teta

teta is a pluggable command-line toolbox written in Python. At its core it is a tiny dispatcher (teta) that discovers commands exported by other packages through the teta.cli_commands entry-point group. This makes it trivial to extend teta with your own functionality while keeping extra dependencies completely optional.

Everything in this table is optional — install only what you need.

teta uses Python's standard entry-point mechanism for plugin discovery. When teta <command> is run, the dispatcher queries the teta.cli_commands entry-point group at runtime and loads the matching plugin on demand. No central registry, no configuration file — just standard packaging metadata.

┌─────────────────────────────────────────────────────────────────┐
│                        teta (core)                              │
│                                                                 │
│  teta <cmd>  ──►  entry_points("teta.cli_commands")            │
│                         │                                       │
│            ┌────────────┼────────────────┐                      │
│            ▼            ▼                ▼                      │
│       teta-config   teta-mcp        teta-camera  …             │
│       (config cmd)  (mcp cmd)       (camera cmd)               │
└─────────────────────────────────────────────────────────────────┘

  teta.cli_commands  ──  CLI sub-commands  (teta <cmd>)
  teta.mcp_commands  ──  MCP tools exposed via teta-mcp

Each plugin package advertises itself with a single line in its pyproject.toml (see Writing your own plugin). No changes to the core teta package are ever needed.

Core only (minimal footprint):

pip install teta

With all official plugins:

pip install "teta[all]"

Individual plugins:

pip install teta-config   # configuration management
pip install teta-camera   # camera streaming
pip install teta-mcp      # Model Context Protocol server
pip install teta-diag     # diagnostics
pip install teta-generate # code generation
pip install teta-app      # Flask web UI

For development (editable install with all plugins):

git clone https://github.com/rios/teta
cd teta
pip install -e ".[all]"

# Show the installed version
teta version

# List all installed plugins
teta plugins

# Show the active configuration file (requires teta-config)
teta config show

# Print diagnostics and verify environment (requires teta-diag)
teta diag

# Start the MCP server (requires teta-mcp)
teta mcp

# Stream frames from a camera (requires teta-camera)
teta camera capture

# Start the Flask web UI (requires teta-app)
teta app run

If a command is not installed, teta will print a clear error and exit. Run teta plugins to see which plugins are currently available.

Usage: teta <command> [options]

Available commands:
  version
  plugins
  config    (if teta-config is installed)
  mcp       (if teta-mcp is installed)
  ...

Prints the version of the installed teta core package.

Lists every registered entry-point in both teta.cli_commands and teta.mcp_commands, along with the package name and version that provides it.

Example output:

[teta.cli_commands]
  config               – teta-config 0.2.1
  mcp                  – teta-mcp 0.2.0
  version              – teta 0.5.0
  plugins              – teta 0.5.0

[teta.mcp_commands]
  (none)

teta-config stores user preferences in JSON. By default the file lives at ~/.teta.json; you can override the location with the TETA_CONFIG environment variable.

{
  "log_level": "INFO",
  "log_format": "%(levelname)s: %(message)s"
}

Programmatic usage:

from teta.config import get_config

cfg = get_config()          # loads and merges defaults
cfg.log_level = "DEBUG"     # mutate values
cfg.save(cfg.get_config_path())

teta-mcp runs a JSON-RPC 2.0 server that exposes teta plugins as MCP tools to AI assistants and agent frameworks.

Start the server:

teta mcp

MCP tools are discovered via the teta.mcp_commands entry-point group. Any package can register additional tools without modifying the core:

[project.entry-points."teta.mcp_commands"]
my_tool = "my_pkg.mcp:my_tool_handler"

Each handler receives a JSONRPCRequest and returns a plain string. See teta-mcp for a full walkthrough.

A teta plugin is a standard Python distribution that advertises one or more commands via an entry point. No changes to teta itself are required.

my-teta-plugin/
├── pyproject.toml
└── my_plugin/
    ├── __init__.py
    └── cli.py

# my_plugin/cli.py
import click

@click.command()
@click.option("--name", default="world", help="Who to greet")
def main(name):
    """A friendly greeting plugin for teta."""
    click.echo(f"Hello, {name}!")

# pyproject.toml
[project.entry-points."teta.cli_commands"]
hello = "my_plugin.cli:main"

pip install -e .
teta hello
# Hello, world!
teta hello --name Alice
# Hello, Alice!

The new command appears automatically in teta — no configuration needed.

git clone https://github.com/rios/teta
cd teta

# Install in editable mode with all plugins
pip install -e ".[all]"

# Run the test suite
make test

# Run with coverage
make coverage

# Format code
make format

# Lint
make lint

python -m unittest discover

make build
# Output goes to dist/

Apache-2.0