Update app.py

app.py

@@ -5,7 +5,8 @@ import os
 import uuid
 import shutil
 import tempfile
-
+import pandas as pd
+import io
 # --- State Management and API Client ---
 
 def get_hf_api(token):
@@ -13,7 +14,29 @@ def get_hf_api(token):
     return HfApi(token=token if token else None)
 
 # --- Core Logic Functions ---
+def is_tabular(filepath):
+    """Checks if the file is a supported tabular dataset."""
+    if not filepath: return False
+    ext = os.path.splitext(filepath)[1].lower()
+    return ext in ['.csv', '.tsv', '.json', '.jsonl', '.parquet']
 
+def load_tabular_content(token, repo_id, repo_type, filepath):
+    """Loads a file into a Pandas DataFrame."""
+    try:
+        api = get_hf_api(token)
+        local_path = api.hf_hub_download(repo_id=repo_id, repo_type=repo_type, filename=filepath, token=token)
+        ext = os.path.splitext(filepath)[1].lower()
+        
+        if ext == '.csv': return pd.read_csv(local_path)
+        if ext == '.tsv': return pd.read_csv(local_path, sep='\t')
+        if ext == '.parquet': return pd.read_parquet(local_path)
+        if ext == '.json': 
+            try: return pd.read_json(local_path)
+            except: return pd.read_json(local_path, lines=True) # Try JSONL
+        if ext == '.jsonl': return pd.read_json(local_path, lines=True)
+        return None
+    except Exception as e:
+        return pd.DataFrame({"Error": [str(e)]})
 def handle_token_change(token):
     """
     Called when the token is entered. Fetches user info, updates UI interactivity,
@@ -112,10 +135,9 @@ def archive_repo(token, repo_id, archive_repo, manage_repo_type_state):
 # --- File Management Functions ---
 
 def show_files_and_load_first(token, repo_id, repo_type):
-    """Lists files in a dropdown and pre-loads the first file's content."""
+    """Lists files and pre-loads the first one (Text OR DataFrame)."""
     if not repo_id:
-        gr.Warning("No repository selected.")
-        return gr.update(visible=False), gr.update(choices=[], value=None), gr.update(value="")
+        return gr.update(visible=False), gr.update(choices=[], value=None), gr.update(visible=False), gr.update(visible=False)
     try:
         api = get_hf_api(token)
         repo_files = api.list_repo_files(repo_id=repo_id, repo_type=repo_type)
@@ -123,16 +145,23 @@ def show_files_and_load_first(token, repo_id, repo_type):
 
         if not filtered_files:
             return (gr.update(visible=True), gr.update(choices=[], value=None),
-                    gr.update(value="## This repository is empty.", language='markdown'))
+                    gr.update(value="## Empty Repo", visible=True), gr.update(visible=False))
         
-        first_file_path = filtered_files[0]
-        content, language = load_file_content_backend(token, repo_id, repo_type, first_file_path)
-
-        return (gr.update(visible=True), gr.update(choices=filtered_files, value=first_file_path),
-                gr.update(value=content, language=language))
+        first_file = filtered_files[0]
+        
+        # Check type and load appropriately
+        if is_tabular(first_file):
+            df = load_tabular_content(token, repo_id, repo_type, first_file)
+            return (gr.update(visible=True), gr.update(choices=filtered_files, value=first_file),
+                    gr.update(visible=False), gr.update(value=df, visible=True))
+        else:
+            content, lang = load_file_content_backend(token, repo_id, repo_type, first_file)
+            return (gr.update(visible=True), gr.update(choices=filtered_files, value=first_file),
+                    gr.update(value=content, language=lang, visible=True), gr.update(visible=False))
+            
     except Exception as e:
-        gr.Error(f"Could not manage files: {e}")
-        return gr.update(visible=False), gr.update(choices=[], value=None), gr.update(value="")
+        gr.Error(f"Error: {e}")
+        return gr.update(visible=False), gr.update(choices=[], value=None), gr.update(visible=True), gr.update(visible=False)
 
 def load_file_content_backend(token, repo_id, repo_type, filepath):
     """Backend logic to fetch and format file content."""
@@ -150,21 +179,43 @@ def load_file_content_backend(token, repo_id, repo_type, filepath):
         return f"Error loading file: {e}", 'python'
 
 def load_file_content_for_editor(token, repo_id, repo_type, filepath):
-    """Gradio wrapper to update the code editor when a file is selected."""
-    content, language = load_file_content_backend(token, repo_id, repo_type, filepath)
-    return gr.update(value=content, language=language)
+    """Decides whether to show the Code Editor or the Data Editor."""
+    if is_tabular(filepath):
+        df = load_tabular_content(token, repo_id, repo_type, filepath)
+        # Hide Code, Show DF
+        return gr.update(visible=False), gr.update(value=df, visible=True)
+    else:
+        content, language = load_file_content_backend(token, repo_id, repo_type, filepath)
+        # Show Code, Hide DF
+        return gr.update(value=content, language=language, visible=True), gr.update(visible=False)
 
