Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -62,31 +62,29 @@ def Dataset_Creator_Function(dataset_name: str, conversation_data: str) -> str:
|
|
| 62 |
URL of the created dataset or error message
|
| 63 |
"""
|
| 64 |
try:
|
| 65 |
-
# Get API key from environment variables
|
| 66 |
api_key = os.getenv("HF_API_KEY") or os.getenv("HUGGINGFACE_API_KEY", "")
|
| 67 |
if not api_key:
|
| 68 |
return "Error: No Hugging Face API key found in environment variables"
|
| 69 |
|
|
|
|
|
|
|
|
|
|
| 70 |
# Initialize Hugging Face API
|
| 71 |
hf_api = HfApi(token=api_key)
|
| 72 |
|
| 73 |
-
# Retrieve username for the repository name
|
| 74 |
-
try:
|
| 75 |
-
user_info = hf_api.whoami()
|
| 76 |
-
username = user_info.get("name", "Misfits-and-Machines") # fallback
|
| 77 |
-
except Exception as e:
|
| 78 |
-
return f"Error authenticating with Hugging Face: {str(e)}"
|
| 79 |
-
|
| 80 |
# Sanitize dataset name and create repo_id
|
| 81 |
safe_dataset_name = dataset_name.replace(" ", "-").lower()
|
| 82 |
repo_id = f"{username}/{safe_dataset_name}"
|
| 83 |
|
| 84 |
-
print(f"Creating
|
| 85 |
-
|
|
|
|
| 86 |
try:
|
| 87 |
hf_api.create_repo(repo_id=repo_id, repo_type="dataset", exist_ok=True)
|
|
|
|
| 88 |
except Exception as repo_error:
|
| 89 |
-
print(f"
|
| 90 |
|
| 91 |
# Build conversation object
|
| 92 |
conversation = {
|
|
@@ -98,16 +96,15 @@ def Dataset_Creator_Function(dataset_name: str, conversation_data: str) -> str:
|
|
| 98 |
# Create a Hugging Face dataset
|
| 99 |
dataset = Dataset.from_dict({"conversations": [conversation]})
|
| 100 |
|
| 101 |
-
# Push to the Hub
|
| 102 |
dataset.push_to_hub(
|
| 103 |
repo_id=repo_id,
|
| 104 |
-
token=api_key
|
| 105 |
-
commit_message="Add conversation dataset"
|
| 106 |
)
|
| 107 |
|
| 108 |
dataset_url = f"https://huggingface.co/datasets/{repo_id}"
|
| 109 |
-
print(f"Dataset successfully pushed to {dataset_url}")
|
| 110 |
-
return dataset_url
|
| 111 |
except Exception as e:
|
| 112 |
import traceback
|
| 113 |
error_trace = traceback.format_exc()
|
|
@@ -121,13 +118,16 @@ def Dataset_Creator_Tool(dataset_name: str, conversation_data: str) -> str:
|
|
| 121 |
Args:
|
| 122 |
dataset_name: Name for the dataset (will be prefixed with username)
|
| 123 |
conversation_data: String content to save to the dataset (no JSON conversion needed)
|
| 124 |
-
"""
|
| 125 |
try:
|
|
|
|
| 126 |
result = Dataset_Creator_Function(dataset_name, conversation_data)
|
| 127 |
-
print(f"
|
| 128 |
return result
|
| 129 |
except Exception as e:
|
| 130 |
-
|
|
|
|
|
|
|
| 131 |
|
| 132 |
|
| 133 |
@tool
|
|
|
|
| 62 |
URL of the created dataset or error message
|
| 63 |
"""
|
| 64 |
try:
|
| 65 |
+
# Get API key from environment variables
|
| 66 |
api_key = os.getenv("HF_API_KEY") or os.getenv("HUGGINGFACE_API_KEY", "")
|
| 67 |
if not api_key:
|
| 68 |
return "Error: No Hugging Face API key found in environment variables"
|
| 69 |
|
| 70 |
+
# Force the username to be the known value
|
| 71 |
+
username = "Misfits-and-Machines"
|
| 72 |
+
|
| 73 |
# Initialize Hugging Face API
|
| 74 |
hf_api = HfApi(token=api_key)
|
| 75 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 76 |
# Sanitize dataset name and create repo_id
|
| 77 |
safe_dataset_name = dataset_name.replace(" ", "-").lower()
|
| 78 |
repo_id = f"{username}/{safe_dataset_name}"
|
| 79 |
|
| 80 |
+
print(f"Creating dataset repository: {repo_id}")
|
| 81 |
+
|
| 82 |
+
# Create the repository explicitly
|
| 83 |
try:
|
| 84 |
hf_api.create_repo(repo_id=repo_id, repo_type="dataset", exist_ok=True)
|
| 85 |
+
print(f"Repository {repo_id} created or confirmed")
|
| 86 |
except Exception as repo_error:
|
| 87 |
+
print(f"Repository creation note: {str(repo_error)}")
|
| 88 |
|
| 89 |
# Build conversation object
|
| 90 |
conversation = {
|
|
|
|
| 96 |
# Create a Hugging Face dataset
|
| 97 |
dataset = Dataset.from_dict({"conversations": [conversation]})
|
| 98 |
|
| 99 |
+
# Push dataset to the Hub with minimal parameters
|
| 100 |
dataset.push_to_hub(
|
| 101 |
repo_id=repo_id,
|
| 102 |
+
token=api_key
|
|
|
|
| 103 |
)
|
| 104 |
|
| 105 |
dataset_url = f"https://huggingface.co/datasets/{repo_id}"
|
| 106 |
+
print(f"Dataset successfully pushed to: {dataset_url}")
|
| 107 |
+
return f"Successfully created dataset: {dataset_url}"
|
| 108 |
except Exception as e:
|
| 109 |
import traceback
|
| 110 |
error_trace = traceback.format_exc()
|
|
|
|
| 118 |
Args:
|
| 119 |
dataset_name: Name for the dataset (will be prefixed with username)
|
| 120 |
conversation_data: String content to save to the dataset (no JSON conversion needed)
|
| 121 |
+
"""
|
| 122 |
try:
|
| 123 |
+
print(f"Creating dataset '{dataset_name}' with {len(conversation_data)} characters of data")
|
| 124 |
result = Dataset_Creator_Function(dataset_name, conversation_data)
|
| 125 |
+
print(f"Result: {result}")
|
| 126 |
return result
|
| 127 |
except Exception as e:
|
| 128 |
+
import traceback
|
| 129 |
+
error_trace = traceback.format_exc()
|
| 130 |
+
return f"Error using Dataset Creator tool: {str(e)}\n{error_trace}"
|
| 131 |
|
| 132 |
|
| 133 |
@tool
|