Update app.py

app.py

@@ -10,17 +10,13 @@ def load_huggingface_dataset(dataset_link, token):
     Load a Hugging Face dataset using the provided link and token.
     """
     # Extract dataset name and config if applicable
-    parts = dataset_link.split("/")
-    dataset_name = parts[-1]
-    if len(parts) > 2:
-        owner = parts[-2]
-        dataset_name = f"{owner}/{dataset_name}"
     
     # Load the dataset
     dataset = load_dataset(dataset_link, split="train")
     
     # Return the dataset as a DataFrame with index and transcription columns
     df = dataset.to_pandas().reset_index()
+    # print(df)
     return df[["index", "transcription"]], dataset
 
 def update_transcriptions(df, dataset, token,dataset_link):
@@ -52,13 +48,32 @@ def main():
     
     dataset = None  # To store the loaded dataset object globally
     
+    original_df = None  # Store the original DataFrame globally for resetting search results
+
     def load_dataset_and_show_table(dataset_link, token):
         """
         Load the dataset and return the DataFrame to display in Gradio.
         """
-        nonlocal dataset
-        df, dataset = load_huggingface_dataset(dataset_link, token)
-        return df
+        nonlocal dataset, original_df
+        original_df, dataset = load_huggingface_dataset(dataset_link, token)
+        return original_df
+    
+    def search_transcriptions(search_term):
+        """
+        Search the transcription column and filter the table based on the search term.
+        """
+        if original_df is None:
+            return pd.DataFrame(columns=["index", "transcription"])  # Empty table if no dataset is loaded
+        filtered_df = original_df[original_df["transcription"].str.contains(search_term, case=False, na=False)]
+        return filtered_df
+    
+    def update_original(df):
+        # Merge modified DataFrame into original DataFrame
+        for _, row in df.iterrows():
+            # Locate the row in the original DataFrame with the same index
+            original_df.loc[original_df["index"] == row["index"], "transcription"] = row["transcription"]
+            
+        return "update Successful"
     
     def submit_changes(df, token,dataset_link):
         """
@@ -67,8 +82,13 @@ def main():
         if dataset is None:
             return "No dataset loaded to update."
         
-        print(df)
-        print(token)
+        print(len(dataset))
+        print(len(df))
+        
+        if len(df) < len(dataset):
+            update_original(df)
+            return update_transcriptions(original_df, dataset, token,dataset_link)
+        
         return update_transcriptions(df, dataset, token,dataset_link)
 
     # Gradio Interface
@@ -82,6 +102,10 @@ def main():
         # Button to load dataset
         load_button = gr.Button("Load Dataset")
         
+        # Search bar
+        search_box = gr.Textbox(label="Search Transcriptions", placeholder="Enter a search term...")
+        
+        
         # Table to display and edit dataset
         table = gr.Dataframe(
             headers=["index", "transcription"],
@@ -90,8 +114,10 @@ def main():
             label="Edit Dataset (Transcriptions are RTL)",
         )
         
+        update_button = gr.Button("Update Table")
         # Button to submit changes
         submit_button = gr.Button("Submit Changes")
+        update_message = gr.Textbox(label="update message")
         output_message = gr.Textbox(label="Message")
         
         # RTL styling for transcription column
@@ -99,6 +125,8 @@ def main():
         
         # Button functionality
         load_button.click(load_dataset_and_show_table, [dataset_link, hf_token], table)
+        search_box.change(search_transcriptions, search_box, table)
+        update_button.click(update_original, [table], update_message)
         submit_button.click(submit_changes, [table, hf_token,dataset_link], output_message)
     
     # Launch Gradio Interface