# `np2ds`

Convert a NumPy array into a URANIE DataServer object.

```python
converter.np2ds(mat, attrib_names, name="", title="")
```

**Parameters:**

```{list-table}
:header-rows: 1
:widths: 18 20 62

* - Parameter
  - Type
  - Description
* - `mat`
  - `numpy.ndarray`
  - Input array with shape `(n_samples, n_attributes)`. 1-D arrays are automatically reshaped to `(1, n_features)`.
* - `attrib_names`
  - `str`
  - Colon-separated attribute names, e.g., `"x1:x2:x3"`. Length must match `mat.shape[1]`.
* - `name`
  - `str`
  - Name for the DataServer object (optional, default: `""`).
* - `title`
  - `str`
  - Title for the DataServer object (optional, default: `""`).
```

**Returns:**

| Type | Description |
|------|-------------|
| `DataServer.TDataServer` | A DataServer object with attributes populated from the array columns. |

## Example

```python
import numpy as np
from uratools import converter

# Create data
X = np.random.randn(100, 3)

# Convert to DataServer
tds = converter.np2ds(X, "x1:x2:x3", name="samples", title="Random samples")

print(f"Patterns: {tds.getNPatterns()}")   # 100
print(f"Attributes: {tds.getNAttributes()}")  # 3
```

## Use Cases

- Prepare NumPy data for URANIE sampling workflows
- Convert preprocessed data into URANIE format
- Create DataServer objects from external data sources (CSV, HDF5, etc.)