vivekprojects-GIT commited on
Commit ·
0973090
1
Parent(s): f2e8372
Update default num_samples to 10 in generate_synthetic_data
Browse files- data_processor.py +28 -0
data_processor.py
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
def generate_synthetic_data(original_data: pd.DataFrame, num_samples: int = 10, output_dir: str = "generated") -> pd.DataFrame:
|
| 2 |
+
"""
|
| 3 |
+
Generates synthetic data using CTGAN based on original data.
|
| 4 |
+
|
| 5 |
+
Args:
|
| 6 |
+
original_data (pd.DataFrame): The original dataset.
|
| 7 |
+
num_samples (int, optional): Number of synthetic samples to generate (default is 10).
|
| 8 |
+
output_dir (str, optional): Directory to save the generated Parquet file.
|
| 9 |
+
|
| 10 |
+
Returns:
|
| 11 |
+
pd.DataFrame: Generated synthetic data.
|
| 12 |
+
"""
|
| 13 |
+
# Train CTGAN model
|
| 14 |
+
model = CTGAN()
|
| 15 |
+
model.fit(original_data)
|
| 16 |
+
|
| 17 |
+
# Generate synthetic data
|
| 18 |
+
synthetic_data = model.sample(num_samples)
|
| 19 |
+
|
| 20 |
+
# Create output directory if it doesn't exist
|
| 21 |
+
os.makedirs(output_dir, exist_ok=True)
|
| 22 |
+
|
| 23 |
+
# Save synthetic data as Parquet file with timestamp
|
| 24 |
+
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
| 25 |
+
file_path = os.path.join(output_dir, f"synthetic_data_{timestamp}.parquet")
|
| 26 |
+
synthetic_data.to_parquet(file_path)
|
| 27 |
+
|
| 28 |
+
return synthetic_data
|