agents.document_modifiers.tnt.branches

Branch decision functions for taxonomy generation workflow.

This module contains functions that determine the flow of the taxonomy generation process by making decisions about which branch of the workflow to follow next.

Example

Basic usage of branch decision functions:

state = TaxonomyGenerationState(...)
next_node = should_review(state)
# Returns either 'update_taxonomy' or 'review_taxonomy'

Functions

should_review(state)

Determines whether to continue refining the taxonomy or proceed to review.

Module Contents

agents.document_modifiers.tnt.branches.should_review(state)

Determines whether to continue refining the taxonomy or proceed to review.

This function compares the number of taxonomy revisions against the number of minibatches to decide if enough iterations have been performed to warrant a final review.

Parameters:

state (TaxonomyGenerationState) – The current state containing: - minibatches: List of document batch indices - clusters: List of taxonomy revisions

Returns:

Next node name in the workflow:

  • ’update_taxonomy’: Continue refining if more minibatches need processing

  • ’review_taxonomy’: Proceed to final review if all minibatches processed

Return type:

str

Example

>>> state = TaxonomyGenerationState(minibatches=[[1,2], [3,4]], clusters=[[...]])
>>> next_node = should_review(state)
>>> print(next_node)
'update_taxonomy'  # Since only one revision exists for two minibatches