Metadata-Version: 2.2
Name: jsongeekai
Version: 0.2.0a1
Summary: High-Performance JSON Parser with AI-Driven Optimization
Author-email: JsonGeekAI Team <team@jsongeekai.org>
License: MIT
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy>=2.1.3
Requires-Dist: numba>=0.61.0
Requires-Dist: psutil>=7.0.0
Requires-Dist: ijson>=3.3.0

# JsonGeekAI

JsonGeekAI是一个智能JSON解析和优化框架，现已集成DeepSeek AI能力，提供更智能的JSON处理和优化建议。

## 🌟 特性

- 🧠 **AI驱动的优化**：集成DeepSeek-Coder模型，提供智能的JSON优化建议
- ⚡ **SIMD加速**：自动识别和建议SIMD优化机会
- 🎯 **智能缓存**：优化建议的智能缓存，提高响应速度
- 🔍 **深度分析**：深入分析JSON结构，识别优化机会
- 🛠️ **易于集成**：简单的API设计，易于集成到现有项目

## 📦 安装

```bash
pip install jsongeekai
```

## 🚀 快速开始

```python
from jsongeekai.ai.optimizers import SimdAdvisor
from jsongeekai.ai.model_manager import ModelLoader

# 初始化模型和优化器
model = ModelLoader()
advisor = SimdAdvisor(model_manager=model)

# 分析JSON并获取优化建议
json_data = {
    "arrays": {
        "numbers": list(range(1000)),
        "floats": [float(x) for x in range(500)]
    }
}

suggestions = advisor.get_optimization_suggestions(json_data)

# 打印优化建议
for suggestion in suggestions:
    print(f"目标模式: {suggestion.target_pattern}")
    print(f"建议指令: {suggestion.simd_instruction}")
    print(f"预期加速: {suggestion.estimated_speedup}x")
    print(f"示例代码:\n{suggestion.code_example}")
    print("-" * 50)
```

## 🏗️ 项目结构

```
jsongeekai/
├── ai/
│   ├── model_manager/
│   │   ├── __init__.py
│   │   ├── loader.py        # 模型加载和管理
│   │   └── tokenizer.py     # 分词和预处理
│   ├── optimizers/
│   │   ├── __init__.py
│   │   ├── json_analyzer.py # JSON结构分析
│   │   └── simd_advisor.py  # SIMD优化建议
│   └── inference/
│       ├── __init__.py
│       ├── engine.py        # 推理引擎
│       └── cache.py         # 优化缓存
└── tests/
    └── ai/
        ├── test_deepseek_integration.py
        └── test_performance.py
```

## 📊 性能基准

| 数据大小 | 推理时间 | GPU内存使用 |
|---------|---------|------------|
| 小型    | <2s     | ~100MB     |
| 中型    | <5s     | ~200MB     |
| 大型    | <10s    | ~500MB     |

## 🔧 配置

### 模型配置

```python
from jsongeekai.ai.model_manager import ModelLoader

# 配置模型加载器
model = ModelLoader(
    model_name="deepseek-ai/deepseek-coder-1.3b-base",
    device="cuda",  # 或 "cpu"
    half_precision=True  # 使用FP16以减少内存使用
)
```

### 缓存配置

```python
from jsongeekai.ai.inference import OptimizationCache

# 配置缓存
cache = OptimizationCache(
    cache_dir="~/.jsongeekai/cache",
    max_size=1000,  # 最大缓存条目数
    ttl=86400      # 缓存有效期（秒）
)
```

## 📝 API文档

### SimdAdvisor

```python
class SimdAdvisor:
    """SIMD优化顾问"""
    
    def get_optimization_suggestions(self, json_data: Any) -> List[SimdOptimizationSuggestion]:
        """获取SIMD优化建议"""
        
    def analyze_json_structure(self, json_data: Any) -> Dict[str, Any]:
        """分析JSON结构"""
```

