]),
+        ])
+
+
+@callback(
+    Output('prediction-result', 'children'),
+    Input('btn-predict', 'n_clicks'),
+    State('slider-scans', 'value'),
+    State('slider-fail-rate', 'value'),
+    prevent_initial_call=True,
+)
+def predict(n_clicks, n_scans, fail_pct):
+    if not model_pkg:
+        return html.Div(style={**card_style, 'borderColor': COLORS['danger']}, children=[
+            html.P("❌ No model loaded.", style={'color': COLORS['danger']}),
+        ])
+
+    from material_predictor import predict_material_needs
+    model = model_pkg['model']
+    feat_cols = model_pkg['feature_cols']
+
+    fail_df = df[df['status'] == 'FAIL']
+    dist = fail_df['defect_type'].value_counts(normalize=True).to_dict()
+
+    pred = predict_material_needs(model, feat_cols, n_scans, fail_pct / 100.0, dist)
+
+    return html.Div(style={
+        **card_style, 'borderColor': COLORS['accent'],
+        'background': f"linear-gradient(135deg, {COLORS['card']}, #1a1040)",
+    }, children=[
+        html.Div(style={'display': 'flex', 'justifyContent': 'space-around', 'flexWrap': 'wrap', 'gap': '16px'}, children=[
+            make_kpi("Daily Production", f"{n_scans}", "wafers", COLORS['accent']),
+            make_kpi("Expected Defect Rate", f"{fail_pct}%", f"~{int(n_scans * fail_pct / 100)} defective", COLORS['danger']),
+            make_kpi("Avg Waste/Wafer", f"{pred['avg_waste_per_wafer']:.1f}%", "per defective wafer loss", COLORS['warning']),
+            make_kpi("Total Daily Waste", f"{pred['total_daily_waste']:.1f} wafers", "total estimated loss", COLORS['danger']),
+        ]),
+    ])
+
+
+if __name__ == '__main__':
+    print("\n" + "=" * 60)
+    print("  WAFER DEFECT ANALYTICS DASHBOARD")
+    print(f"  Data: {total_scans:,} scans | {fail_count:,} defects | {pass_count:,} pass")
+    print(f"  Model: {'Loaded ✅' if model_pkg else 'Not found ❌'}")
+    print("=" * 60)
+    print(f"\n  🌐 Open: http://127.0.0.1:8050\n")
+    app.run(debug=True, port=8050)

middleware/material_model.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47b26c94a138551231a0bb2a9a7b2dfe918c278224048c4fa12a22feacf6a0b2
+size 240685

middleware/material_predictor.py

@@ -0,0 +1,229 @@
+"""
+Material Predictor — Phase 5
+Trains a Random Forest model on historical wafer scan data to predict
+material waste percentage for future production batches.
+"""
+
+import os
+import pickle
+import sqlite3
+import numpy as np
+import pandas as pd
+from sklearn.ensemble import RandomForestRegressor
+from sklearn.model_selection import train_test_split
+from sklearn.metrics import mean_absolute_error, r2_score
+
+# --- CONFIGURATION ---
+DB_PATH = os.path.join(os.path.dirname(__file__), 'wafer_control.db')
+MODEL_PATH = os.path.join(os.path.dirname(__file__), 'material_model.pkl')
+
+# Defect types for feature engineering
+DEFECT_TYPES = ['Center', 'Donut', 'Edge-Loc', 'Edge-Ring', 'Loc', 'Random', 'Scratch', 'Near-full', 'None', 'Undetected']
+
+
+def load_data():
+    """Load wafer logs from the SQLite database into a DataFrame."""
+    conn = sqlite3.connect(DB_PATH)
+    df = pd.read_sql_query("SELECT * FROM wafer_logs", conn)
+    conn.close()
+    df['scan_time'] = pd.to_datetime(df['scan_time'])
+    return df
+
+
+def engineer_features(df):
+    """
+    Build daily-aggregated features from raw scan logs.
+    Each row = one day of production with aggregated metrics.
