Metadata-Version: 2.4
Name: tirx-harness
Version: 0.1.2rc1
Summary: TIRx kernel development, inspection, and simulation tools
Project-URL: Homepage, https://tirxharness.mlc.ai
Project-URL: Repository, https://github.com/mlc-ai/TIRx-harness
Requires-Python: <3.14,>=3.12
Description-Content-Type: text/markdown
Requires-Dist: apache-tvm==0.27.0
Requires-Dist: apache-tvm-ffi==0.1.14.post0
Requires-Dist: numpy>=1.24
Requires-Dist: threadpoolctl<4,>=3.6
Provides-Extra: server
Requires-Dist: kcoral>=0.1.0rc0; extra == "server"
Requires-Dist: fastapi>=0.110; extra == "server"
Requires-Dist: grpcio>=1.66.2; extra == "server"
Requires-Dist: protobuf>=5.27.2; extra == "server"
Requires-Dist: uvicorn>=0.29; extra == "server"
Provides-Extra: test
Requires-Dist: pytest>=8; extra == "test"
Requires-Dist: pytest-xdist>=3; extra == "test"
Requires-Dist: setuptools>=64; extra == "test"
Requires-Dist: pip>=25.1; extra == "test"
Requires-Dist: ml-dtypes; extra == "test"

# TIRx Harness

Use TIRx kernel-development skills in your own agent, or start an
[optimization run](docs/optimization-runs.md) with a prepared task, isolated workspace,
and benchmark.

The [documentation](docs/index.md) covers the components, agent loop, and
guides for adding workloads and contributing kernels and fixes. See
[build and preview instructions](docs/development/build-the-docs.md) to run the site locally.

Start with [Installation](docs/installation.md) for packages and skills, then
follow the [Quick Start](docs/quick-start.md) for an example agent goal.

| Use case | Start here | What is prepared |
| --- | --- | --- |
| Your own agent and project | [Quick Start](docs/quick-start.md) | Installed packages and three copied skills |
| An optimization run | [Optimization Runs](docs/optimization-runs.md) | A task workspace with its own venv and five skills |

See [prerequisites](docs/installation.md#before-you-start) before installing.

## Repository map

| Directory | Owns |
| --- | --- |
| [skills/](skills/README.md) | Three distributable kernel-development skills and the pinned sources of two external skills |
| [tirx_harness/](tirx_harness/README.md) | TIRx Tools, their Python/Rust sources, tests, and design docs |
| [evolution/](evolution/README.md) | Task declarations, run preparation, agent prompts, benchmark implementations, and remote tools |
| [pyproject.toml](pyproject.toml), [uv.lock](uv.lock) | Package metadata, dependency groups, and locked versions |
| [thirdparty/](thirdparty/README.md) | Pinned external source dependencies |
| [docs/](docs/index.md) | Documentation site, Quick Start, and user guides |

Run preparation and execution code lives in `evolution/`. Generated workspaces
and results live under `kda_flow_runs/`; each workspace stores candidate kernels in
`candidates/<workload>/scratch/` and `frontier/`.

## Installation

For your own Python environment, install the harness with
`python -m pip install --group harness .`.
Copy the three local skill directories to your agent's skills directory and run
`tirx-wiki/scripts/fetch_references.py` from the copied wiki directory.
The [Installation guide](docs/installation.md) gives the complete commands.

The reference fetcher also downloads `tirx-kernels` into the wiki's
`references/repos/tirx-kernels/`, including its root README and source docs.
The installed `tirx-kernels` package is used for execution.

For an optimization run, follow [Optimization Runs](docs/optimization-runs.md) to
install the setup dependencies and select a task and GPU/server. Setup creates
the run's worktree and venv and installs its packages and skills. Review the
generated prompt, then launch your agent with the printed command.
