Generate file to download

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.ipynb filter=lfs diff=lfs merge=lfs -text

app.py

@@ -3,8 +3,6 @@ from gradio_huggingfacehub_search import HuggingfaceHubSearch
 import nbformat as nbf
 from huggingface_hub import HfApi
 
-api = HfApi()
-
 def create_notebook_file(cell_commands, notebook_name="generated_notebook.ipynb"):
     nb = nbf.v4.new_notebook()
     nb['cells'] = [nbf.v4.new_code_cell(command) for command in cell_commands]        
@@ -14,11 +12,27 @@ def create_notebook_file(cell_commands, notebook_name="generated_notebook.ipynb"
     
     print(f"Notebook '{notebook_name}' created successfully.")
 
-def generate_notebook(dataset_id):
+def push_notebook(file_path, dataset_id, token):
+    api = HfApi(token=token)
+    api.upload_file(
+        path_or_fileobj=file_path,
+        path_in_repo="dataset_analysis.ipynb",
+        repo_id=dataset_id,
+        repo_type="dataset",
+    )
+    print("Notebook uploaded to Huggingface Hub.")
+    link = f"https://huggingface.co/datasets/{dataset_id}/blob/main/dataset_analyst.ipynb"
+    return f'<a target="_blank" href="{link}"  style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">See notebook</a>'
+
+def generate_notebook(dataset_id, token):
+    api = HfApi(token=token)
+    # TODO: Handle auth error
+    # TODO: Get first config and split? or generate a dataframe per each split maybe?
     commands = [
-        f"from datasets import load_dataset",
-        f"dataset = load_dataset('{dataset_id}')",
-        f"dataset",
+        f"!pip install pandas",
+        f"import pandas as pd",
+        f"df = pd.read_parquet('hf://datasets/{dataset_id}/data/train-00000-of-00001.parquet')",
+        f"df.head()",
     ]
     notebook_name = f"{dataset_id.replace('/', '-')}.ipynb"
     create_notebook_file(commands, notebook_name=notebook_name)
@@ -28,11 +42,9 @@ def generate_notebook(dataset_id):
         repo_id="asoria/en-text",
         repo_type="dataset",
     )
+    # TODO: Handle permission error
     print("Notebook uploaded to Huggingface Hub.")
-    link = f"https://huggingface.co/datasets/{dataset_id}/blob/main/dataset_analyst.ipynb"
-    return f'<a target="_blank" href="{link}"  style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">See notebook</a>'
-
-
+    return notebook_name
 
 with gr.Blocks() as demo:
     gr.Markdown("# 🤖 Dataset auto analyst creator 🕵️")
@@ -57,7 +69,38 @@ with gr.Blocks() as demo:
             """
         return gr.HTML(value=html_code)
 
-    generate_btn = gr.Button("Generate notebook")
+    generate_btn = gr.Button("Generate notebook and push to repo", visible=True)
+
+    download_link = gr.File(label="Download Notebook")
+    generate_btn.click(generate_notebook, inputs=[dataset_name], outputs=[download_link])
+    with gr.Row() as auth_page:
+        with gr.Column():
+            auth_title = gr.Markdown(
+                "Enter your token ([settings](https://huggingface.co/settings/tokens)):"
+            )
+            token_box = gr.Textbox("", label="token", placeholder="hf_xxx", type="password"
+            )
+            auth_error = gr.Markdown("", visible=False)
+
+    def auth(token):
+        if not token:
+            return {
+                    auth_error: gr.Markdown(value="", visible=False),
+                    push_btn: gr.Row(visible=False)
+                }
+        return {
+            auth_error: gr.Markdown(value="", visible=False),
+            push_btn: gr.Row(visible=True)
+        }
+
+
+    push_btn = gr.Button("Push notebook to repo", visible=False)
+    token_box.change(
+        auth,
+        inputs=token_box,
+        outputs=[auth_error, push_btn],
+    )
     output_lbl = gr.HTML(value="")
-    generate_btn.click(generate_notebook, inputs=[dataset_name], outputs=[output_lbl])
+
+    push_btn.click(push_notebook, inputs=[download_link, dataset_name, token_box], outputs=[output_lbl])
 demo.launch()