### InferenceEngine

```python
class InferenceEngine:
    """AI推理引擎"""
    
    def infer(self, context: Dict[str, Any], max_suggestions: int = 5) -> InferenceResult:
        """执行推理"""
```

## 🤝 贡献

欢迎提交Issue和Pull Request！

## 📄 许可证

MIT License

# JsonGeekAI

[![PyPI version](https://img.shields.io/pypi/v/jsongeekai.svg)](https://pypi.org/project/jsongeekai/)
[![Python Versions](https://img.shields.io/pypi/pyversions/jsongeekai.svg)](https://pypi.org/project/jsongeekai/)
[![Downloads](https://img.shields.io/pypi/dm/jsongeekai.svg)](https://pypi.org/project/jsongeekai/)
[![Tests](https://img.shields.io/github/workflow/status/zhanghongping/jsongeekai/Tests)](https://github.com/zhanghongping/jsongeekai/actions)
[![Coverage Status](https://coveralls.io/repos/github/zhanghongping/jsongeekai/badge.svg?branch=main)](https://coveralls.io/github/zhanghongping/jsongeekai?branch=main)
[![License](https://img.shields.io/pypi/l/jsongeekai.svg)](https://github.com/zhanghongping/jsongeekai/blob/main/LICENSE)

A high-performance JSON parser with AI-driven optimizations and multi-format support.

## Quick Start

```python
import jsongeekai

# 基本用法 - 从字符串解析
json_str = '{"name": "JsonGeekAI", "type": "parser"}'
data = jsongeekai.parse(json_str)
print(data)  # {'name': 'JsonGeekAI', 'type': 'parser'}

# 从文件解析
with open('large_file.json', 'rb') as f:
    data = jsongeekai.parse_file(f)

# 流式处理大文件
for item in jsongeekai.parse_stream('large_file.json'):
    print(item)

# 自动检测并处理不同格式
data = jsongeekai.parse_auto('data.jsonl')  # 自动检测JSONL格式

# 启用SIMD优化
jsongeekai.enable_simd()
data = jsongeekai.parse(json_str)  # 现在使用SIMD加速

# 内存效率配置
jsongeekai.set_memory_limit('2GB')  # 限制内存使用
```

## Features

- **SIMD Optimizations**: Utilizes SIMD instructions for faster parsing when available
- **Multi-format Support**: Handles JSON, JSON5, YAML, MessagePack, and JSONL formats
- **Memory Efficient**: Smart memory management with configurable limits
- **Format Auto-detection**: Automatically detects and handles different formats
- **Rich Error Handling**: Detailed error messages with context and documentation
- **Compression Support**: Built-in compression for JSONL format
- **Extensible**: Easy to add new formats and optimizations
- **Streaming Support**: Streaming support for large files

## Installation

### Using pip

```bash
pip install jsongeekai
```

### Using Docker

```bash
# Pull the latest image
docker pull zhanghongping/jsongeekai:latest

# Run tests in container
docker run zhanghongping/jsongeekai:latest

# Run Python with JsonGeekAI
docker run -it zhanghongping/jsongeekai:latest python
```

### Using Conda

```bash
# Install from conda-forge
conda install -c conda-forge jsongeekai

# Or create a new environment
conda create -n jsongeekai-env -c conda-forge jsongeekai
conda activate jsongeekai-env
```

## Version Compatibility

JsonGeekAI requires Python 3.8 or later and is tested on:
- Python 3.8
- Python 3.9
- Python 3.10
- Python 3.11
- Python 3.12

## Performance

JsonGeekAI is designed for performance:

- SIMD acceleration when available
- Efficient memory usage
- Format-specific optimizations
- Streaming support for large files

## Documentation

Full documentation is available at [jsongeekai.readthedocs.io](https://jsongeekai.readthedocs.io/)

## Contributing

We welcome contributions! Please see our [Contributing Guide](CONTRIBUTING.md) for details.

## License

JsonGeekAI is released under the MIT License. See the [LICENSE](LICENSE) file for details.
