Implement data api asset download and supporting functions #707
Conversation
| @respx.mock | ||
| @pytest.mark.asyncio | ||
| async def test_download_asset_img(tmpdir, open_test_img, session): | ||
| dl_url = TEST_DOWNLOAD_URL + '/1?token=IAmAToken' | ||
|
|
||
| img_headers = { | ||
| 'Content-Type': 'image/tiff', | ||
| 'Content-Length': '527', | ||
| 'Content-Disposition': 'attachment; filename="img.tif"' | ||
| } | ||
|
|
||
| async def _stream_img(): | ||
| data = open_test_img.read() | ||
| v = memoryview(data) | ||
|
|
||
| chunksize = 100 | ||
| for i in range(math.ceil(len(v) / (chunksize))): | ||
| yield v[i * chunksize:min((i + 1) * chunksize, len(v))] | ||
|
|
||
| # populate request parameter to avoid respx cloning, which throws | ||
| # an error caused by respx and not this code | ||
| # https://github.com/lundberg/respx/issues/130 | ||
| mock_resp = httpx.Response(HTTPStatus.OK, | ||
| stream=_stream_img(), | ||
| headers=img_headers, | ||
| request='donotcloneme') | ||
| respx.get(dl_url).return_value = mock_resp | ||
|
|
||
| cl = OrdersClient(session, base_url=TEST_URL) | ||
| filename = await cl.download_asset(dl_url, directory=str(tmpdir)) | ||
|
|
||
| assert Path(filename).name == 'img.tif' | ||
| assert os.path.isfile(filename) | ||
|
|
||
|
|
There was a problem hiding this comment.
this is just a move closer to it's sibling test
| print command to report wait status. `download_asset` has reporting built in. | ||
|
|
||
| ```python | ||
| >>> async def download_and_validate(): |
There was a problem hiding this comment.
I am against having the added ">>>"s and "..."s in example code because it makes it difficult for users to copy and paste and immediately use. We're also inconsistent with its usage. We add it to many of our examples, but don't have it in the checksum example (line 262 in this MD).
There was a problem hiding this comment.
I'm against it too. It is included here for consistency and we need to do a sweep of the docs to remove these. I'm not sure why it isn't in the checksum example, it probably slipped through a PR. But that isn't a part of this PR.
There was a problem hiding this comment.
The tricky thing is that >>> is kinda normalized in Python and numpy. See https://docs.python.org/3/tutorial/introduction.html#numbers and https://numpydoc.readthedocs.io/en/latest/format.html#examples. So without some doc linting, this usage might continue to creep in. For what it's worth, the Python doc source includes the >>> prompt (see https://github.com/python/cpython/blob/3.10/Doc/tutorial/introduction.rst#numbers) and the website has a js widget to hide it if you want to copy it without.
|
@kevinlacaille anything else you'd like to bring up in the review or do you feel ok with approving? |
Implements data asset functions, everything needed to download and validate an asset
Closes #499