+    """
+    df['scan_date'] = df['scan_time'].dt.date
+    df['is_fail'] = (df['status'] == 'FAIL').astype(int)
+    df['is_scrap'] = (df['action'] == 'ROUTE_TO_SCRAP').astype(int)
+
+    # One-hot encode defect types per scan
+    for defect in DEFECT_TYPES:
+        col_name = f'is_{defect.lower().replace("-", "_")}'
+        df[col_name] = (df['defect_type'] == defect).astype(int)
+
+    # --- Aggregate by day ---
+    daily = df.groupby('scan_date').agg(
+        total_scans=('id', 'count'),
+        fail_count=('is_fail', 'sum'),
+        scrap_count=('is_scrap', 'sum'),
+        avg_confidence=('confidence', 'mean'),
+        avg_defect_area=('defect_area_px', 'mean'),
+        max_defect_area=('defect_area_px', 'max'),
+        total_waste_pct=('material_wasted_pct', 'sum'),
+        avg_waste_pct=('material_wasted_pct', 'mean'),
+        # Defect type counts per day
+        center_count=('is_center', 'sum'),
+        donut_count=('is_donut', 'sum'),
+        edge_loc_count=('is_edge_loc', 'sum'),
+        edge_ring_count=('is_edge_ring', 'sum'),
+        loc_count=('is_loc', 'sum'),
+        random_count=('is_random', 'sum'),
+        scratch_count=('is_scratch', 'sum'),
+        near_full_count=('is_near_full', 'sum'),
+        pass_count=('is_none', 'sum'),
+    ).reset_index()
+
+    # --- Compute waste among defective wafers only ---
+    defective_daily = df[df['status'] == 'FAIL'].groupby('scan_date').agg(
+        avg_waste_defective=('material_wasted_pct', 'mean'),
+        avg_defect_area_fail=('defect_area_px', 'mean'),
+        avg_confidence_fail=('confidence', 'mean'),
+    ).reset_index()
+
+    daily = daily.merge(defective_daily, on='scan_date', how='left')
+    daily['avg_waste_defective'] = daily['avg_waste_defective'].fillna(0)
+    daily['avg_defect_area_fail'] = daily['avg_defect_area_fail'].fillna(0)
+    daily['avg_confidence_fail'] = daily['avg_confidence_fail'].fillna(0)
+
+    # Derived ratios
+    daily['fail_rate'] = daily['fail_count'] / daily['total_scans']
+    daily['scrap_rate'] = daily['scrap_count'] / daily['total_scans']
+
+    # Time features
+    daily['scan_date'] = pd.to_datetime(daily['scan_date'])
+    daily['day_of_week'] = daily['scan_date'].dt.dayofweek
+    daily['day_index'] = (daily['scan_date'] - daily['scan_date'].min()).dt.days
+
+    return daily
+
+
+def train_model(daily):
+    """Train a Random Forest to predict avg material waste among defective wafers."""
+    feature_cols = [
+        'total_scans', 'fail_count', 'scrap_count', 'avg_confidence',
+        'avg_defect_area', 'max_defect_area', 'fail_rate', 'scrap_rate',
+        'avg_defect_area_fail', 'avg_confidence_fail',
+        'center_count', 'donut_count', 'edge_loc_count', 'edge_ring_count',
+        'loc_count', 'random_count', 'scratch_count', 'near_full_count',
+        'pass_count', 'day_of_week', 'day_index'
+    ]
+
+    target = 'avg_waste_defective'
+
+    X = daily[feature_cols]
+    y = daily[target]
+
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
+
+    model = RandomForestRegressor(
+        n_estimators=100,
+        max_depth=10,
+        random_state=42,
+        n_jobs=-1
+    )
+    model.fit(X_train, y_train)
+
+    # Evaluate
+    y_pred = model.predict(X_test)
+    mae = mean_absolute_error(y_test, y_pred)
+    r2 = r2_score(y_test, y_pred)
+
+    print(f"\n{'=' * 50}")
+    print(f"  MODEL EVALUATION (target: avg waste % per wafer)")
+    print(f"  Mean Absolute Error:  {mae:.2f}%")
+    print(f"  R² Score:             {r2:.4f}")
+    print(f"{'=' * 50}")
+
+    # Feature importance
+    importance = pd.Series(model.feature_importances_, index=feature_cols).sort_values(ascending=False)
+    print(f"\n  Top 5 Feature Importances:")
+    for feat, imp in importance.head(5).items():
+        print(f"    {feat:25s} {imp:.4f}")
+
+    return model, feature_cols, {'mae': mae, 'r2': r2, 'importances': importance.to_dict()}
+
+
+def predict_material_needs(model, feature_cols, total_scans, fail_rate, defect_distribution):
+    """
+    Predict material waste for a hypothetical future production day.
