annotate.classify_text

annotate.classify_text(
    chain,
    text,
    codebook,
    prompt_type='standard',
    use_examples=False,
    char_counts=None,
    timing_data=None,
    process_textbox=False,
    row_num=None,
    progress_bar=None,
    total_rows=None,
    process_span=False,
    retries=1,
    retry_strategy='identical',
    retry_chain=None,
    chat_session=None,
    reasoning_traces=None,
    chat_mode=None,
)

Annotate one text row across all sections in a codebook.

Parameters

Name	Type	Description	Default
chain		Runnable returned by setup_model.	required
text		Raw source text to annotate.	required
codebook		Parsed codebook dictionary.	required
prompt_type		Registered prompt wrapper name or callable wrapper.	'standard'
use_examples		Whether codebook examples should be included in prompts.	False
char_counts		Optional mutable counter dict for prompt/response characters.	None
timing_data		Optional mutable timing dict for inference statistics.	None
process_textbox		Whether textbox annotations should be generated.	False
row_num		Optional 1-based row number for progress logging.	None
progress_bar		Optional progress-bar helper updated after each annotation.	None
total_rows		Optional total row count for progress rendering.	None
chat_session		Optional retained chat history for this text or run.	None
reasoning_traces		Optional list that receives per-query reasoning records.	None
chat_mode		Normalized chat-history policy for trace metadata.	None

Returns

Name	Type	Description
		Tuple of (responses, char_counts, timing_data).