ViaRAG SDK

Minimal Python SDK for interacting with the ViaRAG API. Designed for developers building RAG pipelines, chatbots, and AI-native workflows.

---

📦 Installation

pip install viarag

---

🚀 Quickstart

from viarag import ViaRAGClient

client = ViaRAGClient(api_key="your_api_key")
print(client.health_check())

---

🔧 Class: ViaRAGClient

ViaRAGClient(api_key: str, timeout: int = 30)

Creates a new client.

---

📡 Endpoints

✅ 1. health_check()

Returns the API's current health status.

client.health_check()

Returns:

{"status": "ok"}

---

🤖 2. simple_query(prompt: str, top_k: int = 5)

Runs a retrieval-augmented generation (RAG) query.

client.simple_query("What is ViaRAG?")

Returns:

{
  "response": "ViaRAG is an API for retrieval-augmented generation...",
  "contexts": [...],
  "prompt": "..."
}

---

💬 3. direct_query(prompt: str)

Runs a prompt directly through the LLM, no retrieval.

client.direct_query("Tell me a joke.")

---

🔍 4. match_context(prompt: str, top_k: int = 5)

Returns top-k context chunks that match your prompt (no generation).

client.match_context("What is ViaVeri?")

Returns:

[
  {"content": "...", "score": 0.92},
  {"content": "...", "score": 0.87}
]

---

📄 5. upload_document(file_path: str, metadata: dict = None, chunking_config: dict = None)

Uploads a document and indexes it.

client.upload_document(
    file_path="my_notes.pdf",
    metadata={"source": "user-upload"},
    chunking_config={"chunk_size": 1000, "chunk_overlap": 200}
)

Supports: .pdf, .docx, .txt

---

✂️ 6. chunk_text(text: str, chunk_size: int = 1000, chunk_overlap: int = 200, splitter: str = "recursive")

Chunks raw text without uploading a file.

chunks = client.chunk_text("A long string of text...")

Returns:

["Chunk 1", "Chunk 2", ...]

---

📚 7. list_documents()

Lists all documents you’ve uploaded.

client.list_documents()

Returns:

[
  {"doc_id": "abc123", "filename": "my_notes.pdf"},
  ...
]

---

❌ 8. delete_document_by_id(doc_id: str)

Deletes all chunks associated with a document.

client.delete_document_by_id("abc123")

---

🔐 Authentication

All API calls (except health_check) require a Bearer token:

client = ViaRAGClient(api_key="sk-...")

---

🧪 Development

Clone the repo and install locally:

git clone https://github.com/YOUR_USERNAME/viarag-sdk.git
cd viarag-sdk
pip install -e .

---

📄 License

GNU General Purpose License. See LICENSE file.

---

👋 About

Built by ViaVeri Technologies to empower developers with simple, powerful RAG tools.