Text Sentiment Analysis Service

A high-performance sentiment analysis service that supports both Chinese and English text analysis, built with FastAPI and PaddlePaddle.

Features

• Real-time sentiment analysis for single texts
• Batch processing for multiple texts
• Automatic language detection (Chinese/English)
• High concurrency support
• Containerized deployment
• RESTful API interface

Prerequisites

• Python 3.9+
• Docker (for containerized deployment)
• PaddlePaddle models (Senta)

Installation

1. Clone the repository:

git clone <repository-url>
cd NLPService

2. Install dependencies:

pip install -r requirements.txt

3. Download models: Place the Senta models in the models/senta directory with the following structure:

models/senta/
├── cn_model/
│   ├── model.pdmodel
│   └── model.pdiparams
└── en_model/
    ├── model.pdmodel
    └── model.pdiparams

Running the Service

Local Development

uvicorn main:app --reload

Docker Deployment

docker build -t sentiment-service .
docker run -p 8000:8000 sentiment-service

API Usage

Single Text Analysis

curl -X POST "http://localhost:8000/analyze" \
     -H "Content-Type: application/json" \
     -d '{"text": "我很喜欢这个产品"}'

Response:

{
    "label": "positive",
    "confidence": 0.95
}

Batch Analysis

curl -X POST "http://localhost:8000/batch_analyze" \
     -H "Content-Type: application/json" \
     -d '{"texts": ["我很喜欢这个产品", "差评", "还不错"]}'

Response:

[
    {"label": "positive", "confidence": 0.95},
    {"label": "negative", "confidence": 0.87},
    {"label": "neutral", "confidence": 0.65}
]

API Documentation

Once the service is running, visit:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

Monitoring

The service logs are stored in logs/app.log. Key metrics are exposed for Prometheus monitoring.

License

[Your License]