perf: Clean up resources with docling-parse v4, no parsed_page output by default (#2105)

* Call PdfDocument.unload_pages from the pipelines where needed, delete parsed_page data unless requested to keep

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* pin docling-parse and update lock

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* Reinstate pipeline_options.generate_parsed_page

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

---------

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>
Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
Co-authored-by: Michele Dolfi <dol@zurich.ibm.com>
This commit is contained in:
Christoph Auer
2025-08-20 10:46:31 +02:00
committed by GitHub
parent c5f2e2fdd6
commit 5f57ff2a45
8 changed files with 110 additions and 65 deletions

View File

@@ -27,6 +27,7 @@ def get_converter():
pipeline_options.do_table_structure = True
pipeline_options.table_structure_options.do_cell_matching = True
pipeline_options.accelerator_options.device = AcceleratorDevice.CPU
pipeline_options.generate_parsed_pages = True
converter = DocumentConverter(
format_options={