From 23e3cc23a9b4235a306dfa70f37535e7670899ca Mon Sep 17 00:00:00 2001
From: plpycoin <103234125+plpycoin@users.noreply.github.com>
Date: Sun, 8 Dec 2024 07:57:04 +0800
Subject: [PATCH] local GPU's conversion of PDFs is really too slow. Just can't
 bear it anymore (#229)

time consumption has been reduced from around 200s to around 15s

-----
* turn off OCR
* switch to Pdfium for the backend
*  turn on cell matching
---
 server/app/routes/convert.py | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/server/app/routes/convert.py b/server/app/routes/convert.py
index a896bb37..2c680831 100644
--- a/server/app/routes/convert.py
+++ b/server/app/routes/convert.py
@@ -64,8 +64,23 @@ async def convert_documents(files: List[UploadFile] = File(...)):
             print(f"Modal endpoint failed: {str(e)}. Falling back to local processing...")
     
     # If Modal fails, fall back to local processing
-    from docling.document_converter import DocumentConverter
-    doc_converter = DocumentConverter()
+    from docling.datamodel.base_models import InputFormat
+    from docling.document_converter import DocumentConverter, PdfFormatOption
+    from docling.datamodel.pipeline_options import PdfPipelineOptions
+    from docling.backend.pypdfium2_backend import PyPdfiumDocumentBackend
+    
+    pipeline_options = PdfPipelineOptions()
+    pipeline_options.do_ocr = False
+    pipeline_options.do_table_structure = True
+    pipeline_options.table_structure_options.do_cell_matching = True
+
+    doc_converter = DocumentConverter(
+        format_options={
+            InputFormat.PDF: PdfFormatOption(
+                pipeline_options=pipeline_options, backend=PyPdfiumDocumentBackend
+            )
+        }
+    )
     
     # Create a temporary directory to store uploaded files
     with tempfile.TemporaryDirectory() as temp_dir: