Metadata-Version: 2.4
Name: nlqcat
Version: 0.1.0
Summary: Natural Language Question Answering Toolkit (Hybrid NLP + GenAI)
Home-page: https://github.com/Anirbansarkars/nlqcat
Author: Anirban-QuantumCAT
Author-email: Anirban Sarkar <sanirbansarkar@gmail.com>
Project-URL: Homepage, https://github.com/AnirbansarkarS/NLQcat
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: spacy>=3.0.0
Requires-Dist: chromadb>=0.4.0
Requires-Dist: sentence-transformers>=2.2.0
Requires-Dist: numpy
Provides-Extra: openai
Requires-Dist: openai>=1.0.0; extra == "openai"
Dynamic: author
Dynamic: home-page
Dynamic: license-file
Dynamic: requires-python

# NLQcat: Natural Language Question Answering Toolkit

**NLQcat** is a production-ready, hybrid NLP + GenAI library that unifies classic linguistic analysis (spaCy) with modern semantic search (Vector Databases) and Large Language Models (LLMs) into a single, easy-to-use pipeline.

## 📦 Installation

```bash
pip install nlqcat
python -m spacy download en_core_web_sm
```

## 🚀 Quick Start (3-Line Example)

```python
from nlqcat.core.pipeline import Pipeline

pipe = Pipeline(vector_store_type="chroma")
pipe.add_documents(["NLQcat combines linguistic NLP with semantic RAG."])
print(pipe.query("What does NLQcat do?")['answer'])
```

## ✨ Features

- **Hybrid Analysis**: Seamlessly mixes spaCy's linguistic features (NER, POS tagging) with semantic embeddings.
- **Unified RAG Pipeline**: Built-in Retrieve-Augmented Generation flow (Query -> Retrieve -> Answer).
- **Vector Store Support**: Integrated ChromaDB client for local vector storage.
- **Modular Design**: Plug-and-play components for Loaders, Preprocessors, and Agents.
- **LLM Ready**: Simple `OpenAIAgent` integration for generative answers.

## 🗺️ Roadmap

- [ ] Support for FAISS and Pinecone vector stores.
- [ ] Integration with HuggingFace and Anthropic LLMs.
- [ ] Advanced RAG techniques (Re-ranking, HyDE).
- [ ] Async support for high-throughput pipelines.
