Update app.py

app.py

@@ -3,7 +3,39 @@ from database import save_to_db
 from preprocessing import read_file
 
 def process_file(file, topics):
-    """Process uploaded file and save to database."""
+    """
+    Processes an uploaded file, extracts its text content, and saves it to the database.
+
+    This function performs the following steps:
+    1. Reads the content of the uploaded file using the `read_file` function.
+       - Supports `.docx`, `.txt`, and `.pdf` file formats.
+    2. Splits the extracted text into chunks (if applicable).
+    3. Saves the processed text and associated topics to the database using the `save_to_db` function.
+    4. Returns a success message if the file is processed and saved successfully.
+
+    If any error occurs during processing, the function catches the exception and returns an error message.
+
+    Parameters:
+    ----------
+    file : object
+        The uploaded file object. The file's name (`file.name`) is used to determine the file path.
+    topics : list or str
+        A list of topics or a single topic string associated with the file. These are saved to the database along with the file content.
+
+    Returns:
+    -------
+    str
+        - A success message indicating that the file was processed and saved successfully.
+        - An error message if an exception occurs during processing.
+
+    Example:
+    --------
+    >>> process_file(uploaded_file, ["Persian Literature", "History"])
+    'File processed successfully! File saved to the database.'
+
+    >>> process_file(unsupported_file, ["Science"])
+    'Error processing file: Unsupported file format. Only .docx, .txt, and .pdf are allowed.'
+    """
     try:
         # Read the file content
         file_path = file.name
@@ -21,7 +53,7 @@ def process_file(file, topics):
 with gr.Blocks() as demo:
     gr.Markdown("# Dataset Upload Interface")
     with gr.Row():
-        file_input = gr.File(label="Upload File (.docx or .txt)")
+        file_input = gr.File(label="Upload File (.docx or .txt or .pdf)")
         topics_input = gr.Textbox(label="Topics (comma-separated)", placeholder="e.g., science, technology, law, medicin")
     submit_button = gr.Button("Upload and Process")
     output_text = gr.Textbox(label="Status")