Update app.py
Browse files
app.py
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
import pandas as pd
|
| 3 |
import io
|
| 4 |
-
from huggingface_hub import HfApi, HfFolder, Repository
|
| 5 |
import os
|
|
|
|
| 6 |
|
| 7 |
# Global state
|
| 8 |
df = pd.DataFrame()
|
|
@@ -10,35 +10,25 @@ df = pd.DataFrame()
|
|
| 10 |
def upload_csv(file):
|
| 11 |
global df
|
| 12 |
df = pd.read_csv(file.name)
|
| 13 |
-
|
| 14 |
if "text" not in df.columns or "label" not in df.columns:
|
| 15 |
-
return
|
| 16 |
-
|
| 17 |
-
# Fill label column if empty
|
| 18 |
df["label"] = df["label"].fillna("")
|
| 19 |
-
|
| 20 |
-
# Return the editable table
|
| 21 |
-
return gr.Dataframe(
|
| 22 |
-
value=df,
|
| 23 |
-
headers=["text", "label"],
|
| 24 |
-
interactive=True,
|
| 25 |
-
label="Edit labels below"
|
| 26 |
-
), "File uploaded successfully."
|
| 27 |
|
| 28 |
-
def
|
| 29 |
global df
|
| 30 |
df = pd.DataFrame(updated_table, columns=["text", "label"])
|
| 31 |
-
return "Changes saved."
|
| 32 |
|
| 33 |
def download_csv():
|
| 34 |
-
|
| 35 |
-
csv_bytes = df.to_csv(index=False).encode()
|
| 36 |
return gr.File.update(value=io.BytesIO(csv_bytes), filename="annotated_data.csv")
|
| 37 |
|
| 38 |
def push_to_hub(repo_name, hf_token):
|
| 39 |
-
|
| 40 |
repo_url = f"https://huggingface.co/datasets/{repo_name}"
|
| 41 |
-
local_path = f"./{repo_name}"
|
| 42 |
|
| 43 |
if os.path.exists(local_path):
|
| 44 |
os.system(f"rm -rf {local_path}")
|
|
@@ -46,35 +36,42 @@ def push_to_hub(repo_name, hf_token):
|
|
| 46 |
api = HfApi()
|
| 47 |
api.create_repo(repo_id=repo_name, token=hf_token, repo_type="dataset", exist_ok=True)
|
| 48 |
repo = Repository(local_dir=local_path, clone_from=repo_url, token=hf_token)
|
|
|
|
| 49 |
df.to_csv(f"{local_path}/data.csv", index=False)
|
| 50 |
repo.push_to_hub()
|
| 51 |
|
| 52 |
-
return f"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
|
| 54 |
-
with gr.Blocks() as demo:
|
| 55 |
-
gr.Markdown("## 🏷️ Label it! Text Labeling Tool")
|
| 56 |
-
|
| 57 |
with gr.Row():
|
| 58 |
-
|
| 59 |
upload_btn = gr.Button("Upload")
|
| 60 |
|
| 61 |
-
|
| 62 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
|
|
|
|
|
|
|
| 64 |
with gr.Row():
|
| 65 |
-
save_btn = gr.Button("Save
|
| 66 |
-
download_btn = gr.Button("Download CSV")
|
| 67 |
-
download_file = gr.File(label="
|
| 68 |
-
|
| 69 |
with gr.Row():
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
push_btn = gr.Button("Push to
|
| 73 |
-
push_status = gr.Textbox(interactive=False)
|
| 74 |
-
|
| 75 |
-
upload_btn.click(fn=upload_csv, inputs=csv_input, outputs=[df_output, upload_status])
|
| 76 |
-
save_btn.click(fn=save_edits, inputs=df_output, outputs=upload_status)
|
| 77 |
-
download_btn.click(fn=download_csv, outputs=download_file)
|
| 78 |
-
push_btn.click(fn=push_to_hub, inputs=[hf_repo, hf_token], outputs=push_status)
|
| 79 |
|
| 80 |
-
|
|
|
|
|
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
import pandas as pd
|
| 3 |
import io
|
|
|
|
| 4 |
import os
|
| 5 |
+
from huggingface_hub import HfApi, Repository
|
| 6 |
|
| 7 |
# Global state
|
| 8 |
df = pd.DataFrame()
|
|
|
|
| 10 |
def upload_csv(file):
|
| 11 |
global df
|
| 12 |
df = pd.read_csv(file.name)
|
|
|
|
| 13 |
if "text" not in df.columns or "label" not in df.columns:
|
| 14 |
+
return None, "❌ CSV must have 'text' and 'label' columns."
|
|
|
|
|
|
|
| 15 |
df["label"] = df["label"].fillna("")
|
| 16 |
+
return df[["text", "label"]], "✅ File uploaded. You can now edit labels."
