Spaces:
Sleeping
Sleeping
Fix style and parameters
Browse files
app.py
CHANGED
@@ -3,15 +3,17 @@ from gradio_huggingfacehub_search import HuggingfaceHubSearch
|
|
3 |
import nbformat as nbf
|
4 |
from huggingface_hub import HfApi
|
5 |
|
|
|
6 |
def create_notebook_file(cell_commands, notebook_name="generated_notebook.ipynb"):
|
7 |
nb = nbf.v4.new_notebook()
|
8 |
-
nb[
|
9 |
|
10 |
-
with open(notebook_name,
|
11 |
nbf.write(nb, f)
|
12 |
-
|
13 |
print(f"Notebook '{notebook_name}' created successfully.")
|
14 |
|
|
|
15 |
def push_notebook(file_path, dataset_id, token):
|
16 |
api = HfApi(token=token)
|
17 |
api.upload_file(
|
@@ -20,40 +22,36 @@ def push_notebook(file_path, dataset_id, token):
|
|
20 |
repo_id=dataset_id,
|
21 |
repo_type="dataset",
|
22 |
)
|
|
|
23 |
print("Notebook uploaded to Huggingface Hub.")
|
24 |
-
link =
|
|
|
|
|
25 |
return f'<a target="_blank" href="{link}" style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">See notebook</a>'
|
26 |
|
27 |
-
|
28 |
-
|
29 |
-
# TODO: Handle auth error
|
30 |
# TODO: Get first config and split? or generate a dataframe per each split maybe?
|
31 |
commands = [
|
32 |
-
|
33 |
-
|
34 |
f"df = pd.read_parquet('hf://datasets/{dataset_id}/data/train-00000-of-00001.parquet')",
|
35 |
-
|
36 |
]
|
37 |
notebook_name = f"{dataset_id.replace('/', '-')}.ipynb"
|
38 |
create_notebook_file(commands, notebook_name=notebook_name)
|
39 |
-
api.upload_file(
|
40 |
-
path_or_fileobj=notebook_name,
|
41 |
-
path_in_repo="dataset_analysis.ipynb",
|
42 |
-
repo_id="asoria/en-text",
|
43 |
-
repo_type="dataset",
|
44 |
-
)
|
45 |
-
# TODO: Handle permission error
|
46 |
print("Notebook uploaded to Huggingface Hub.")
|
47 |
return notebook_name
|
48 |
|
|
|
49 |
with gr.Blocks() as demo:
|
50 |
gr.Markdown("# 🤖 Dataset auto analyst creator 🕵️")
|
51 |
dataset_name = HuggingfaceHubSearch(
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
|
58 |
@gr.render(inputs=dataset_name)
|
59 |
def embed(name):
|
@@ -72,28 +70,30 @@ with gr.Blocks() as demo:
|
|
72 |
generate_btn = gr.Button("Generate notebook and push to repo", visible=True)
|
73 |
|
74 |
download_link = gr.File(label="Download Notebook")
|
75 |
-
generate_btn.click(
|
|
|
|
|
76 |
with gr.Row() as auth_page:
|
77 |
with gr.Column():
|
78 |
auth_title = gr.Markdown(
|
79 |
"Enter your token ([settings](https://huggingface.co/settings/tokens)):"
|
80 |
)
|
81 |
-
token_box = gr.Textbox(
|
|
|
82 |
)
|
83 |
auth_error = gr.Markdown("", visible=False)
|
84 |
|
85 |
def auth(token):
|
86 |
if not token:
|
87 |
return {
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
return {
|
92 |
auth_error: gr.Markdown(value="", visible=False),
|
93 |
-
push_btn: gr.Row(visible=True)
|
94 |
}
|
95 |
|
96 |
-
|
97 |
push_btn = gr.Button("Push notebook to repo", visible=False)
|
98 |
token_box.change(
|
99 |
auth,
|
@@ -102,5 +102,9 @@ with gr.Blocks() as demo:
|
|
102 |
)
|
103 |
output_lbl = gr.HTML(value="")
|
104 |
|
105 |
-
push_btn.click(
|
106 |
-
|
|
|
|
|
|
|
|
|
|
3 |
import nbformat as nbf
|
4 |
from huggingface_hub import HfApi
|
5 |
|
6 |
+
|
7 |
def create_notebook_file(cell_commands, notebook_name="generated_notebook.ipynb"):
|
8 |
nb = nbf.v4.new_notebook()
|
9 |
+
nb["cells"] = [nbf.v4.new_code_cell(command) for command in cell_commands]
|
10 |
|
11 |
+
with open(notebook_name, "w") as f:
|
12 |
nbf.write(nb, f)
|
13 |
+
|
14 |
print(f"Notebook '{notebook_name}' created successfully.")