+    """
+    fail_count = int(total_scans * fail_rate)
+    pass_count = total_scans - fail_count
+
+    features = {
+        'total_scans': total_scans,
+        'fail_count': fail_count,
+        'scrap_count': int(fail_count * defect_distribution.get('Center', 0) +
+                          fail_count * defect_distribution.get('Near-full', 0)),
+        'avg_confidence': 0.95,
+        'avg_defect_area': 1500,
+        'max_defect_area': 2704,
+        'fail_rate': fail_rate,
+        'scrap_rate': defect_distribution.get('Center', 0) + defect_distribution.get('Near-full', 0),
+        'avg_defect_area_fail': 1500,
+        'avg_confidence_fail': 0.85,
+        'center_count': int(fail_count * defect_distribution.get('Center', 0)),
+        'donut_count': int(fail_count * defect_distribution.get('Donut', 0)),
+        'edge_loc_count': int(fail_count * defect_distribution.get('Edge-Loc', 0)),
+        'edge_ring_count': int(fail_count * defect_distribution.get('Edge-Ring', 0)),
+        'loc_count': int(fail_count * defect_distribution.get('Loc', 0)),
+        'random_count': int(fail_count * defect_distribution.get('Random', 0)),
+        'scratch_count': int(fail_count * defect_distribution.get('Scratch', 0)),
+        'near_full_count': int(fail_count * defect_distribution.get('Near-full', 0)),
+        'pass_count': pass_count,
+        'day_of_week': 2,
+        'day_index': 30,
+    }
+
+    X = pd.DataFrame([features])[feature_cols]
+    avg_waste_per_wafer = model.predict(X)[0]
+    total_waste_wafers = (avg_waste_per_wafer / 100.0) * fail_count
+    return {
+        'avg_waste_per_wafer': round(avg_waste_per_wafer, 2),
+        'total_daily_waste': round(total_waste_wafers, 1),
+        'total_scans': total_scans,
+        'fail_rate': fail_rate,
+    }
+
+
+if __name__ == '__main__':
+    print("=" * 50)
+    print("  MATERIAL WASTE PREDICTOR — Training")
+    print("=" * 50)
+
+    # 1. Load and engineer features
+    print("\nLoading scan data...")
+    raw_df = load_data()
+    print(f"  Total records: {len(raw_df)}")
+    print(f"  PASS: {len(raw_df[raw_df['status'] == 'PASS'])}")
+    print(f"  FAIL: {len(raw_df[raw_df['status'] == 'FAIL'])}")
+
+    print("Engineering daily features...")
+    daily_df = engineer_features(raw_df)
+    print(f"  Training days: {len(daily_df)}")
+
+    # 2. Train
+    print("\nTraining Random Forest model...")
+    trained_model, feat_cols, metrics = train_model(daily_df)
+
+    # 3. Save
+    model_package = {
+        'model': trained_model,
+        'feature_cols': feat_cols,
+        'metrics': metrics,
+    }
+    with open(MODEL_PATH, 'wb') as f:
+        pickle.dump(model_package, f)
+    print(f"\nModel saved to: {MODEL_PATH}")
+
+    # 4. Demo prediction
+    print(f"\n{'=' * 50}")
+    print("  DEMO PREDICTION")
+    print(f"{'=' * 50}")
+
+    demo_distribution = {
+        'Center': 0.15, 'Edge-Ring': 0.37, 'Edge-Loc': 0.06,
+        'Donut': 0.23, 'Random': 0.03, 'Scratch': 0.03,
+        'Loc': 0.10, 'Near-full': 0.01
+    }
+
+    pred = predict_material_needs(trained_model, feat_cols,
+                                   total_scans=1300, fail_rate=0.97,
+                                   defect_distribution=demo_distribution)
+    print(f"  Scenario: 1,300 wafers/day, 97% defect rate")
+    print(f"  Predicted avg waste per wafer: {pred['avg_waste_per_wafer']:.2f}%")
+    print(f"  Predicted total daily waste:   {pred['total_daily_waste']:.1f} equivalent wafers")
+    print(f"{'=' * 50}")

middleware/robot_controller.py

@@ -0,0 +1,278 @@
+"""
+Robotic Control Middleware — Full Production Scan
+Phase 1: Loads Mixed-type Wafer Defect Dataset, runs YOLOv8 on defective wafers.
+Phase 2: Loads ALL passed wafers from WM-811K dataset (direct insert, no YOLO needed).
