ubelt.util_download_manager module

A simple download manager

class ubelt.util_download_manager.DownloadManager(download_root=None, mode='thread', max_workers=None, cache=True)[source]

Bases: object

Simple implementation of the download manager

Example

>>> # xdoctest: +REQUIRES(--network)
>>> import ubelt as ub
>>> # Download a file with a known hash
>>> manager = ub.DownloadManager()
>>> job = manager.submit(
>>>     'http://i.imgur.com/rqwaDag.png',
>>>     hash_prefix='31a129618c87dd667103e7154182e3c39a605eefe90f84f2283f3c87efee8e40'
>>> )
>>> fpath = job.result()
>>> print('fpath = {!r}'.format(fpath))

Example

>>> # Does not require network
>>> import ubelt as ub
>>> manager = ub.DownloadManager()
>>> for i in range(100):
...     job = manager.submit('localhost/might-not-exist-i-{}'.format(i))
>>> file_paths = []
>>> for job in manager.as_completed(prog=True):
...     try:
...         fpath = job.result()
...         file_paths += [fpath]
...     except Exception:
...         pass
>>> print('file_paths = {!r}'.format(file_paths))

Example

>>> # xdoctest: +REQUIRES(--network)
>>> import pytest
>>> import ubelt as ub
>>> manager = ub.DownloadManager()
>>> item1 = {
>>>     'url': 'https://data.kitware.com/api/v1/item/5b4039308d777f2e6225994c/download',
>>>     'dst': 'forgot_what_the_name_really_is',
>>>     'hash_prefix': 'c98a46cb31205cf',
>>>     'hasher': 'sha512',
>>> }
>>> item2 = {
>>>     'url': 'http://i.imgur.com/rqwaDag.png',
>>>     'hash_prefix': 'f79ea24571da6ddd2ba12e3d57b515249ecb8a35',
>>>     'hasher': 'sha1',
>>> }
>>> item1 = item2  # hack around SSL error
>>> manager.submit(**item1)
>>> manager.submit(**item2)
>>> for job in manager.as_completed(prog=True, verbose=3):
>>>     fpath = job.result()
>>>     print('fpath = {!r}'.format(fpath))
Parameters:
  • download_root (str | PathLike) – default download location

  • mode (str) – either thread, process, or serial

  • cache (bool) – defaults to True

  • max_workers (int | None) – maximum concurrent tasks

Todo

  • [ ] Will likely have to initialize and store some sort of

    “connection state” objects.

submit(url, dst=None, hash_prefix=None, hasher='sha256')[source]

Add a job to the download Queue

Parameters:
  • url (str | PathLike) – pointer to the data to download

  • dst (str | None) – The relative or absolute path to download to. If unspecified, the destination name is derived from the url.

  • hash_prefix (str | None) – If specified, verifies that the hash of the downloaded file starts with this.

  • hasher (str, default=’sha256’) – hashing algorithm to use if hash_prefix is specified.

Returns:

a Future object that will point to the downloaded location.

Return type:

concurrent.futures.Future

as_completed(prog=None, desc=None, verbose=1)[source]

Generate completed jobs as they become available

Example

>>> import pytest
>>> import ubelt as ub
>>> download_root = ub.ensure_app_config_dir('ubelt', 'dlman')
>>> manager = ub.DownloadManager(download_root=download_root,
>>>                              cache=False)
>>> for i in range(3):
>>>     manager.submit('localhost')
>>> results = list(manager)
>>> print('results = {!r}'.format(results))
>>> manager.shutdown()
shutdown()[source]

Cancel all jobs and close all connections.