|
15 |
|
16 |
+
|
17 |
def push_notebook(file_path, dataset_id, token):
|
18 |
api = HfApi(token=token)
|
19 |
api.upload_file(
|
|
|
22 |
repo_id=dataset_id,
|
23 |
repo_type="dataset",
|
24 |
)
|
25 |
+
# TODO: Handle permission error
|
26 |
print("Notebook uploaded to Huggingface Hub.")
|
27 |
+
link = (
|
28 |
+
f"https://huggingface.co/datasets/{dataset_id}/blob/main/dataset_analyst.ipynb"
|
29 |
+
)
|
30 |
return f'<a target="_blank" href="{link}" style="color: var(--link-text-color); text-decoration: underline;text-decoration-style: dotted;">See notebook</a>'
|
31 |
|
32 |
+
|
33 |
+
def generate_notebook(dataset_id):
|
|
|
34 |
# TODO: Get first config and split? or generate a dataframe per each split maybe?
|
35 |
commands = [
|
36 |
+
"!pip install pandas",
|
37 |
+
"import pandas as pd",
|
38 |
f"df = pd.read_parquet('hf://datasets/{dataset_id}/data/train-00000-of-00001.parquet')",
|
39 |
+
"df.head()",
|
40 |
]
|
41 |
notebook_name = f"{dataset_id.replace('/', '-')}.ipynb"
|
42 |
create_notebook_file(commands, notebook_name=notebook_name)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
print("Notebook uploaded to Huggingface Hub.")
|
44 |
return notebook_name
|
45 |
|
46 |
+
|
47 |
with gr.Blocks() as demo:
|
48 |
gr.Markdown("# 🤖 Dataset auto analyst creator 🕵️")
|
49 |
dataset_name = HuggingfaceHubSearch(
|
50 |
+
label="Hub Dataset ID",
|
51 |
+
placeholder="Search for dataset id on Huggingface",
|
52 |
+
search_type="dataset",
|
53 |
+
value="",
|
54 |
+
)
|
55 |
|
56 |
@gr.render(inputs=dataset_name)
|
57 |
def embed(name):
|
|
|
70 |
generate_btn = gr.Button("Generate notebook and push to repo", visible=True)
|
71 |
|
72 |
download_link = gr.File(label="Download Notebook")
|
73 |
+
generate_btn.click(
|
74 |
+
generate_notebook, inputs=[dataset_name], outputs=[download_link]
|
75 |
+
)
|
76 |
with gr.Row() as auth_page:
|
77 |
with gr.Column():
|
78 |
auth_title = gr.Markdown(
|
79 |
"Enter your token ([settings](https://huggingface.co/settings/tokens)):"
|
80 |
)
|
81 |
+
token_box = gr.Textbox(
|
82 |
+
"", label="token", placeholder="hf_xxx", type="password"
|
83 |
)
|
84 |
auth_error = gr.Markdown("", visible=False)
|
85 |
|
86 |
def auth(token):
|
87 |
if not token:
|
88 |
return {
|
89 |
+
auth_error: gr.Markdown(value="", visible=False),
|
90 |
+
push_btn: gr.Row(visible=False),
|
91 |
+
}
|
92 |
return {
|
93 |
auth_error: gr.Markdown(value="", visible=False),
|
94 |
+
push_btn: gr.Row(visible=True),
|
95 |
}
|
96 |
|
|
|
97 |
push_btn = gr.Button("Push notebook to repo", visible=False)
|
98 |
token_box.change(
|
99 |
auth,
|
|
|
102 |
)
|
103 |
output_lbl = gr.HTML(value="")
|
104 |
|
105 |
+
push_btn.click(
|
106 |
+
push_notebook,
|
107 |
+
inputs=[download_link, dataset_name, token_box],
|
108 |
+
outputs=[output_lbl],
|
109 |
+
)
|
110 |
+
demo.launch()
|