oaas-tool 0.1.0

A lightweight observability wrapper for LangSmith, Langfuse, Sentry, Phoenix, and OpenTelemetry.

oaas-tool

oaas-tool is a small Python observability wrapper that lets you plug one or more providers into any project with a simple config.

Supported tools:

• langsmith
• langfuse
• sentry
• phoenix
• otel

Install

After this repository is on GitHub, anyone can install it with:

pip install "git+https://github.com/MeghanaTathireddy/oaas-tool.git"

If you later publish it to PyPI, the install becomes:

pip install oaas-tool

Quick Start

from oaas import init, Event

obs = init(
    langsmith={
        "api_key": "your-langsmith-api-key",
        "project": "my-app",
    }
)

obs.push(Event("app_started", {"env": "dev"}))

You can also use environment variables instead of passing keys directly.

from oaas import init, Event

obs = init(
    langsmith={},
    sentry={},
)

obs.push(Event("request_received", {"route": "/chat"}))

Environment Variables

LangSmith

• LANGSMITH_API_KEY
• LANGSMITH_PROJECT

Langfuse

• LANGFUSE_PUBLIC_KEY
• LANGFUSE_SECRET_KEY
• LANGFUSE_HOST

Sentry

• SENTRY_DSN

Tool Selection

Users only enable the tools they want by passing those tool names into init(...).

from oaas import init

obs = init(
    sentry={"dsn": "your-dsn"},
    phoenix={},
)

If a tool needs credentials, the library expects the user's own keys through config or environment variables.

API

from oaas import init, Event, available_tools, node_start, node_end, error_event

• init(config=None, **tool_configs) creates and starts the configured providers.
• available_tools() returns the supported tool names.
• Event(...) creates a custom event.
• node_start(...), node_end(...), and error_event(...) are helper events.

Example

from oaas import init, Event

obs = init(
    langfuse={
        "public_key": "your-public-key",
        "secret_key": "your-secret-key",
    }
)

obs.push(Event("generation_completed", {"tokens": 128}))

Publish To GitHub

1. Create a new GitHub repository.
2. Add that GitHub repo as the remote for this project.
3. Push your code.
4. Create a release when you want to publish to PyPI.

Publish To PyPI Later

Once you are ready, users will be able to run pip install oaas-tool.

GitHub Actions

This repository includes:

• CI on push and pull request
• package build verification
• a PyPI publish workflow triggered by GitHub releases

For PyPI publishing, configure Trusted Publishing for this GitHub repository on PyPI.