artasupport 0.1.0

Python SDK for ArtaSupport TSE copilot and batch APIs.

artasupport Python SDK

artasupport is a Python SDK for ArtaSupport TSE APIs (copilot and batch).

Requirements

• Python 3.10+
• An API key (arta_live_...)
• Backend base URL (http://localhost:8000 or your deployed URL)

Install

pip install artasupport

Local development install:

pip install -e .

Environment setup

export ARTASUPPORT_API_KEY="arta_live_...<secret>"
export ARTASUPPORT_BASE_URL="http://localhost:8000"

You can also pass api_key= and base_url= directly in code.

How it works

Each SDK call:

1. Builds auth headers (x-api-key + Bearer)
2. Sends request to backend (/api/tse, /v1/models, /v1/sessions, /api/ingest)
3. Parses typed response (text, session_id, usage, next_steps, etc.)
4. Raises typed exceptions for API/network/config errors

Usage mode 1: Quick one-shot calls

Use top-level helper functions for simple scripts.

from artasupport import tse_copilot, tse_batch

resp = tse_copilot(user_input="internet is slow")
print(resp.text)

resp2 = tse_batch(text="summarize tunnel issue")
print(resp2.text)

Usage mode 2: Sync conversation with context

Use ArtaSupportMessage when you want multi-turn chat. Conversation context is tied to session_id.

from artasupport import ArtaSupportMessage

with ArtaSupportMessage() as client:
    sid = client.create_session().session_id

    r1 = client.tse_copilot(user_input="internet is slow", session_id=sid)
    r2 = client.tse_copilot(user_input="only one location affected", session_id=sid)

    print(r1.text)
    print(r2.text)

    # optional cleanup
    client.delete_session(sid)

Usage mode 3: Async conversation

Use AsyncArtaSupportMessage in async apps.

import asyncio
from artasupport import AsyncArtaSupportMessage

async def main():
    async with AsyncArtaSupportMessage() as client:
        sid = (await client.create_session()).session_id
        r = await client.tse_batch(user_input="summarize issue", session_id=sid)
        print(r.text)

asyncio.run(main())

AsyncArtaSupportMessage is non-blocking async I/O, not token streaming.

Usage mode 4: File ingestion

Upload PDF or Excel files for runbook extraction and cases ingestion.

from artasupport import tse_ingest

# Quick one-shot upload
response = tse_ingest("runbook.pdf")
print(response.status)  # "completed" or "processing"

# With explicit client
from artasupport import ArtaSupportMessage

with ArtaSupportMessage() as client:
    response = client.tse_ingest("cases.xlsx")
    print(f"Ingestion {response.status}")

Async version:

import asyncio
from artasupport import AsyncArtaSupportMessage

async def upload_file():
    async with AsyncArtaSupportMessage() as client:
        response = await client.tse_ingest("runbook.pdf")
        print(f"Status: {response.status}")

asyncio.run(upload_file())

Supported file types: .pdf, .xlsx, .xls Folders are not supported; tse_ingest(...) expects a single file path.

Input styles

For tse_copilot(...) and tse_batch(...), provide exactly one of:

• user_input="..."
• text="..."
• message="..."
• messages=[{"role": "user", "content": "..."}]

Error handling

from artasupport import (
    ArtaSupportMessage,
    APIError,
    AuthenticationError,
    RateLimitError,
)

try:
    with ArtaSupportMessage() as client:
        resp = client.tse_copilot(user_input="help", session_id="sess_...")
except AuthenticationError as exc:
    print("Invalid/revoked API key:", exc)
except RateLimitError as exc:
    print("Rate limited:", exc)
except APIError as exc:
    print("API failure:", exc.status_code, exc.request_id, exc.message)

API surface

• Top-level helpers:
  • tse_copilot(...)
  • tse_batch(...)
  • tse_ingest(file_path)
• Sync client:
  • ArtaSupportMessage
  • list_models(), create_session(), delete_session()
  • tse_copilot(...), tse_batch(...), tse_ingest(file_path)
• Async client:
  • AsyncArtaSupportMessage
  • await list_models(), await create_session(), await delete_session()
  • await tse_copilot(...), await tse_batch(...), await tse_ingest(file_path)