synapse-plugin-executionlisted

# Plugin Execution Synapse SDK provides multiple ways to execute plugin actions programmatically. ## run_plugin() Function The simplest way to run a plugin action: ```python from synapse_sdk.plugins.runner import run_plugin, ExecutionMode # Local execution (in-process, good for dev) result = run_plugin( plugin_code='/path/to/plugin', action='train', params={'epochs': 100}, mode=ExecutionMode.LOCAL, ) # Ray Actor execution (fast startup) result = run_plugin( plugin_code='/path/to/plugin', action='train', params={'epochs': 100}, mode=ExecutionMode.TASK, ) # Ray Job execution (heavy workloads) job_id = run_plugin( plugin_code='/path/to/plugin', action='train', params={'epochs': 100}, mode=ExecutionMode.JOB, ) ``` ## Execution Modes | Mode | Class | Use Case | Returns | |------|-------|----------|---------| | `LOCAL` | `LocalExecutor` | Development, testing | Result dict | | `TASK` | `RayActorExecutor` | Fast startup, medium work | Result dict | | `JOB` | `RayJobExecutor` | Heavy workloads, isolation | Job ID string | ## Plugin Discovery Discover and inspect plugins before execution: ```python from synapse_sdk.plugins.discovery import PluginDiscovery # From filesystem path discovery = PluginDiscovery.from_path('/path/to/plugin') # From Python module import my_plugin discovery = PluginDiscovery.from_module(my_plugin) # List available actions actions = discovery.list_actions() # ['train', 'inference'] # Get action cla