allow user to specify key-value pairs. Insert these into DO.metadata. do nothing with them, but they will be written into the header of the output chi files.

90% of tests will pass without the --is-test flag set, but for consistency, I suggest we add this to all the tests. It will create less confusion in the future .

specify an inputs directory name. Default is cwd.

use diffpy.utils as a template

user can specify wavelength. If it is specified have it overload the wavelength pulled from the WAVELENGTHS set

The functionality works like this: first it looks for a file, then it looks for a directory and glob all files in it. It also accepts wild card and a file list (default file may be like input_file_list that contains a list of things to run on).

Examples:
lpd 2.5 jins_data.chi
lpd 2.5 ./input_dir/jins_data.chi
lpd 2.5 .
lpd 2.5 ./input_dir
lpd 2.5 .chi
lpd 2.5 input_dir/.chi
lpd 2.5 file1.chi file10.chi
lpd 2.5 input_file_list.txt

file1.chi 2.5
file2.chi 3.5 (multiple files with different muD)
-file7.chi (specify files don’t want to read)

implement x-axes other than tth grid

specify an output directory name. Expected behavior: code checks the dir is there, creates it if it isn't then and writes corrected files in there. Files exists/force behavior updated for this eventuality. Default could be "_corrected" or sthg.

load the username into the metadata (try and get it from system but allow user to overload that?)

expected behavior:
user specifies sample composition. Given known wavelength the program computes a best effort muD. at this point make muD an optional argument. if muD and composition are given muD = args.mud, but composition is written to metadata.

If possible reuse other code for parsing/handling composition. We are doing this in getx3 already. This cannot be a dependency because it is proprietary, so maybe pull that code out into diffpy.utils?

use regolith helper_gui as a template, though this implementation is more complicated than we will need here.

Not needed for the next release

tests for everything....except the argparsing.

• For lists and globs, the best behavior would be to read as many files as possible and skip over the others, letting the user know which have been skipped.
• We would like to have tests for all the same sets of inputs that were used in #48 and #55 but that check that the right correced files are written to the right places (so glob the directories after the corrections are applied and check the list is as expected).
• Also, let's make sure that the args that are written to the file headers are correct, where the parent file is correctly identified (i.e., not just the input file list saved to the header, but also the actual file that was used for that particular correction. This will have to be set inside the loop).

UC 1: single file uc.

1. Jin has measured an x-ray pattern and wants to apply the absorption correction.
2. Jin navigates to the folder with the file in it
3. Jin runs pdfapp giving the name of the file
4. pdfapp computes the cve, applies it to the datafile and writes out a new data file

UC 2: single file, remote directory

1. as UC1.1
2. Jin cd's to his home directory
3. as UC1.3-1.4 except he also specifies the location of the file as well as the name

check if args.input_file is specified. If not, glob the inputs directory for all readable files and correct them. skip non-readable (by loadData (or DO.load when it is written). Write outputs to output directory check if files are there or not and whether force is set. Only compute cve one time and then apply it to all the files in teh glob list.

I added some functionality in the parser to make the app more flexible to use. These need to be implemented:

1. specify an output directory name. Expected behavior: code checks the dir is there, creates it if it isn't then and writes corrected files in there. Files exists/force behavior updated for this eventuality. Default could be "_corrected" or sthg.
2. specify an inputs directory name. Default is cwd.
3. check if args.input_file is specified. If not, glob the inputs directory for all readable files and correct them. skip non-readable (by loadData (or DO.load when it is written). Write outputs to output directory check if files are there or not and whether force is set. Only compute cve one time and then apply it to all the files in teh glob list.
4. user can specify wavelength. If it is specified have it overload the wavelength pulled from the WAVELENGTHS set
5. allow user to specify key-value pairs. Insert these into DO.metadata. do nothing with them, but they will be written into the header of the output chi files.
6. load all the input parameters into metada.
7. load the name of the package and the version number into metadata
8. load the datetime of the analysis into metadata
9. load the username into the metadata (try and get it from system but allow user to overload that?)
10. as above but email of user