Create app.py

app.py

@@ -0,0 +1,95 @@
+import subprocess
+import json
+import os
+import gradio as gr
+
+# Set JAVA_HOME environment variable
+os.environ['JAVA_HOME'] = '/usr/lib/jvm/java-11-openjdk-amd64'
+
+def grade_pdf(input_pdf, pdffigures2_jar):
+    """
+    Process a PDF file using pdffigures2 and return the extracted JSON output.
+
+    Args:
+        input_pdf (str): Path to the input PDF file.
+        output_dir (str): Path to the directory where JSON output will be saved.
+        pdffigures2_jar (str): Path to the pdffigures2 JAR file.
+
+    Returns:
+        dict: Parsed JSON data extracted by pdffigures2.
+
+    Raises:
+        FileNotFoundError: If the input PDF or JAR file does not exist.
+        Exception: If the command fails or the JSON output is not generated.
+    """
+    # Check if input file and JAR file exist
+    if not os.path.exists(input_pdf):
+        raise FileNotFoundError(f"Input PDF file not found: {input_pdf}")
+    if not os.path.exists(pdffigures2_jar):
+        raise FileNotFoundError(f"pdffigures2 JAR file not found: {pdffigures2_jar}")
+
+
+    # Command to execute pdffigures2
+    command = [
+        "java", "-jar", pdffigures2_jar,
+        input_pdf,
+        "-d", '_'
+    ]
+    print(input_pdf)
+
+    try:
+        # Run the command
+        subprocess.run(command, capture_output=True, text=True, check=True)
+
+        # Construct the output JSON file path
+        # print()
+        output_json_file = "_"+os.path.splitext(os.path.basename(input_pdf))[0] +".json"
+
+        # Read and return the JSON data
+        if os.path.exists(output_json_file):
+            with open(output_json_file, "r") as f:
+                return json.load(f)
+        else:
+            raise Exception(f"JSON output file not found: {output_json_file}")
+
+    except subprocess.CalledProcessError as e:
+        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.
+
+    Args:
+        file (str): Path to the uploaded PDF file.
+
+    Returns:
+        str: JSON string of extracted data.
+    """
+    # Define paths
+    pdffigures2_jar_path = "pdffigures2.jar"  # Path to pdffigures2 JAR file
+
+    try:
+        # Process the PDF and get JSON data
+        # print(file.name)
+        extracted_data = grade_pdf(file.name, pdffigures2_jar_path)
+        
+
+        # Return JSON as a formatted string
+        return json.dumps(extracted_data, indent=4)
+    except Exception as e:
+        return f"An error occurred: {e}"
+
+# Gradio Interface
+def gradio_interface():
+    interface = gr.Interface(
+        fn=process_file,
+        inputs=gr.File(label="Upload PDF File"),
+        outputs=gr.Textbox(label="Extracted JSON Data", lines=20),
+        title="PDF Grading Interface",
+        description="Upload a PDF file, and this tool will extract figures and captions as JSON."
+    )
+    return interface
+
+if __name__ == "__main__":
+    # Launch the Gradio app
+    gradio_interface().launch()