FITS loader - Changing handling of header keys

[John Hill]

The FITS loader uses fixed header values which will change as the usage requirements for the IMAT instruments becomes more defined. Some of the keys can currently be overridden from the load dialogue, or from a selected map file.

When header keys are finalised it may be desirable to remove the load dialogue properties for overriding them, specifically in the case of BITPIX and NAXIS which should be universal (defined in the FITS standard and are in theory not subject to change).

Information required for use in tomography are: ImageKey (key being the type of image it is open/dark field or sample) Rotation (current sample rotation when image was taken) Intensity (beam intensity at the time the image was taken)

Ideally this information will be available from the FITS header, but actual key names to look for are unknown at this time. And the defaults currently specified may need changing.

[Federico Montesino Pouzols]

mark some required changes/additions, re #10231

Changeset: 112cdd71f4eab23d51c86c64abf2898ef3d02d0e

[Federico Montesino Pouzols]

use '_FITS:' prefix for log entries from the FITS header, re #10231

Changeset: 111b7fb4726cf99d0c7a0f7d9daf3398d8cef6a4

[Federico Montesino Pouzols]

make sure FITS entries are not added twice, re #10231

Changeset: 32bfecd7f354a35c59991fcf110effcf88b15ac0

[Federico Montesino Pouzols]

slight speed up, change header process closer to IMAP FITS, re #10231

Changeset: c51776eaf7634221ae13969b32d2e9e281f494d7

[Federico Montesino Pouzols]

old input properties finally gone, re #10231

Changeset: 95870913ac6ab241ee6e05ef4f32e9e1f1ea84f5

[Federico Montesino Pouzols]

Drop _ prefix for logs, better property help, re #10231

Changeset: 0b79383b4ee2e85e285f13e3edcadb12bd9d4c4b

[Federico Montesino Pouzols]

develop test, check basics and FITS headers, re #10231

Changeset: 7cbf9cf30b5611d12f8119bd30b307bcc5dde229

[Federico Montesino Pouzols]

revamp, better error checks/msgs, new headers, re #10231

Changeset: 1731035b8430f4088740a7fff478c439e447d9d8

[Federico Montesino Pouzols]

This is being verified as ​pull request #482.

[Federico Montesino Pouzols]

use fstream rather than FILE - so avoid leaks, re #10231

Changeset: 1b268b546f97047569904079cf890ef9812a0b8a

[Federico Montesino Pouzols]

It seems that the system test MERLINReduction is failing on auriga because of a missing/broken file MER11060.raw. Jenkins, retest this please

[Federico Montesino Pouzols]

Jenkins, retest this please

[Federico Montesino Pouzols]

Jenkins, retest this please

[Federico Montesino Pouzols]

Jenkins, retest this please

[Federico Montesino Pouzols]

Jenkins, retest this please

[Federico Montesino Pouzols]

Jenkins, retest this please

[Federico Montesino Pouzols]

Jenkins, I'm begging you... retest this please

[Federico Montesino Pouzols]

Jenkins retest this please

[Federico Montesino Pouzols]

Jenkins retest this please

[Karl Palmen]

Only Filename & OutputWorkspace should remain on the LoadFITS dialog box, when I test. Please let me know if I'm wrong.

[Federico Montesino Pouzols]

There is one more (optional) field with the name of a "headers map" file. This is still there, as we will need to play with non-standard and changeable headers for a while. Not sure at this point if it will stay there or die when the tomography code stabilizes. It doesn't hurt anyway, and most users will never need to care about it.

[Federico Montesino Pouzols]

Oh, and also note that there is another PR with documentation for this algorithm: #471.

[Federico Montesino Pouzols]

Both are open at the moment, and there could be merge conflicts (they shouldn't be really important). Let me know if something happens.

[Karl Palmen]

When I test it, the LoadFITS algorithm is not available for selection and so I can't access the dialog box.

[Federico Montesino Pouzols]

There's a merge conflict after another PR was merged. So probably in your test branch you didn't get the new code or an inconsistent version. I'll sort it out and let you know.

[Federico Montesino Pouzols]

Merge remote-tracking branch 'origin/master' into feature/10231_FITS_loader_change_and_add_header_keys

Conflicts:

Code/Mantid/Framework/DataHandling/src/LoadFITS.cpp

Sort out merge conflict with another PR related to LoadFITS, re #10231

Changeset: 310a48a0c150deedbcaeea1a2a768a065cda7553

[Federico Montesino Pouzols]

I just fixed the merge conflict. Unless there's any unexpected issue this should be ready for testing again when the CI builds finish (crossing all fingers and toes).

[Federico Montesino Pouzols]

Jenkins, retest this please

[Federico Montesino Pouzols]

_ prefix not used any longer for FITS headers, re #10231

Changeset: 941b74f3bbc4e1d5342c8ffe8410290f056cc1bf

[Federico Montesino Pouzols]

Oops, the builds failed on osx and rhel6 for unrelated random CI issues, but there's actually an issue in the LoadFITS doc test where I forgot to update after the merger of another branch. It's now or never...

[Karl Palmen]

Merge pull request #482 from mantidproject/feature/10231_FITS_loader_change_and_add_header_keys

LoadFITS: change and add header keys, and other improvements

Full changeset: ​2217b576871ce96f4f465f371b97f690fd94f2fc

[Stuart Campbell]

This ticket has been transferred to github issue ​11073