Write Document Files¶
Write a standard document array to Catalog files, retaining content, metadata, and associated parsing artifacts.
Use it to save parsed or chunked results for later flows. To select a destination and handle other content such as text or JSON, use Save to MOI Volume.
Configure content¶
documents must be nonempty. Optionally set output_file_name; otherwise, the node uses a document filename containing a time identifier.
Bind the full parsed or chunked array to retain content and source information.
File format and count¶
Ordinary documents can be written as JSONL. References to parsing artifacts such as Markdown or layout data, or an explicitly specified ZIP filename, produce ZIP output.
Documents from multiple sources are grouped by source, so one execution can return several files.
Use the output¶
count counts written documents. file_ids and file_names cover all results; file_id and file_name identify the primary result.
Use file_ids for later batch processing and retain all files and their source relationships.
Example: save chunks¶
Flow: General Document Parsing → Segment → Write Document Files.
Bind the complete chunked documents. Check the document count, file list, and actual format before passing the files to a later flow.