Allow multiple sources for the dataset #2276
Conversation
|
|
||
| _dataset_path = dataset_path[:-c] | ||
| path, base = posixpath.split(_dataset_path) | ||
| self._dataset_paths = {} |
There was a problem hiding this comment.
This could go with the modifs l.155 to keep the relevant code together.
There was a problem hiding this comment.
Can make sense ... my idea was to keep the context manager part as small as possible, i.e. put this outside the context manager. No problem for moving it into the with hdf5file section.
src/pyFAI/gui/pilx/MainWindow.py
Outdated
| c=0 | ||
| for i in dataset_path[-1::-1]: | ||
| if i.isdigit(): | ||
| c+=1 | ||
| else: | ||
| break |
There was a problem hiding this comment.
I guess we can assume more on the dataset path?
Do something like i = dataset_path.find("images_"); _dataset_path = dataset_path[i+6:];
There was a problem hiding this comment.
I would prefer assume as little as reasonably possible. Why enforce the basename if not needed ?
There was a problem hiding this comment.
I mean, we are still enforcing a particular format of basename here (it must have digits at the end and be contained in a group). So I thought that putting stronger conditions would allow the parsing to be more explicit.
There was a problem hiding this comment.
it would suppose changing the initData attributes again, but if we pass group_path and dataset_name on the initData method, with dataset_name None as default, and in this case, just iterates to get every dataset inside the group. That would be a general solution.
There was a problem hiding this comment.
Loic, it just imposes all datasets have the same prefix. If not, the sortoperation will not work as expected.
It is really a pity there is no way to enforce the order of dataset/groups in an HDF5 file. Those are just work around for this weakness.
Edgar, I am not in favour of changing the signature of methods which are supposed to be in the public API, even though I have no clue if anybody else is using it. As Linus said in https://lkml.org/lkml/2012/12/23/75 ... don't break users' programs. I like Linus because I feel really moderated in comparison with him :)
| status_bar = self.statusBar() | ||
| if status_bar: | ||
| status_bar.showMessage(error_msg) | ||
| else: |
There was a problem hiding this comment.
It is really worth continuing the execution of the function if the image group cannot be accessed ?
There was a problem hiding this comment.
There are plenty of self.blabla initialized later on. Putting a return there would prevent them from being initialized. I have no clue what would be the behaviour of the application. You are the author more than me.
If you believe it is safe to return there, I see no objection.
There was a problem hiding this comment.
Let's leave it like that, I'll check later.
There was a problem hiding this comment.
Sometimes when I was writing and testing the widget, I moved the file or deleted the raw data, so there was no 2D pattern displayed but I still could test the map and the integrated pattern. I think the program should continue the execution, at least if we cannot see the data, we know immediately that there is a problem with the path of the data
Co-authored-by: Loïc Huder <42204205+loichuder@users.noreply.github.com>
Co-authored-by: Loïc Huder <42204205+loichuder@users.noreply.github.com>
| status_bar = self.statusBar() | ||
| if status_bar: | ||
| status_bar.showMessage(error_msg) | ||
| else: |
There was a problem hiding this comment.
Let's leave it like that, I'll check later.
To be merged after silx-kit#2276
close #2275