-def commit_file(token, repo_id, repo_type, filepath, content, commit_message):
-    """Commits a file to the repository."""
-    if not token: gr.Error("A write-enabled token is required."); return
+def commit_file(token, repo_id, repo_type, filepath, code_content, df_content, commit_message):
+    """Smart commit: saves text OR dataframe based on file extension."""
+    if not token: gr.Error("Write-token required."); return
     if not filepath: gr.Warning("No file selected."); return
-    if not commit_message: gr.Warning("Commit message cannot be empty."); return
+    
     try:
         api = get_hf_api(token)
-        api.upload_file(bytes(content, 'utf-8'), path_in_repo=filepath,
+        
+        if is_tabular(filepath):
+            # Handle DataFrame Saving
+            buffer = io.BytesIO()
+            ext = os.path.splitext(filepath)[1].lower()
+            if ext == '.csv': df_content.to_csv(buffer, index=False)
+            elif ext == '.tsv': df_content.to_csv(buffer, sep='\t', index=False)
+            elif ext == '.parquet': df_content.to_parquet(buffer, index=False)
+            elif ext == '.json': df_content.to_json(buffer, orient='records')
+            elif ext == '.jsonl': df_content.to_json(buffer, orient='records', lines=True)
+            
+            data_to_upload = buffer.getvalue()
+        else:
+            # Handle Text Saving
+            data_to_upload = bytes(code_content, 'utf-8')
+
+        api.upload_file(path_or_fileobj=data_to_upload, path_in_repo=filepath,
                         repo_id=repo_id, repo_type=repo_type, commit_message=commit_message)
-        gr.Info(f"Successfully committed '{filepath}' to '{repo_id}'!")
-    except Exception as e: gr.Error(f"Failed to commit file: {e}")
+        gr.Info(f"Successfully committed '{filepath}'!")
+    except Exception as e: gr.Error(f"Failed to commit: {e}")
 
 # --- Download Tab Functions ---
 
@@ -229,6 +280,7 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue"), title="Hugging Face Hub
                         gr.Markdown("### 3. Edit Files")
                         file_selector = gr.Dropdown(label="Select File", interactive=True)
                         code_editor = gr.Code(label="File Content", interactive=True)
+                        dataframe_editor = gr.Dataframe(label="Dataset Editor", interactive=True, visible=False, wrap=True) 
                         commit_message_input = gr.Textbox(label="Commit Message", placeholder="e.g., Update README.md")
                         commit_btn = gr.Button("Commit Changes", variant="primary", interactive=False)
             # Now, attach event handlers
@@ -261,7 +313,8 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue"), title="Hugging Face Hub
     
     manage_files_btn.click(fn=show_files_and_load_first,
                            inputs=[hf_token_state, manage_repo_dropdown, manage_repo_type_state],
-                           outputs=[editor_panel, file_selector, code_editor])
+                           outputs=[editor_panel, file_selector, code_editor, dataframe_editor]) # Added dataframe_editor
+
     
     archive_repo_btn.click(fn=archive_repo, inputs=[hf_token_state, selected_repo, archive_repo_name, manage_repo_type_state],
                           outputs=[manage_repo_dropdown, action_panel, editor_panel],
@@ -274,10 +327,14 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue"), title="Hugging Face Hub
 
     file_selector.change(fn=load_file_content_for_editor,
                          inputs=[hf_token_state, manage_repo_dropdown, manage_repo_type_state, file_selector],
-                         outputs=code_editor)
+                         outputs=[code_editor, dataframe_editor]) # Added dataframe_editor
+
 
     commit_btn.click(fn=commit_file,
-                     inputs=[hf_token_state, manage_repo_dropdown, manage_repo_type_state, file_selector, code_editor, commit_message_input])
+                     inputs=[hf_token_state, manage_repo_dropdown, manage_repo_type_state, file_selector, 
+                             code_editor, dataframe_editor, # Added dataframe_editor
+                             commit_message_input])
+
 
     download_btn.click(fn=download_repos_as_zip,
                        inputs=[hf_token_state, download_repo_dropdown, download_repo_type_state],