PR #24 should fix the problem with the test_element_xray_transitions unit test. I also created issue #23 to allow values from several references to be returned.

I've committed the changes. Will the pull request be automatically updated?

I will wait for PR #24 to go through and then merge your PR.

I think the fix in PR #24 was to get only distinct transitions. The test which is failing with my changes now might continue to fail for two of the cases (Z=3 and Z=6) after PR #24 . It is because "elam2002" and "perkins1991" have transition probability data for different number of unique transitions for some elements. Expected values in the test have to updated I think.

Also on a side note, I calculated transition probability for "elam2002" by multiplying the fluorescence yield and the relative intensity data provided for each family of lines (transition lines resulting from a hole in one particular shell). The relaive intensities add up to 1 for each family of lines. From what I have researched, fluorescence yield * relative intensity should be the same as radiative transition probability. Please let me know if you think otherwise.

For "chantler2005", I'm using shelldata.csv file available here https://github.com/openmicroanalysis/FFAST.jl/blob/master/data/shelldata.csv . I just figured I could stream the data directly from the website. I couldn't find a way to access the data directly from the NIST website. Should I do any changes? And should any liscensing data be included?

I think the fix in PR #24 was to get only distinct transitions. The test which is failing with my changes now might continue to fail for two of the cases (Z=3 and Z=6) after PR #24 . It is because "elam2002" and "perkins1991" have transition probability data for different number of unique transitions for some elements. Expected values in the test have to updated I think.

Yes that's correct. The tests need to be updated, but at least now the values are correct.

Also on a side note, I calculated transition probability for "elam2002" by multiplying the fluorescence yield and the relative intensity data provided for each family of lines (transition lines resulting from a hole in one particular shell). The relaive intensities add up to 1 for each family of lines. From what I have researched, fluorescence yield * relative intensity should be the same as radiative transition probability. Please let me know if you think otherwise.

I think the transition probability should include the Coster-Kronig transitions. Is the fluorescence yield discussed in Elam (2002) reference?

For "chantler2005", I'm using shelldata.csv file available here https://github.com/openmicroanalysis/FFAST.jl/blob/master/data/shelldata.csv . I just figured I could stream the data directly from the website. I couldn't find a way to access the data directly from the NIST website.

Have you look at this Julia script in FFAST.jl project?
https://github.com/openmicroanalysis/FFAST.jl/blob/master/support/scrape.jl

Should I do any changes? And should any liscensing data be included?

There is no licence or reference document for all the databases included in pyxray, but that would be a good thing to add. Could you either add a section in the README.rst or simply add a new issue to add references at a later point?

Have you look at this Julia script in FFAST.jl project?
https://github.com/openmicroanalysis/FFAST.jl/blob/master/support/scrape.jl

Thanks, I had not seen that script. I can implement something similar.

Also on a side note, I calculated transition probability for "elam2002" by multiplying the fluorescence yield and the relative intensity data provided for each family of lines (transition lines resulting from a hole in one particular shell). The relaive intensities add up to 1 for each family of lines. From what I have researched, fluorescence yield * relative intensity should be the same as radiative transition probability. Please let me know if you think otherwise.

I think the transition probability should include the Coster-Kronig transitions. Is the fluorescence yield discussed in Elam (2002) reference?

I checked the sources used by the author in Elam 2002. In all of them [1, 2, 3] Coster-Kronig yields are considered separately, i.e. fluorescence yield + Auger yield + Croster-Kronig yield = 1. Conster-Kronig yield covers both radiative and radiationless Coster-Kronig transitions. Both fluorescence and auger yields cover purely inter-shell radiative and radiationless transitions, respectively. So I think my calculation (fluorescence yield * relative intensity) must yield the radiative transition probability for inter-shell transitions correctly.

[1] Krause, M.O., 1979. Atomic radiative and radiationless yields for K and L shells. J. Phys. Chem. Ref. Data 8, 307–327
[2] McGuire, E.J., 1972. Atomic Coster–Kronig, auger, and radiative rates, and fluorescence yields for Ca-Th. Phys. Rev. A 5, 1043–1047.
[3] McGuire, E.J., 1974. Atomic N-shell Coster–Kronig, auger, and radiative rates and fluorescence yields for 384Z4103. Phys. Rev. A 9, 1840–1851.