You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Given a lot of microscope output has very long names, I think a deep TIFF stack would quickly hit shell limitations on command length. Two ways round this:
Accept input from stdin
Accept a directory and discover files within it (sorting them) - probably better handled by piping fd | sort into option 1, but might be convenient for simple cases
Here's a repo which takes Vec<OsString>, where those OsStrings represent a mixture of file paths, directory paths, and - for stdin, and produces Vec<PathBuf> to files. Recursion into directories is optional and configurable.
Good issue. One mitigation and otherwise complementary, useful feature would just be to have a block offset to import into a (potentially pre-existing) dataset, with manually specified total dimension. This would make it easier to distribute parallel imports and limit the number of files given to any one command.
Eventually I see import going away in its current form once ndgest succeeds n5gest, because image stacks will just be another chunknd-implementing backend which can be "imported" via convert, and it will be easy to specify arbitrary sources via stdin streaming or autodiscovery like you suggest.
Given a lot of microscope output has very long names, I think a deep TIFF stack would quickly hit shell limitations on command length. Two ways round this:
fd | sort
into option 1, but might be convenient for simple casesHere's a repo which takes
Vec<OsString>
, where thoseOsString
s represent a mixture of file paths, directory paths, and-
for stdin, and producesVec<PathBuf>
to files. Recursion into directories is optional and configurable.https://github.com/clbarnes/filelist
The text was updated successfully, but these errors were encountered: