vectornest-sdk 0.1.0

Python SDK for the Vectornest / RAG Studio API

vectornest

Official Python SDK for the Vectornest / RAG Studio API.

Installation

pip install vectornest-sdk

Quickstart

from vectornest import Vectornest

client = Vectornest(api_key="vn_your_key_here")

# Create and activate a collection
client.collections.create("research-docs")
client.collections.activate("research-docs")

# Upload documents and wait for ingestion
task = client.documents.upload(["report.pdf", "data.csv"])
task.wait(timeout=120)  # blocks; raises TaskFailedError on failure

# Query
resp = client.chat.query("summarize the key findings", session_id="session-1")
print(resp.answer)
for source in resp.sources:
    print(f"  [{source.score:.2f}] {source.filename}")

Async Usage

import asyncio
from vectornest import AsyncVectornest

async def main():
    async with AsyncVectornest(api_key="vn_your_key_here") as client:
        await client.collections.activate("research-docs")
        task = await client.documents.upload(["report.pdf"])
        await task.wait_async(timeout=120)
        resp = await client.chat.query("what are the conclusions?")
        print(resp.answer)

asyncio.run(main())

Error Handling

from vectornest.exceptions import BudgetExceededError, AuthenticationError, TaskFailedError

try:
    task = client.documents.upload(["big_file.pdf"])
    task.wait()
except AuthenticationError:
    print("Invalid API key")
except BudgetExceededError:
    print("Usage limit reached — upgrade your plan")
except TaskFailedError as e:
    print(f"Ingestion failed: {e.message}")

API Reference

Vectornest(api_key, base_url, timeout)

Namespace	Methods
client.collections	create(), list(), active(), activate(), delete()
client.documents	upload(), list(), delete()
client.chat	query()

UploadTask

Method	Description
.wait(timeout=300, poll_interval=3)	Block until ingestion completes
.wait_async(timeout=300, poll_interval=3)	Async variant

QueryResponse

Field	Type	Description
answer	str	Generated LLM answer
sources	list[SourceChunk]	Retrieved context chunks
llm_metrics	LLMMetrics	Token counts, latency
fallback_used	bool	Whether fallback model was used

License

MIT