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
+
def save_changes(updated_table):
|
| 19 |
global df
|
| 20 |
df = pd.DataFrame(updated_table, columns=["text", "label"])
|
| 21 |
+
return "✅ Changes saved."
|
| 22 |
|
| 23 |
def download_csv():
|
| 24 |
+
global df
|
| 25 |
+
csv_bytes = df.to_csv(index=False).encode("utf-8")
|
| 26 |
return gr.File.update(value=io.BytesIO(csv_bytes), filename="annotated_data.csv")
|
| 27 |
|
| 28 |
def push_to_hub(repo_name, hf_token):
|
| 29 |
+
global df
|
| 30 |
repo_url = f"https://huggingface.co/datasets/{repo_name}"
|
| 31 |
+
local_path = f"./{repo_name.replace('/', '_')}"
|
| 32 |
|
| 33 |
if os.path.exists(local_path):
|
| 34 |
os.system(f"rm -rf {local_path}")
|
|
|
|
| 36 |
api = HfApi()
|
| 37 |
api.create_repo(repo_id=repo_name, token=hf_token, repo_type="dataset", exist_ok=True)
|
| 38 |
repo = Repository(local_dir=local_path, clone_from=repo_url, token=hf_token)
|
| 39 |
+
|
| 40 |
df.to_csv(f"{local_path}/data.csv", index=False)
|
| 41 |
repo.push_to_hub()
|
| 42 |
|
| 43 |
+
return f"✅ Successfully pushed to: {repo_url}"
|
| 44 |
+
|
| 45 |
+
with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue")) as app:
|
| 46 |
+
gr.Markdown("# 🏷️ Label it! Text Annotation Tool")
|
| 47 |
+
gr.Markdown("Upload your CSV with `text` and `label` columns. Edit labels, save your work, and export or publish.")
|
| 48 |
|
|
|
|
|
|
|
|
|
|
| 49 |
with gr.Row():
|
| 50 |
+
file_input = gr.File(label="📁 Upload CSV", file_types=[".csv"])
|
| 51 |
upload_btn = gr.Button("Upload")
|
| 52 |
|
| 53 |
+
dataframe = gr.Dataframe(
|
| 54 |
+
headers=["text", "label"],
|
| 55 |
+
label="📝 Annotate Labels Below",
|
| 56 |
+
interactive=True,
|
| 57 |
+
visible=False,
|
| 58 |
+
row_count=10,
|
| 59 |
+
col_count=(2, "fixed")
|
| 60 |
+
)
|
| 61 |
|
| 62 |
+
status = gr.Textbox(visible=True, label="Status", interactive=False)
|
| 63 |
+
|
| 64 |
with gr.Row():
|
| 65 |
+
save_btn = gr.Button("💾 Save")
|
| 66 |
+
download_btn = gr.Button("⬇️ Download CSV")
|
| 67 |
+
download_file = gr.File(label="📥 Downloaded File")
|
| 68 |
+
|
| 69 |
with gr.Row():
|
| 70 |
+
repo_input = gr.Textbox(label="📦 Hugging Face Dataset Repo (e.g. username/my-dataset)")
|
| 71 |
+
token_input = gr.Textbox(label="🔑 Hugging Face Token", type="password")
|
| 72 |
+
push_btn = gr.Button("🚀 Push to HF Hub")
|
| 73 |
+
push_status = gr.Textbox(label="Push Status", interactive=False)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
|
| 75 |
+
upload_btn.click(fn=upload_csv, inputs=file_input, outputs=[dataframe, status])
|
| 76 |
+
save_btn.click(fn=save_changes, inputs=dataframe, outputs=status)
|
| 77 |
+
download_btn.click(fn=download_csv, outputs=download_file)_
|