Simulating Q space data from crystal info.

[samueljackson92]

I've created a first attempt at this. I used SXD data because it's smaller and therefore easier to work with at this early stage. The sample use in this demo is NaCl which can be directly compared to the sample in SXD23767.

This first version can generate gaussian peaks in HKL space and crop the space to the instrument. Below is the output compared to some real SXD data:

In the next steps we will need to:

• Modify FakeMDEventData to be able to apply a skew and rotate peaks towards the sample
  • It would also be useful to be able to create multiple peaks at once to prevent slow down from many, many threads.
• Modify FakeMDEventData to provide a background that drops off with distance from the sample
• Add support for simulating in Q space rather than HKL.
• Adjust data to have correct intensity distribution, especially for peaks.
• Refactor & generalise code.

[samueljackson92]

I've added an update with some further improvements to this:

• Includes the debye-waller factor for the background
• Now uses Gaussian instead of uniform background
• Massively sped up the script
• Now relying on numpy to generate peaks instead of Mantid both for speed and because I can in future adjust the size/shape of the peaks
• Now supports reducing multiple CIF files and outputting the MD data, a peaks workspace and a UB file.
• All MD data is now in Q instead of HKL. This means the mask only needs to be generated once.

[samueljackson92]

Things still to do with this:

• [x] Add command line tools for the output directory and input files
• [x] Make a system for config files.
  • a python file with import config?
  • a parameter when you pass a .json file?
• [x] Fix bug with not finding all peaks. (mask dilation)
• [x] Fix underscore in simulation output suffix
• [x] Fix imports for running via command line
• [ ] Better intensity simulation
• [x] Separate mask generation for simulation
  • I just want to supply a pre-generated mask file instead.
• [ ] Option to pass a UB matrix with each sample