Update app.py

app.py

@@ -5,6 +5,9 @@ import gradio as gr
 import tempfile
 
 def grade_pdf(input_pdf, pdffigures2_jar):
+    """
+    Process a PDF file using pdffigures2 and return the extracted JSON output.
+    """
     if not os.path.exists(input_pdf):
         raise FileNotFoundError(f"Input PDF file not found: {input_pdf}")
     if not os.path.exists(pdffigures2_jar):
@@ -20,64 +23,71 @@ def grade_pdf(input_pdf, pdffigures2_jar):
         subprocess.run(command, capture_output=True, text=True, check=True)
 
         output_json_file = "_" + os.path.splitext(os.path.basename(input_pdf))[0] + ".json"
-
         if os.path.exists(output_json_file):
             with open(output_json_file, "r") as f:
                 regions = json.load(f)
 
-            issue_list = []
+            issueList = []
 
             for region in regions:
                 issues = {}
-                if region["captionBoundary"]["y1"] > region["regionBoundary"]["y1"]:
-                    region["captionLocation"] = "Below"
-                else:
-                    region["captionLocation"] = "Above"
+                caption_y = region["captionBoundary"]["y1"]
+                region_y = region["regionBoundary"]["y1"]
+                region["captionLocation"] = "Below" if caption_y > region_y else "Above"
 
                 if region["captionLocation"] == "Below" and region["figType"] == "Table":
-                    issues["page_number"] = region["page"]+1
+                    issues["page_number"] = region["page"]
                     issues["caption_coordinate"] = region["captionBoundary"]
                     issues["fig_type"] = region["figType"]
                     issues["caption_location"] = region["captionLocation"]
                     issues["description"] = "Location of the caption for tables must be above the table."
-                    issue_list.append(issues)
+                    issueList.append(issues)
 
                 if region["captionLocation"] == "Above" and region["figType"] == "Figure":
-                    issues["page_number"] = region["page"]+1
+                    issues["page_number"] = region["page"]
                     issues["caption_coordinate"] = region["captionBoundary"]
                     issues["fig_type"] = region["figType"]
                     issues["caption_location"] = region["captionLocation"]
                     issues["description"] = "Location of the caption for figures must be below the figure."
-                    issue_list.append(issues)
+                    issueList.append(issues)
 
-            return issue_list
+            return issueList
         else:
             raise Exception(f"JSON output file not found: {output_json_file}")
 
     except subprocess.CalledProcessError as e:
-        raise Exception(f"Error while running pdffigures2:\nSTDOUT:\n{e.stdout}\nSTDERR:\n{e.stderr}")
+        raise Exception(f"Error while running pdffigures2: {e.stderr}")
 
 def process_file(file):
+    """
+    Wrapper to process a file via grade_pdf and return JSON data or error as a string.
+    """
     pdffigures2_jar_path = "pdffigures2.jar"
 
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf", mode="wb") as temp_input:
-        temp_input.write(file.read())
+    temp_input = tempfile.NamedTemporaryFile(delete=False, suffix='.pdf')
+    try:
+        temp_input.write(file)
+        temp_input.close()
         temp_input_path = temp_input.name
 
-    try:
         extracted_data = grade_pdf(temp_input_path, pdffigures2_jar_path)
-        os.unlink(temp_input_path)
+
         return json.dumps(extracted_data, indent=4)
+
     except Exception as e:
-        return f"An error occurred:\n{str(e)}"
+        return f"An error occurred: {str(e)}"
+
+    finally:
+        if os.path.exists(temp_input.name):
+            os.unlink(temp_input.name)
 
 def gradio_interface():
     interface = gr.Interface(
         fn=process_file,
-        inputs=gr.File(label="Upload PDF File", file_types=[".pdf"]),
-        outputs=gr.Textbox(label="Extracted JSON Data", lines=20),
-        title="PDF Grading Interface",
-        description="Upload a PDF to detect improper caption positions for tables and figures."
+        inputs=gr.File(label="Upload PDF File", file_types=[".pdf"], type="binary"),
+        outputs=gr.Textbox(label="Output", lines=20),
+        title="PDF Caption Grader",
+        description="Detects misplacement of captions in figures and tables in PDF files."
     )
     return interface