Classes
PropositionList
A Pydantic model to ensure the LLM returns a list of propositions.Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
propositions | List[str] | - | A list of simple, self-contained factual statements extracted from the source text |
Topic
Represents a single, emergent topic with its assigned propositions.Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
topic_id | int | - | A unique integer ID for this topic |
propositions | List[str] | - | A list of the propositions that belong to this topic |
TopicAssignmentList
The structure for the LLM’s response during the batch clustering stage.Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
topics | List[Topic] | - | A list of all identified topics, each containing its assigned propositions |
RefinedTopic
The final, refined metadata for a completed chunk.Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
title | str | - | A concise, human-readable title for the topic cluster (e.g., ‘Candidate Skills and Experience’) |
summary | str | - | A one-sentence summary that encapsulates the core information of all propositions in the cluster |