Chunks

The output of a task is encapsulated in an OutputResponse, which contains a list of chunks. Each chunk is a collection of one or more segments.

The grouping of segments into chunks depends on the target_chunk_length parameter. Click here to learn more about how to configure it.

​

Output Response

The OutputResponse struct is the main output of a task. It is defined as:

#[derive(Serialize, Deserialize, Debug, Clone, ToSchema)]
pub struct OutputResponse {
    pub chunks: Vec<Chunk>,
    pub extracted_json: Option<ExtractedJson>,
}

• chunks: This is a vector of Chunk objects, representing the output of a task.
• extracted_json: This is an optional field that may contain additional extracted JSON data.

​

Chunk Model

The Chunk model represents a collection of segments. It consists of:

#[derive(Serialize, Deserialize, Debug, Clone, ToSchema)]
pub struct Chunk {
    pub segments: Vec<Segment>,
    pub chunk_length: i32,
}

• segments: This is a vector of Segment objects that make up the chunk.
• chunk_length: This is an integer representing the actual length of the chunk, i.e., the number of words in the chunk. This may differ from the target_chunk_length as segments are not split into smaller segments.

​

Segments

This will contain an array of segments that make up the chunk.

​

Chunk Length

This will contain the actual length of the chunk, i.e., the number of words in the chunk. This may be different from the target_chunk_length as the segments are not split into smaller segments.