Upload 6 files

.dockerignore

@@ -0,0 +1,13 @@
+__pycache__
+*.pyc
+*.pyo
+*.pyd
+.env
+.git
+.gitignore
+*.keras
+*.h5
+*.ckpt
+*.ckpt.index
+.vscode
+.idea

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+efficientnet_b0_best.keras filter=lfs diff=lfs merge=lfs -text

Dockerfile

@@ -0,0 +1,42 @@
+# Use official Python slim image
+FROM python:3.11-slim
+
+# Prevent Python from writing .pyc files and buffer stdout/stderr
+ENV PYTHONDONTWRITEBYTECODE=1
+ENV PYTHONUNBUFFERED=1
+
+# Set working dir
+WORKDIR /app
+
+# Install system dependencies required for OpenCV and fonts, etc.
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    build-essential \
+    wget \
+    ca-certificates \
+    ffmpeg \
+    libsm6 \
+    libxext6 \
+    libxrender1 \
+    libglib2.0-0 \
+    && apt-get clean && rm -rf /var/lib/apt/lists/*
+
+# Copy requirements first for better caching
+COPY requirements.txt /app/requirements.txt
+
+# Upgrade pip, install dependencies
+RUN pip install --upgrade pip
+RUN pip install --no-cache-dir -r /app/requirements.txt
+
+# Copy app source code
+COPY . /app
+
+# Create model directory (if you will mount model at runtime)
+RUN mkdir -p /app/model && chmod -R 755 /app/model
+
+# Expose port that HF Spaces expects
+EXPOSE 8080
+
+# Use gunicorn to run the Flask app
+# --bind 0.0.0.0:8080 to listen on all interfaces
+# Adjust workers/threads according to memory (lower for HF free spaces)
+CMD ["gunicorn", "--bind", "0.0.0.0:8080", "app:app", "--workers", "2", "--threads", "2", "--timeout", "120"]

