tokamunch-pyuda-datasource

Python data source plugin for tokamunch. Fetches data from a UDA server via pyuda.

Available on PyPI.

Installation

pip install tokamunch
pip install tokamunch-pyuda-datasource

Configuration

Register the plugin in your munchi.toml and supply the connection args. The table key is the data-source name used in your libtokamap mapping files.

[data_sources.UDA]
plugin = "tokamunch_pyuda_datasource"
enabled = true

args.host = "localhost"
args.port = 56565
args.plugin_name = "UDA"

Any other key in args becomes a default for the per-call UDA query args, which an individual mapping can override.

Mapping example

Per-call arguments come from the mapping's ARGS block, plus the runtime attributes tokamunch passes to Mapper.map (notably shot):

"ip/data": {
    "MAP_TYPE": "DATA_SOURCE",
    "DATA_SOURCE": "UDA",
    "ARGS": {
        "signal": "AMC_PLASMA CURRENT"
    }
}

Any additional key/value pairs are passed through as UDA query parameters.

Args: config vs per-call

• Config args (host, port, plugin_name, function) are set once when the plugin is constructed from munchi.toml.
• Per-call args (signal, source, and any extra UDA query parameters) are supplied per mapping call by the tokamunch mapping engine.

host and port are forwarded into every UDA query from the plugin config, so mapping files no longer need to repeat them in their ARGS blocks. source falls back to the runtime shot when the mapping does not set it. Both remain overridable per mapping.

Notes on thread safety

This plugin is not thread-safe. pyuda.Client stores server and port as class-level attributes, so concurrent threads sharing a process will interfere with each other. The plugin re-applies its own server/port under a lock before each request — which keeps several data sources (e.g. UDA and GEOMETRY on different ports) correct within one process — but pyuda itself offers no thread-safety guarantee. Use run.concurrency.mode = "process" if you need parallel requests.