> ## Documentation Index
> Fetch the complete documentation index at: https://nixtlaverse.nixtla.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Utils | DatasetsForecast

> Utility functions for datasetsforecast

##

### `download_file`

```python theme={null}
download_file(directory, source_url, decompress=False, filename=None, max_retries=3)
```

Download data from source\_url inside directory.

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `directory` | <code>([str](#str), [Path](#pathlib.Path))</code> | Custom directory where data will be downloaded. | *required* |
| `source_url` | <code>[str](#str)</code> | URL where data is hosted. | *required* |
| `decompress` | <code>[bool](#bool)</code> | Whether to decompress downloaded file. Default False. | <code>False</code> |
| `filename` | <code>[str](#str)</code> | Override filename for the downloaded file. If None, the filename is derived from the URL. | <code>None</code> |
| `max_retries` | <code>[int](#int)</code> | Maximum number of retry attempts on transient errors. | <code>3</code> |

### `extract_file`

```python theme={null}
extract_file(filepath, directory)
```

### `async_download_files`

```python theme={null}
async_download_files(path, urls)
```

Asynchronously download files from urls inside path.

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `path` | <code>([str](#str), [Path](#pathlib.Path))</code> | Directory where files will be downloaded. | *required* |
| `urls` | <code>[Iterable](#typing.Iterable)\[[str](#str)]</code> | Iterable of URLs to download. | *required* |

Example:

```python theme={null}
import os
import tempfile

import requests
gh_url = 'https://api.github.com/repos/Nixtla/datasetsforecast/contents/'
base_url = 'https://raw.githubusercontent.com/Nixtla/datasetsforecast/main'

headers = {}
gh_token = os.getenv('GITHUB_TOKEN')
if gh_token is not None:
    headers = {'Authorization': f'Bearer: {gh_token}'}
resp = requests.get(gh_url, headers=headers)
if resp.status_code != 200:
    raise Exception(resp.text)
urls = [f'{base_url}/{e["path"]}' for e in resp.json() if e['type'] == 'file']
with tempfile.TemporaryDirectory() as tmp:
    tmp = Path(tmp)
    await async_download_files(tmp, urls)
    files = list(tmp.iterdir())
    assert len(files) == len(urls)
```

### `download_files`

```python theme={null}
download_files(directory, urls)
```

Download files from urls inside directory.

**Parameters:**

| Name | Type | Description | Default |
| - | - | - | - |
| `directory` | <code>([str](#str), [Path](#pathlib.Path))</code> | Directory where files will be downloaded. | *required* |
| `urls` | <code>[Iterable](#typing.Iterable)\[[str](#str)]</code> | Iterable of URLs to download. | *required* |

Example:

```python theme={null}
with tempfile.TemporaryDirectory() as tmp:
    tmp = Path(tmp)
    fname = tmp / 'script.py'
    fname.write_text(f'''
from datasetsforecast.utils import download_files

download_files('{tmp.as_posix()}', {urls})
    ''')
    !python {fname}
    fname.unlink()
    files = list(tmp.iterdir())
    assert len(files) == len(urls)
```


## Related topics

- [datasetsforecast](/datasetsforecast/index.html.md)
- [Utils](/coreforecast/utils.md)
- [Nixtlaverse](/index.md)
- [Data augmentation with real datasets](/synforecast/docs/capabilities/augmentation_real_data.html.md)
- [Long-Horizon Original Datasets](/datasetsforecast/long_horizon2.html.md)