+"""
+
+import os
+import sys
+import pickle
+import sqlite3
+import random
+import cv2
+import time
+import numpy as np
+from datetime import datetime, timedelta
+from ultralytics import YOLO
+
+# Fix for old Pandas architecture in WM-811K pickle
+import pandas.core.indexes
+sys.modules['pandas.indexes'] = pandas.core.indexes
+import pandas as pd
+
+# --- CONFIGURATION ---
+NPZ_PATH = os.path.expanduser(
+    '~/.cache/kagglehub/datasets/co1d7era/mixedtype-wafer-defect-datasets/versions/4/Wafer_Map_Datasets.npz'
+)
+WM811K_PATH = os.path.expanduser(
+    '~/.cache/kagglehub/datasets/qingyi/wm811k-wafer-map/versions/1/LSWMD.pkl'
+)
+MODEL_PATH = 'middleware/best.pt'
+DB_PATH = os.path.join('middleware', 'wafer_control.db')
+
+# Defect names matching the 8-dim one-hot encoding order in the dataset
+DEFECT_NAMES = ['Center', 'Donut', 'Edge-Loc', 'Edge-Ring', 'Loc', 'Near-full', 'Random', 'Scratch']
+
+
+def setup_database():
+    """Creates a fresh wafer_logs table with ground_truth column."""
+    os.makedirs('middleware', exist_ok=True)
+    conn = sqlite3.connect(DB_PATH)
+    cursor = conn.cursor()
+
+    cursor.execute('DROP TABLE IF EXISTS wafer_logs')
+    cursor.execute('''
+        CREATE TABLE wafer_logs (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            wafer_id TEXT,
+            batch_id TEXT,
+            scan_time TEXT,
+            status TEXT,
+            ground_truth TEXT,
+            defect_type TEXT,
+            action TEXT,
+            confidence REAL,
+            roi_coordinates TEXT,
+            defect_area_px INTEGER,
+            material_wasted_pct REAL
+        )
+    ''')
+    conn.commit()
+    return conn
+
+
+def decode_label(one_hot):
+    """Convert 8-dim one-hot label to human-readable defect string."""
+    active = np.where(one_hot == 1)[0]
+    if len(active) == 0:
+        return 'Normal'
+    return '+'.join([DEFECT_NAMES[i] for i in active])
+
+
+def wafer_to_image(wafer_map):
+    """Convert a 52x52 wafer map array to a 3-channel BGR image for YOLOv8."""
+    img = np.zeros(wafer_map.shape, dtype=np.uint8)
+    img[wafer_map == 1] = 127   # Normal die → gray
+    img[wafer_map == 2] = 255   # Broken die → white
+    img[wafer_map == 3] = 255   # Treat 3 as defect too (rare edge artifact)
+    # YOLOv8 expects 3-channel (BGR) images
+    img_bgr = cv2.cvtColor(img, cv2.COLOR_GRAY2BGR)
+    return img_bgr
+
+
+def compute_defect_area(coords):
+    """Calculate bounding box area in pixels from [x1, y1, x2, y2]."""
+    if not coords or len(coords) != 4:
+        return 0
+    x1, y1, x2, y2 = coords
+    return max(0, (x2 - x1) * (y2 - y1))
+
+
+def run_production_scan(conn, model, wafer_maps, labels, batch_id, start_time):
+    """Process all wafers: YOLO inference on defective, direct insert for normals."""
+    cursor = conn.cursor()
+    total = len(wafer_maps)
+
+    for i in range(total):
+        wafer_id = f"wafer_{i}"
+        ground_truth = decode_label(labels[i])
+
+        # Distribute realistic timestamps with Gaussian noise to create natural defect spikes
+        base_day = (i / total) * 30
+        noisy_day = base_day + random.gauss(0, 4) # High variance for defects
+        days_offset = int(max(0, min(29, noisy_day)))
+        scan_time = start_time + timedelta(
+            days=days_offset,
+            seconds=random.randint(0, 68)
+        )
+        scan_time_str = scan_time.strftime("%Y-%m-%d %H:%M:%S")
+
+        if ground_truth == 'Normal':
+            # PASS wafer — no YOLO needed
+            cursor.execute('''
+                INSERT INTO wafer_logs 
+                (wafer_id, batch_id, scan_time, status, ground_truth, defect_type, action, confidence, roi_coordinates, defect_area_px, material_wasted_pct)
+                VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+            ''', (wafer_id, batch_id, scan_time_str, "PASS", ground_truth, "None", "ROUTE_TO_ASSEMBLY", 1.0, "[]", 0, 0.0))
+        else:
+            # Defective wafer — convert to image and run YOLO
+            img = wafer_to_image(wafer_maps[i])
+            wafer_area_px = img.shape[0] * img.shape[1]  # 52*52 = 2704
+
