Addition of mib importer

[chrisallen80]

Importer for .mib files from Merlin/Medipix direct electron detectors. This is a hack of the raw/rpl importer. mib files are always saved in mib/hdr pairs. When loading into hyperspy it is the hdr file which must be loaded.

• Added mib to io_plugins init.py

• Added mib.py

• Read meta-data from .hdr file

• Import various bit depth data (1,6, 12,24)

• split inter-frame headers from data

• lazy loading of data

• get frame time from inter-frame headers

• update docstring

• update user guide

• add tests

• ready for review.

Minimal example of the bug fix or the new feature

>>> import hyperspy.api as hs
>>> import numpy as np
>>> s = hs.load("filename.hdr")

[magnunor]

Nice!

This is the same code which pyXem uses?

In the example you use the header file to load the data, does the reader itself require the header file? I think all the data you really need is in the binary files.

[chrisallen80]

@magnunor it is similar to the pyXem code, with some updates to deal will all bit depths.

You are correct that all the relevant information is in the inter-frame headers in the binary files, but prior to attempting to open the binary file you don't know the bit depth. As the hdr file is always there and states the bit-depth of the mib file I felt this was simpler. I'm happy to try to get it working without using the hdr files if this is preferable.

[magnunor]

I do it without the header file in my simple reader in pixStem, feel free to use that code: https://gitlab.com/pixstem/pixstem/blob/master/pixstem/io_tools.py#L14

See https://gitlab.com/pixstem/pixstem/blob/master/pixstem/io_tools.py#L76 for how to grab just the header part in the binary file.

It might also be nice to be able to specify what the probe dimensions should be, and removal of things like flyback pixels. This can be done as optional parameters. Possible implementation of that: https://gitlab.com/pixstem/pixstem/blob/master/pixstem/io_tools.py#L85

No worries if you don't have time to do all these things in this pull request. The basic functionality can be added first, and subsequent pull requests can expand upon the functionality later.

[sem-geologist]

It is not review, just want to point a few caveats

[sem-geologist]

these two lines.
Please review the dependency, it should not depend from hyperspy (exception is hyperspy.misc.io.utils_readfile as it will go together with io_plugins to splited library).
It would be great idea not to use DictionaryTreeBrowser in io library for metadata tree building. It is better to expose all metadata as original_metadata and use mapping dictionary to build proper hyperspy tree.

[sem-geologist]

all these ifs could be simplified with using mapping

[dnjohnstone]

I think this is now out of scope until after the split of hyperspy - can we close it and re-visit after a split?

[chrisallen80]

Agreed, closing.