Skip to main content

mellea.plugins.hooks.generation

Generation pipeline hook payloads.

Classes​

CLASS GenerationPreCallPayload ​

Payload for generation_pre_call — before LLM backend call.

Attributes:

  • action: The Component or CBlock about to be sent to the backend.
  • context: The Context being used for this generation call.
  • model_options: Dict of model options (writable — plugins may adjust temperature, etc.).
  • format: Optional BaseModel subclass for constrained decoding (writable).
  • tool_calls: Whether tool calls are enabled for this generation (writable).
  • generation_id: Mellea-side hook correlation ID, distinct from the provider-assigned GenerationMetadata.response_id. None when the firing site does not generate one.

CLASS GenerationPostCallPayload ​

Payload for generation_post_call — fires once the model output is fully computed.

For lazy ModelOutputThunk objects this hook fires inside ModelOutputThunk.astream after post_process completes, so model_output.value is guaranteed to be available. For already-computed thunks (e.g. cached responses) it fires before generate_from_context returns.

Attributes:

  • prompt: The formatted prompt sent to the backend (str or list of message dicts).
  • model_output: The fully-computed ModelOutputThunk.
  • latency_ms: Elapsed milliseconds from the generate_from_context call to when the value was fully materialized.
  • generation_id: Mellea-side hook correlation ID matching the corresponding pre_call payload, distinct from the provider-assigned GenerationMetadata.response_id. None when the firing site did not generate one.

CLASS GenerationErrorPayload ​

Payload for generation_error — fires when the LLM backend raises an exception.

This hook fires inside ModelOutputThunk.astream just before the exception is re-raised, giving plugins a chance to observe (but not suppress) the error.

Attributes:

  • exception: The exception raised by the backend.
  • model_output: The ModelOutputThunk at the time of the error. model and provider are set when the backend set them early (before the async task); otherwise they are None.
  • generation_id: Mellea-side hook correlation ID matching the corresponding pre_call payload, distinct from the provider-assigned GenerationMetadata.response_id. None when the firing site did not generate one.

CLASS GenerationBatchPreCallPayload ​

Payload for generation_batch_pre_call — fires once before a batch generation request.

Carries the action sequence being sent in the batch alongside batch-level fields (model, provider, num_actions) describing the single API call.

Attributes:

  • actions: The action sequence being sent in the batch.
  • generation_id: Correlation identifier set by the firing backend; matches the corresponding generation_batch_post_call / generation_batch_error payloads for the same call.
  • format: Optional structured-output format applied to the batch.
  • tool_calls: Whether tool calling is enabled (typically False for raw).
  • num_actions: Convenience copy of len(actions).
  • model: Model identifier the backend is calling.
  • provider: Provider name (e.g. "openai", "ollama").

CLASS GenerationBatchPostCallPayload ​

Payload for generation_batch_post_call — fires once after a batch generation succeeds.

Carries the list of ModelOutputThunk instances produced by the batch alongside batch-level fields (usage, model, provider, latency_ms) describing the single API call.

Attributes:

  • generation_id: Correlation identifier from the matching pre_call.
  • model_outputs: The list of ModelOutputThunk instances built from the API response, in batch order.
  • usage: Aggregate token-usage dict (OpenAI-shape) for the whole batch.
  • model: Model identifier from the call.
  • provider: Provider name.
  • latency_ms: Wall-clock duration of the API call in milliseconds.

CLASS GenerationBatchErrorPayload ​

Payload for generation_batch_error — fires once when a batch generation request fails.

Carries the exception alongside batch-level fields (model, provider, latency_ms) describing the failed API call. No ModelOutputThunk instances are present.

Attributes:

  • generation_id: Correlation identifier from the matching pre_call.
  • exception: The exception raised by the backend.
  • model: Model identifier from the call.
  • provider: Provider name.
  • latency_ms: Wall-clock time-until-error in milliseconds.