+            results = model.predict(source=img, conf=0.25, verbose=False)
+            boxes = results[0].boxes
+
+            if len(boxes) > 0:
+                box = boxes[0]
+                class_id = int(box.cls[0].item())
+                defect_type = model.names[class_id]
+                confidence = round(box.conf[0].item(), 2)
+
+                x1, y1, x2, y2 = [int(x) for x in box.xyxy[0].tolist()]
+                coords = [x1, y1, x2, y2]
+
+                status = "FAIL"
+                action = "ROUTE_TO_SCRAP" if defect_type in ["Center", "Near-full"] else "MOVE_TO_MICRO_STAGE"
+            else:
+                # YOLO didn't detect anything (could be mixed pattern it can't see)
+                status = "FAIL"
+                defect_type = "Undetected"
+                action = "MOVE_TO_MICRO_STAGE"
+                confidence = 0.0
+                coords = []
+
+            defect_area = compute_defect_area(coords)
+            material_wasted_pct = round((defect_area / wafer_area_px) * 100, 2) if defect_area > 0 else 0.0
+
+            cursor.execute('''
+                INSERT INTO wafer_logs 
+                (wafer_id, batch_id, scan_time, status, ground_truth, defect_type, action, confidence, roi_coordinates, defect_area_px, material_wasted_pct)
+                VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+            ''', (wafer_id, batch_id, scan_time_str, status, ground_truth, defect_type, action, confidence, str(coords), defect_area, material_wasted_pct))
+
+        # Commit in batches
+        if (i + 1) % 500 == 0:
+            conn.commit()
+            print(f"  Processed {i + 1}/{total} wafers...")
+
+    conn.commit()
+
+
+def insert_wm811k_passed(conn, batch_id, start_time):
+    """Load ALL passed wafers from WM-811K and insert directly into DB."""
+    print("Loading WM-811K dataset...")
+    with open(WM811K_PATH, 'rb') as f:
+        wm_df = pickle.load(f, encoding='latin1')
+
+    wm_df['failure_class'] = wm_df['failureType'].apply(lambda x: x[0][0] if len(x) > 0 else 'None')
+    passed = wm_df[(wm_df['failure_class'] == 'None') | (wm_df['failure_class'] == 'none')]
+    passed = passed[passed['waferMap'].apply(lambda x: isinstance(x, np.ndarray) and x.size > 0)]
+
+    total = len(passed)
+    print(f"  Found {total:,} passed wafers in WM-811K")
+    print(f"  Inserting all into database...")
+
+    cursor = conn.cursor()
+    rows = []
+    for i, (index, row) in enumerate(passed.iterrows()):
+        # Stable production schedule with low variance for normal wafers
+        base_day = (i / total) * 30
+        noisy_day = base_day + random.gauss(0, 0.5) 
+        days_offset = int(max(0, min(29, noisy_day)))
+        scan_time = start_time + timedelta(
+            days=days_offset,
+            seconds=random.randint(0, 3)
+        )
+        rows.append((
+            f"wm811k_{index}", batch_id, scan_time.strftime("%Y-%m-%d %H:%M:%S"),
+            "PASS", "Normal", "None", "ROUTE_TO_ASSEMBLY", 1.0, "[]", 0, 0.0
+        ))
+
+        if len(rows) >= 10000:
+            cursor.executemany('''
+                INSERT INTO wafer_logs
+                (wafer_id, batch_id, scan_time, status, ground_truth, defect_type, action, confidence, roi_coordinates, defect_area_px, material_wasted_pct)
+                VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+            ''', rows)
+            conn.commit()
+            rows = []
+            print(f"  Inserted {i + 1:,}/{total:,} passed wafers...")
+
+    if rows:
+        cursor.executemany('''
+            INSERT INTO wafer_logs
+            (wafer_id, batch_id, scan_time, status, ground_truth, defect_type, action, confidence, roi_coordinates, defect_area_px, material_wasted_pct)
+            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+        ''', rows)
+        conn.commit()
+
+    print(f"  Done! Inserted {total:,} passed wafers.")
+    return total
+
+
+if __name__ == '__main__':
+    print("=" * 60)
+    print("  ROBOTIC CONTROL MIDDLEWARE — HYBRID PRODUCTION SCAN")
+    print("=" * 60)
+
+    # 1. Load Mixed-type dataset
+    print("\nPhase 1: Loading Mixed-type Wafer Defect Dataset...")
