A service to detect if a given image of palm is ripe or not. A R&D initiated for Swopt

Dr-Swopt 328e3e8b22 Initialize: Add requirements.txt and MIT LICENSE		vor 3 Tagen
src	5f0bb61282 Feature: Add /analyze endpoint for full processing and refactor /detect for simple detection	vor 3 Tagen
.env	930fbc5999 Refactor: Implement Domain-Driven Design (DDD) with Vertex AI and MongoDB integration	vor 3 Tagen
.gitignore	ebcabef0d7 basic setup and trained data	vor 4 Tagen
LICENSE	328e3e8b22 Initialize: Add requirements.txt and MIT LICENSE	vor 3 Tagen
README.md	328e3e8b22 Initialize: Add requirements.txt and MIT LICENSE	vor 3 Tagen
Streamlit.md	ebcabef0d7 basic setup and trained data	vor 4 Tagen
best.pt	ebcabef0d7 basic setup and trained data	vor 4 Tagen
demo_app.py	ebcabef0d7 basic setup and trained data	vor 4 Tagen
main.py	14213a73a9 Refactor: Redirect root main.py to src.api.main and migrate confidence features	vor 3 Tagen
requirements.txt	328e3e8b22 Initialize: Add requirements.txt and MIT LICENSE	vor 3 Tagen
test_model.py	ebcabef0d7 basic setup and trained data	vor 4 Tagen
train_script.py	6dc23e13aa simple setup	vor 4 Tagen
yolov8n.pt	6dc23e13aa simple setup	vor 4 Tagen

🌴 Palm Oil Ripeness Agent (n8n + YOLOv8)

This project uses a custom-trained YOLOv8 model to detect the ripeness of Palm Oil Fresh Fruit Bunches (FFB). It features a local Python FastAPI server and a Streamlit Dashboard, both integrated into an agentic n8n workflow.

🚀 Project Overview

Vision Engine: YOLOv8 Nano (Custom-trained on MPOB-standard datasets).
Inference Server: FastAPI (Python) for n8n integration.
Demo Dashboard: Streamlit UI for drag-and-drop batch testing.
Intelligence: MongoDB Atlas Vector Search for similarity-based reasoning.

🛠 Prerequisites

Python 3.10+
n8n (Desktop or Self-hosted)
MongoDB Atlas Account (with Vector Search index enabled)

📦 Setup Instructions

1. Environment Setup

git clone <your-repo-url>
cd palm-oil-ai
python -m venv venv
# Windows: venv\Scripts\activate | Mac: source venv/bin/activate
pip install ultralytics fastapi uvicorn streamlit python-multipart pillow

2. Dataset & Training

Download the dataset from Roboflow *Or you can also find your own separate source of datasets. Make sure the file/folder format structure is consistent, especially the .yaml file.
Extract into /datasets.
Train the model:
```
python train_script.py
```
1. Copy the resulting best.pt from runs/detect/train/weights/ to the project root.
Running the API Server (DDD Structure)

The new architecture decouples the vision logic from the API entry point. You can run it either via the root wrapper or directly as a module:
```
# Option 1: Using the root wrapper
python main.py

# Option 2: Running as a module
python -m src.api.main
```

By default, the server runs on http://localhost:8000.

Available Endpoints

Endpoint	Method	Description
`/detect`	`POST`	Simple YOLO detection (Returns JSON)
`/analyze`	`POST`	Detection + Vertex Vectorization + MongoDB Archival
`/get_confidence`	`GET`	Returns the current model confidence setting
`/set_confidence`	`POST`	Updates the global model confidence setting

Running the Streamlit Dashboard

The Streamlit app still provides the user interface for manual testing.

# Run the Streamlit app
streamlit run demo_app.py
---

## 📂 Repository Structure

```text
├── datasets/           # Labeled images (Train/Valid/Test)
├── runs/               # YOLO training logs and output weight├── src
│   ├── api                 # FastAPI entry points
│   ├── application         # Use Cases (Orchestration)
│   ├── domain              # Business Logic & Entities
│   └── infrastructure      # External Services (MongoDB, VertexAI)
├── best.pt                 # YOLOv8 Trained weights
├── requirements.txt        # Python dependencies
├── .env                    # Configuration (Mongo, Vertex)
├── LICENSE                 # MIT License
└── README.md

README.md