app.py

@@ -0,0 +1,173 @@
+import os
+os.environ["TF_CPP_MIN_LOG_LEVEL"] = "2"  # reduce TF logs
+
+import matplotlib
+matplotlib.use('Agg')  # headless backend
+
+import cv2
+import numpy as np
+import pandas as pd
+from datetime import datetime
+from flask import Flask, render_template, request, redirect, send_file
+from tensorflow.keras.applications.mobilenet_v2 import preprocess_input
+from tensorflow.keras.models import load_model
+from fpdf import FPDF
+import matplotlib.pyplot as plt
+
+# Optional: ability to download model from Hugging Face Hub if not present
+# Uncomment if you want runtime download (requires huggingface_hub in requirements)
+# from huggingface_hub import hf_hub_download
+
+# -----------------------------
+# Flask Config
+# -----------------------------
+app = Flask(__name__)
+app.config["UPLOAD_FOLDER"] = "static/uploads"
+os.makedirs(app.config["UPLOAD_FOLDER"], exist_ok=True)
+
+# -----------------------------
+# Model path
+MODEL_DIR = "model"
+os.makedirs(MODEL_DIR, exist_ok=True)
+MODEL_PATH = os.path.join(MODEL_DIR, "efficientnet_b0_best.keras")
+
+# If you want to download the model from HF Hub at runtime (optional)
+# Replace "username/repo" and "efficientnet_b0_best.keras" below with your model repo
+# try:
+#     if not os.path.exists(MODEL_PATH):
+#         hf_hub_download(repo_id="your-username/your-model-repo", filename="efficientnet_b0_best.keras", local_dir=MODEL_DIR)
+# except Exception as e:
+#     print("Could not download model from HF Hub:", e)
+
+# Load Keras model (ensure the .keras file exists in model/)
+if not os.path.exists(MODEL_PATH):
+    raise FileNotFoundError(f"Model file not found at {MODEL_PATH}. Place your .keras model in the /model folder.")
+
+best_model = load_model(MODEL_PATH)
+
+IMG_SIZE = 128
+
+# Class labels (no label encoder needed)
+CLASS_LABELS = ['biological', 'brown-glass', 'cardboard', 'green-glass', 
+                'metal', 'paper', 'plastic', 'shoes', 'trash', 'white-glass']
+
+# Waste categories
+RECYCLABLE = ["brown-glass", "green-glass", "white-glass", "metal", "plastic", "paper", "cardboard"]
+NON_RECYCLABLE = ["trash", "biological", "shoes"]
+
+# Initialize statistics
+stats = {}
+
+# -----------------------------
+# Preprocess image
+def preprocess_image(file_path):
+    img = cv2.imread(file_path)
+    img_rgb = cv2.cvtColor(cv2.resize(img, (IMG_SIZE, IMG_SIZE)), cv2.COLOR_BGR2RGB)
+    img_input = preprocess_input(img_rgb.astype("float32"))
+    img_input = np.expand_dims(img_input, axis=0)
+    return img_rgb, img_input
+
+# -----------------------------
+# Log predictions
+def log_prediction(class_label):
+    log_df = pd.DataFrame([[datetime.now(), class_label]], columns=["Timestamp", "Class"])
+    try:
+        old_df = pd.read_csv("waste_log.csv")
+        new_df = pd.concat([old_df, log_df], ignore_index=True)
+    except FileNotFoundError:
+        new_df = log_df
+    new_df.to_csv("waste_log.csv", index=False)
+
+# -----------------------------
+# PDF Report
+def generate_pdf_report():
+    pdf = FPDF()
+    pdf.add_page()
+    pdf.set_font("Arial", size=14)
+    pdf.cell(200, 10, txt="Waste Classification Report", ln=True, align="C")
+    pdf.ln(10)
+
+    total_items = sum(stats.values()) if stats else 0
+    pdf.set_font("Arial", size=12)
+    pdf.cell(0, 10, txt=f"Total Items Processed: {total_items}", ln=True)
+
+    for category, count in stats.items():
+        pdf.cell(0, 10, txt=f"{category}: {count}", ln=True)
+
+    pdf_file = "waste_report.pdf"
+    pdf.output(pdf_file)
+    return pdf_file
+
+# -----------------------------
+# Routes
+@app.route("/", methods=["GET", "POST"])
+def index():
+    if request.method == "POST":
+        file = request.files.get("file")
+        if file is None or file.filename == "":
+            return redirect(request.url)
+
+        file_path = os.path.join(app.config["UPLOAD_FOLDER"], file.filename)
+        file.save(file_path)
+
+        # Preprocess & predict
+        img_rgb, img_input = preprocess_image(file_path)
+        preds = best_model.predict(img_input)
+        class_idx = np.argmax(preds, axis=1)[0]
+        class_label = CLASS_LABELS[class_idx]
+        confidence = preds[0][class_idx]
+
+        # Update stats & log
+        stats[class_label] = stats.get(class_label, 0) + 1
+        log_prediction(class_label)
+
+        # Determine bin type
+        if class_label in RECYCLABLE:
+            bin_type = "Recyclable ♻️"
+        elif class_label in NON_RECYCLABLE:
+            bin_type = "Non-Recyclable 🗑️"
+        else:
+            bin_type = "Unknown ⚠️"
+
+        return render_template(
+            "result.html",
+            image=file.filename,
+            label=class_label,
+            confidence=f"{confidence*100:.2f}%",
+            bin_type=bin_type
+        )
+
+    return render_template("index.html")
+
+
+@app.route("/stats")
+def show_stats():
+    if stats:
+        categories = list(stats.keys())
+        counts = list(stats.values())
+        plt.figure(figsize=(6, 4))
+        plt.bar(categories, counts)
+        plt.xlabel("Category")
+        plt.ylabel("Count")
+        plt.title("Waste Classification Statistics")
+        plt.tight_layout()
+        plt.savefig("static/stats_chart.png")  # Safe with Agg backend
+        plt.close()
+    return render_template("report.html", stats=stats)
+
+
+@app.route("/download_pdf")
+def download_pdf():
+    pdf_path = generate_pdf_report()
+    return send_file(pdf_path, as_attachment=True)
+
+
+@app.route("/download_csv")
+def download_csv():
+    return send_file("waste_log.csv", as_attachment=True)
+
+# -----------------------------
+# Run Flask (useful locally)
+if __name__ == "__main__":
+    port = int(os.environ.get("PORT", 8080))
+    app.run(host="0.0.0.0", port=port, debug=True)

efficientnet_b0_best.keras

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b32878e752dc023742c2622159c2df364490e26db97d36b61c5d12d6fe281c1e
+size 29783632

mobilenet_trash.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83be04a781b14ae307e6144c8a6371301e897938c4680681eaacb02588a5877b
+size 9244779

requirement.txt

@@ -0,0 +1,10 @@
+Flask==2.3.3
+numpy
+opencv-python-headless
+tensorflow-cpu==2.12.0
+pandas
+matplotlib
+fpdf
+scikit-learn
+gunicorn
+huggingface-hub     # optional, only if you plan to download model at runtime from HF Hub