+    data = np.load(NPZ_PATH)
+    X = data['arr_0']
+    Y = data['arr_1']
+
+    normals_mixed = sum(1 for y in Y if np.sum(y) == 0)
+    defective = len(X) - normals_mixed
+
+    print(f"  Mixed-type: {len(X):,} total ({defective:,} defective + {normals_mixed:,} normal)")
+    print(f"  WM-811K:    ~786K passed wafers")
+
+    # 2. Setup
+    db_connection = setup_database()
+    wafer_model = YOLO(MODEL_PATH)
+
+    batch_id = f"BATCH_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
+    start_time = datetime.now() - timedelta(days=30)
+
+    print(f"\n  Batch ID: {batch_id}")
+    print(f"  Scan window: {start_time.strftime('%Y-%m-%d')} → {datetime.now().strftime('%Y-%m-%d')}")
+
+    try:
+        # PHASE 1: YOLOv8 on Mixed-type defective wafers
+        print(f"\n{'=' * 60}")
+        print(f"  PHASE 1: YOLOv8 Inference ({len(X):,} Mixed-type wafers)")
+        print(f"{'=' * 60}\n")
+
+        t0 = time.time()
+        run_production_scan(db_connection, wafer_model, X, Y, batch_id, start_time)
+        t1 = time.time()
+        print(f"\n  Phase 1 complete: {t1 - t0:.1f}s")
+
+        # PHASE 2: All passed wafers from WM-811K
+        print(f"\n{'=' * 60}")
+        print(f"  PHASE 2: WM-811K Passed Wafers (all ~786K)")
+        print(f"{'=' * 60}\n")
+
+        passed_count = insert_wm811k_passed(db_connection, batch_id, start_time)
+        t2 = time.time()
+
+        total = len(X) + passed_count
+        print(f"\n{'=' * 60}")
+        print(f"  SCAN COMPLETE")
+        print(f"  Defective (Mixed-type): {defective:,}")
+        print(f"  Normal (Mixed-type):    {normals_mixed:,}")
+        print(f"  Passed (WM-811K):       {passed_count:,}")
+        print(f"  Total records:          {total:,}")
+        print(f"  Pass rate:              {(normals_mixed + passed_count) / total * 100:.1f}%")
+        print(f"  Time elapsed:           {t2 - t0:.1f}s")
+        print(f"  Database:               {DB_PATH}")
+        print(f"{'=' * 60}")
+
+    except Exception as e:
+        print(f"\nError during scan: {e}")
+        import traceback
+        traceback.print_exc()
+
+    finally:
+        db_connection.close()
+        print("Database connection closed.")

middleware/wafer_control.db

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f38b7b11d6ac3e7f8292a4bc48c23de272ae29c0c728ef62f62a42a4bc42406b
+size 90316800

notebooks/01_data_exploration.ipynb

@@ -0,0 +1,399 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "03c79b89",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "e517de18",
+   "metadata": {},
+   "outputs": [
+    {
+     "ename": "ModuleNotFoundError",
+     "evalue": "No module named 'kagglehub'",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mModuleNotFoundError\u001b[0m                       Traceback (most recent call last)",
+      "Cell \u001b[1;32mIn[1], line 2\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;21;01msys\u001b[39;00m\n\u001b[1;32m----> 2\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;21;01mkagglehub\u001b[39;00m\n\u001b[0;32m      3\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;21;01mpandas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;21;01mpd\u001b[39;00m\n\u001b[0;32m      4\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;21;01mnumpy\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;28;01mas\u001b[39;00m\u001b[38;5;250m \u001b[39m\u001b[38;5;21;01mnp\u001b[39;00m\n",
+      "\u001b[1;31mModuleNotFoundError\u001b[0m: No module named 'kagglehub'"
+     ]
+    }
+   ],
+   "source": [
+    "import sys\n",
+    "import kagglehub\n",
+    "import pandas as pd\n",
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "import os\n",
+    "import pickle"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "f7ac0941",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas.core.indexes\n",
+    "sys.modules['pandas.indexes'] = pandas.core.indexes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "82aec67e",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Downloading dataset from Kaggle...\n"
+     ]
+    },
+    {
+     "ename": "NameError",
+     "evalue": "name 'kagglehub' is not defined",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
+      "Cell \u001b[1;32mIn[3], line 2\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[38;5;28mprint\u001b[39m(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mDownloading dataset from Kaggle...\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m----> 2\u001b[0m path \u001b[38;5;241m=\u001b[39m \u001b[43mkagglehub\u001b[49m\u001b[38;5;241m.\u001b[39mdataset_download(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mqingyi/wm811k-wafer-map\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[0;32m      3\u001b[0m \u001b[38;5;28mprint\u001b[39m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mDataset downloaded to: \u001b[39m\u001b[38;5;132;01m{\u001b[39;00mpath\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m)\n",
+      "\u001b[1;31mNameError\u001b[0m: name 'kagglehub' is not defined"
+     ]
+    }
+   ],
+   "source": [
+    "print(\"Downloading dataset from Kaggle...\")\n",
