Base knife-edge scan analysis implementation

Hi @lleguy @mercadil,

This is an example of a base implementation, specifically the knife edge scan analysis from Laurent's tb.knife_edge. This is agnostic from xarray data format, thus this can be easily used on other means, such as on a Karabo device.

I also have found out that there could be problems with using bisect especially with repetitive elements. I use slicing instead to find the valid range.

added 1 commit

• 1552b7b7 - Rename methods, fix and test finite_array

Compare with previous version

added 1 commit

• 41fcdc78 - Better finite array retrieval

Compare with previous version

resolved all discussions

I am not sure whether this is the right place to do general ideas, because you are just cleaning up existing code, and my points are actually on the original code:

while I like your erfc generator, we do not actually need the two different functions. The parameter a is just positive or negative, and once we are done fitting, we take the absolute value. Sure that should be tested, but now we have tests, init?

Ha! I knew I was doing something silly with this approach. You are right, and this means there is no need to estimate the slope anymore --> this can also be removed and save some computing time.

TIL :D I attempted to test the erfc without the estimate, but I am not sure if I went for the right approach.. would be happy for any suggestions.

added 2 commits

• eb35f355 - Support unknown signal dims
• 796f088f - Try adding more tests

Compare with previous version

added 1 commit

• df6c4b81 - Correct width absolute value when flipped

Compare with previous version

I think your test is fine.

As a purist, I would actually prefer if we used the erf function instead of the erfc function, why should we use the complementary function if the original function is as good? But this is just purist's nitpicking.

added 2 commits

• b98f8523 - Add generic fit function, generalize prepare_arrays()
• 63107bdb - Add generic fit function, generalize prepare_arrays()

Compare with previous version

added 2 commits

• b98f8523 - Add generic fit function, generalize prepare_arrays()
• 63107bdb - Add generic fit function, generalize prepare_arrays()

Compare with previous version

added 1 commit

• 22ec09a8 - Add docstrings

Compare with previous version

added 20 commits

• 22ec09a8...4c1e5e5f - 2 commits from branch master
• 7218b61e - Refactor detectors module importing
• 4889de70 - Refactor misc module importing
• 739ac0c2 - Refactor routines module importing
• 9a8291a1 - Refactor top-level module importing
• 67b84d7f - Use absolute import on module alias
• 4e1b1c1e - List complete required package
• f3a52c4a - Use extra-requires to sort libraries according to usage
• b86aaffa - Update docs and remove some detector modules from init
• b480b8f1 - Add PES detector
• d22f5551 - Base knife-edge scan analysis implementation
• 79cf3f72 - Rename methods, fix and test finite_array
• ba43eb53 - Better finite array retrieval
• aff1d06d - Support unknown signal dims
• 5031350b - Try adding more tests
• 3c4d3669 - Correct width absolute value when flipped
• 42b9876c - Add generic fit function, generalize prepare_arrays()
• e3189cfa - Add docstrings
• 3911af0c - Add __all__ and adapt __init__ files

Compare with previous version

changed the description

changed target branch from master to refactor-packaging

It's now ready to be used :) (and merge after considering using erf and after merging !106 (merged))

@mercadil, should we integrate it on the existing tb.knife_edge? If so, you can use it with:

from toolbox_scs.base import knife_edge as <alias>

assigned to @mercadil

added 2 commits

• 3911af0c...87b28ade - 2 commits from branch refactor-packaging

Compare with previous version

changed target branch from refactor-packaging to master

added 8 commits

• 87b28ade...023a300b - 7 commits from branch master
• 21474d27 - Solve merge conflict

Compare with previous version

@carinanc I realised that range_mask will return an empty array if minimum == maximum which will give an error later on. Shouldn't the error be raised inside range_mask with a more explicit message instead?

added 14 commits

• 2002897e - Refactor routines module importing
• 6c158b05 - Refactor top-level module importing
• 69f4f353 - Use absolute import on module alias
• 9b67d0ca - Base knife-edge scan analysis implementation
• ffd845fb - Rename methods, fix and test finite_array
• 149312f5 - Better finite array retrieval
• c92205fa - Support unknown signal dims
• eb356efd - Try adding more tests
• d807ecd9 - Correct width absolute value when flipped
• 22cfe1fa - Add generic fit function, generalize prepare_arrays()
• c62cd2cb - Add docstrings
• cae94825 - Add __all__ and adapt __init__ files
• 8197fdd4 - Use knife edge base in routine function
• 3e7d8775 - Catch invalid range exception

Compare with previous version

mentioned in commit 03569509

merged