dataexamples tidy #1752
dataexamples tidy #1752
Conversation
casperdcl
force-pushed
the
dataexamples-tidy
branch
from
1681b09
to
479b83e
Compare
casperdcl
force-pushed
the
dataexamples-tidy
branch
3 times, most recently
from
976dd26
to
bc3bef2
Compare
casperdcl
force-pushed
the
dataexamples-tidy
branch
from
bc3bef2
to
017a4cf
Compare
| ''' | ||
| URL = 'https://zenodo.org/records/4912435/files/small.zip' | ||
| FILE_SIZE = '227 MB' |
There was a problem hiding this comment.
@hrobarts I think you're missing a @classmethod def get(...) here
There was a problem hiding this comment.
It's not obvious which file you would want to get from the sandstone dataset. I was thinking of just having the download function then the user can load whichever one they need. That's how I would use it in the how-to. But maybe that's a bad idea. What do you think @paskino ?
There was a problem hiding this comment.
Could add an extra argument to the get function here to select the file?
| loader = ZEISSDataReader(file_name=filepath) | ||
| return loader.read() | ||
| except FileNotFoundError as exc: | ||
| raise ValueError(f"Specify a different data_dir or download data with `{cls.__name__}({data_dir}).download_data()`") from exc |
There was a problem hiding this comment.
btw @hrobarts why not this instead? It will prompt if not already downloaded...
| try: | |
| loader = ZEISSDataReader(file_name=filepath) | |
| return loader.read() | |
| except FileNotFoundError as exc: | |
| raise ValueError(f"Specify a different data_dir or download data with `{cls.__name__}({data_dir}).download_data()`") from exc | |
| self.download_data() | |
| loader = ZEISSDataReader(file_name=filepath) | |
| return loader.read() |
There was a problem hiding this comment.
We had some discussion about wanting to keep the download separate from get so the user has to actively request the download. But I agree as we have the prompt before downloading this would be neater...
There was a problem hiding this comment.
also could have def get(..., missing='prompt' / 'download' / 'error') to control (default) behaviour
Co-authored-by: Casper da Costa-Luis <[email protected]> Signed-off-by: Hannah Robarts <[email protected]>
| return res == "y" | ||
|
|
||
| def download_data(self): |
There was a problem hiding this comment.
I think we sometimes want to just download the data without getting it e.g. in the how-to for the readers, the purpose is to show the loading step. In that case I think we need to keep data_dir as an argument and maybe keep this as a class method??
|
FYI |
Changes
OOP tidy of #1712
abc.ABC->BaseTestData->TestData,NexusTestData,RemoteTestData**kwargs(Use of **kwargs in methods and functions signature is confusing #1115)_download_and_extract_from_urlTesting you performed
Related issues/links
Checklist
I have updated the relevant documentationCHANGELOG.md has been updated with any functionality changeContribution Notes
Please read and adhere to the developer guide and local patterns and conventions.