+    "path = kagglehub.dataset_download(\"qingyi/wm811k-wafer-map\")\n",
+    "print(f\"Dataset downloaded to: {path}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "4dea4d86",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Loading dataset with latin1 encoding (this might take a minute)...\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/var/folders/_b/4pz7ygss6y7c564mcqz3grmw0000gn/T/ipykernel_3677/1757211552.py:6: VisibleDeprecationWarning: dtype(): align should be passed as Python or NumPy boolean but got `align=0`. Did you mean to pass a tuple to create a subarray type? (Deprecated NumPy 2.4)\n",
+      "  df = pickle.load(f, encoding='latin1')\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Success! Total wafers in dataset: 811457\n"
+     ]
+    }
+   ],
+   "source": [
+    "#Making the file path and loading the file in this venv\n",
+    "file_path = os.path.join(path, 'LSWMD.pkl')\n",
+    "\n",
+    "print(\"Loading dataset with latin1 encoding (this might take a minute)...\")\n",
+    "with open(file_path, 'rb') as f:\n",
+    "    df = pickle.load(f, encoding='latin1')\n",
+    "print(f\"Success! Total wafers in dataset: {len(df)}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "bd97a8e5",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Data cleaned! Now you can view it.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 1. Clean the nested failure type column to create 'failure_class'\n",
+    "df['failure_class'] = df['failureType'].apply(lambda x: x[0][0] if len(x) > 0 else 'None')\n",
+    "\n",
+    "# 2. Filter out the perfect wafers to create the 'defective_wafers' subset\n",
+    "defective_wafers = df[(df['failure_class'] != 'None') & (df['failure_class'] != 'none')]\n",
+    "\n",
+    "print(\"Data cleaned! Now you can view it.\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "780e3a68",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Top 5 rows of the dataset:\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>waferMap</th>\n",
+       "      <th>dieSize</th>\n",
+       "      <th>failure_class</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...</td>\n",
+       "      <td>1683.0</td>\n",
+       "      <td>none</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...</td>\n",
+       "      <td>1683.0</td>\n",
+       "      <td>none</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...</td>\n",
+       "      <td>1683.0</td>\n",
+       "      <td>none</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...</td>\n",
+       "      <td>1683.0</td>\n",
+       "      <td>none</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>[[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...</td>\n",
+       "      <td>1683.0</td>\n",
+       "      <td>none</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                            waferMap  dieSize failure_class\n",
+       "0  [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...   1683.0          none\n",
+       "1  [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...   1683.0          none\n",
+       "2  [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...   1683.0          none\n",
+       "3  [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...   1683.0          none\n",
+       "4  [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,...   1683.0          none"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Shape of the first defective wafer array: (45, 48)\n",
+      "\n",
+      "The raw 2D array data (Notice the 0s, 1s, and 2s):\n",
+      "[[0 0 0 ... 0 0 0]\n",
+      " [0 0 0 ... 0 0 0]\n",
+      " [0 0 0 ... 0 0 0]\n",
+      " ...\n",
+      " [0 0 0 ... 0 0 0]\n",
+      " [0 0 0 ... 0 0 0]\n",
+      " [0 0 0 ... 0 0 0]]\n"
+     ]
+    }
+   ],
+   "source": [
+    "#first look at the data\n",
+    "print(\"Top 5 rows of the dataset:\")\n",
+    "display(df[['waferMap', 'dieSize', 'failure_class']].head())\n",
+    "\n",
+    "# Look at exactly what the 2D array looks like for the first defective wafer\n",
+    "first_defect_index = defective_wafers.index[0]\n",
+    "first_defect_array = defective_wafers.loc[first_defect_index, 'waferMap']\n",
+    "\n",
+    "print(\"\\nShape of the first defective wafer array:\", first_defect_array.shape)\n",
+    "print(\"\\nThe raw 2D array data (Notice the 0s, 1s, and 2s):\")\n",
+    "print(first_defect_array)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "899f308a",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "The FULL raw 2D array data:\n",
+      "[[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 2 1 1 2 0 0 0 0 0 0 0 0 0\n",
+      "  0 0 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0\n",
+      "  0 0 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0\n",
+      "  0 0 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1\n",
+      "  0 0 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 2 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 2 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2\n",
+      "  1 1 1 2 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1\n",
