a

app.py

@@ -10,9 +10,20 @@ from marker.pdf.extract_text import get_length_of_text
 from marker.models import load_all_models
 from marker.settings import settings
 from marker.logger import configure_logging
+import traceback
+
+
+configure_logging()
+MAX_PAGES = 20
+MIN_LENGTH=200
+settings.EXTRACT_IMAGES = False
+settings.DEBUG = False
+
+model_refs = load_all_models()
+metadata = {}
 
-model_name = "maxidl/arena-test"
 
+model_name = "maxidl/arena-test"
 model = AutoModelForCausalLM.from_pretrained(
     model_name,
     torch_dtype=torch.bfloat16,
@@ -25,8 +36,33 @@ title = "# Placeholder Title"
 steps = """Placeholder Description"""
 # steps = """1. Converts uploaded pdf file to markdown. You can edit the intermediate markdown output.\n2. Generates a review for the paper"""
 
+@spaces.GPU(duration=60)
+def convert_file(filepath):
+    full_text, images, out_metadata = convert_single_pdf(
+            filepath, model_refs, metadata=metadata, max_pages=MAX_PAGES
+    )
+    return full_text.strip()
+
 def process_file(file):
-    return "Processed file"
+    print(file.name)
+    filepath = file.name
+    try:
+        if MIN_LENGTH:
+            filetype = find_filetype(filepath)
+            if filetype == "other":
+                raise ValueError()
+
+            length = get_length_of_text(filepath)
+            if length < MIN_LENGTH:
+                raise ValueError()
+        paper_text = convert_file(filepath)
+        if not len(paper_text) > MIN_LENGTH:
+            raise ValueError()
+    except Exception as e:
+        print(traceback.format_exc())
+        print(f"Error converting {filepath}: {e}")
+        return "Error processing pdf"
+    return paper_text
 
 
 @spaces.GPU(duration=60)