From 4bae02c2951f3d6fce9dfc2cd5ccdedae42be662 Mon Sep 17 00:00:00 2001
From: myhloli <moe@myhloli.com>
Date: Wed, 27 May 2026 23:19:24 +0800
Subject: [PATCH] feat: refactor finalize_middle_json to support server-side
 post-processing and enhance title leveling

---
 mineru/backend/hybrid/hybrid_analyze.py | 3 ++-
 mineru/cli/client.py                    | 6 +++++-
 mineru/cli/gradio_app.py                | 4 ++--
 3 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/mineru/backend/hybrid/hybrid_analyze.py b/mineru/backend/hybrid/hybrid_analyze.py
index 0c84c455..da70bbbd 100644
--- a/mineru/backend/hybrid/hybrid_analyze.py
+++ b/mineru/backend/hybrid/hybrid_analyze.py
@@ -989,7 +989,8 @@ async def aio_doc_analyze(
             )
 
         if client_side_output_generation:
-            apply_server_side_postprocess(
+            await asyncio.to_thread(
+                apply_server_side_postprocess,
                 middle_json["pdf_info"],
                 hybrid_pipeline_model,
                 _ocr_enable,
diff --git a/mineru/cli/client.py b/mineru/cli/client.py
index d90727d3..0af1f7d6 100644
--- a/mineru/cli/client.py
+++ b/mineru/cli/client.py
@@ -820,7 +820,11 @@ async def run_planned_task(
                 parse_method,
                 is_office=document.suffix in office_suffixes,
             )
-            regenerate_client_side_outputs(parse_dir, document.stem)
+            await asyncio.to_thread(
+                regenerate_client_side_outputs,
+                parse_dir,
+                document.stem,
+            )
     completed_tasks, completed_pages = await mark_task_completed(
         progress,
         planned_task.total_pages,
diff --git a/mineru/cli/gradio_app.py b/mineru/cli/gradio_app.py
index b3ff0c54..509a3095 100644
--- a/mineru/cli/gradio_app.py
+++ b/mineru/cli/gradio_app.py
@@ -1073,8 +1073,9 @@ async def _run_to_markdown_job(
         parse_method,
         allow_office_fallback=True,
     )
+    emit_status(STATUS_PROCESSING_OUTPUT)
     if use_client_side_output_generation:
-        regenerate_client_side_outputs(local_md_dir, file_name)
+        await asyncio.to_thread(regenerate_client_side_outputs, local_md_dir, file_name)
 
     preview_pdf_path = maybe_generate_local_preview(
         extract_root=extract_root,
@@ -1084,7 +1085,6 @@ async def _run_to_markdown_job(
         parse_method=parse_method,
     )
 
-    emit_status(STATUS_PROCESSING_OUTPUT)
     zip_archive_success = compress_directory_to_zip(local_md_dir, archive_zip_path)
     if zip_archive_success == 0:
         logger.info('Compression successful')