+      "  1 2 1 1 1 1 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 2 1\n",
+      "  1 1 1 1 1 1 1 0 0 0 0 0]\n",
+      " [0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2\n",
+      "  1 1 1 1 1 1 2 1 0 0 0 0]\n",
+      " [0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 1 1 1\n",
+      "  1 1 1 1 2 2 1 1 0 0 0 0]\n",
+      " [0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 2 2 2 2 1 1 2 1 0 0 0]\n",
+      " [0 0 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1\n",
+      "  2 2 2 2 2 1 1 1 1 2 0 0]\n",
+      " [0 0 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 1 2 2\n",
+      "  2 2 2 2 2 2 1 1 1 2 0 0]\n",
+      " [0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2\n",
+      "  2 2 2 2 2 2 2 2 2 1 0 0]\n",
+      " [0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2\n",
+      "  2 2 2 2 1 1 1 1 1 1 1 0]\n",
+      " [0 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2\n",
+      "  2 2 2 2 2 2 2 2 1 1 1 0]\n",
+      " [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1\n",
+      "  2 2 2 2 2 2 2 1 2 1 1 0]\n",
+      " [1 1 1 1 2 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1\n",
+      "  1 1 1 2 1 1 1 1 1 1 1 2]\n",
+      " [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  2 1 1 1 1 1 1 1 1 1 1 2]\n",
+      " [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 2 1\n",
+      "  1 1 1 1 1 1 1 1 1 1 1 2]\n",
+      " [2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 2 1 1 1 2 1 1 2 2]\n",
+      " [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 2 1 1 1 2]\n",
+      " [2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 2 1 1 1 1 1 1 1 2]\n",
+      " [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 2 1 2 1 1 1 1 1 1 2]\n",
+      " [2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 1 1 1 0]\n",
+      " [2 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 1 1 2 0]\n",
+      " [0 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2\n",
+      "  1 1 1 1 1 1 1 1 1 1 1 0]\n",
+      " [0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 1 1 1 0]\n",
+      " [0 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 1 1 0 0]\n",
+      " [0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 1 1 0 0]\n",
+      " [0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 1 0 0 0]\n",
+      " [0 0 0 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 1 0 0 0]\n",
+      " [0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 1 0 0 0 0]\n",
+      " [0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 1 1 1 0 0 0 0 0]\n",
+      " [0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2\n",
+      "  1 1 1 1 1 1 1 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 2 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 2 1 1 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1\n",
+      "  1 1 2 1 1 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1\n",
+      "  1 1 1 1 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 2 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 2 1 1 1 1\n",
+      "  2 1 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1\n",
+      "  2 0 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0\n",
+      "  0 0 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0\n",
+      "  0 0 0 0 0 0 0 0 0 0 0 0]\n",
+      " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 2 0 0 0 0 0 0\n",
+      "  0 0 0 0 0 0 0 0 0 0 0 0]]\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Force NumPy to print the entire array without truncation\n",
+    "np.set_printoptions(threshold=10000)\n",
+    "\n",
+    "print(\"The FULL raw 2D array data:\")\n",
+    "print(first_defect_array)\n",
+    "\n",
+    "# Reset it back to default right after so future arrays don't break your terminal\n",
+    "np.set_printoptions(threshold=1000)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "wafer_gpu",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.25"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

requirements.txt

@@ -0,0 +1,16 @@
+# Data handling & EDA
+kagglehub
+pandas
+numpy
+matplotlib
+jupyter
+
+# Computer Vision & AI Tools
+opencv-python
+ultralytics
+scikit-learn
+
+# Web Application & AI Chatbot
+fastapi
+uvicorn[standard]
+google-genai

src/batch_inference.py

@@ -0,0 +1,19 @@
+from ultralytics import YOLO
+
+def test_all_validation_images():
+    print("Loading your custom-trained wafer brain...")
+    # Pointing to your exact model path from earlier
+    model_path = 'middleware/best.pt'
+    model = YOLO(model_path)
+
+    print("Running inference on ALL validation images...")
+    # Instead of one image, we hand it the entire validation folder
+    val_dir = 'data/yolo_dataset/images/val'
+
+    # The AI will automatically loop through all 5,000+ images!
+    results = model.predict(source=val_dir, save=True, conf=0.25)
+    
+    print("\nMassive inference complete! Look in the newest 'predict' folder to see thousands of drawn bounding boxes.")
+
+if __name__ == '__main__':
+    test_all_validation_images()