test_images_metadata with ome.xml?

[wmoore]

Discussed briefly with Melissa:

• Need a small git repository of ome.xml for our handful of test images
• On our local machines we can include the PFFs themselves in the hierarchy, but tell git to ignore them (not committed to repo).
• Melissa has a way to compare the DOM tree and element attributes of ome.xml for validation.

Need to:

• Set up a git repository where we can start to put ome.xml
• Then I can continue to look for bugs in ome.xml generated from PFF.
• NB: It would also be nice if we could turn off the Structured Annotations in ome.xml generation. This makes the files 10 times bigger and makes working with them & comparing them much harder.

[jmoore]

In discussions with Melissa & Scott (separately), it came out that unless we use symlinking, the use of separate directories is going to start to make repository VERY difficult to use. I would argue for cleaning the directories up globally using something like:

.
├── SVSReader
│   └── some-project
└── TiffReader
    ├── 2010-martha
    └── 2011-george

i.e. something relatively flat, top-level something like the reader name in question, sub-directories something unique and memorable.

Then we could go about extending the metadata in the .bioformats files which we could then script with python to come up with lists like:

 ./bfrepo find "images_metadata"

(or whatever).

[jmoore]

Moved from sprint 2011-11-29 (3)

[jburel]

Moved from sprint 2011-12-13 (4)

[jmoore]

Moved from sprint 2012-01-31 (7)

[wmoore]

I have started working on liff/fabrice/Rosier.liff and lei/brian/tropo.lei committing to a new git repo on my machine. But I will hold off now since I may end up cloning from Melissa's end?

[wmoore]

Josh's instructions: create a git repo on necromancer. Push stuff there. Can move to a different location as needed.

[wmoore]

I have created a git repoistory on necromancer and have pushed Rosier.liff to master.

jrs-macbookpro-25107:metadata will$ git remote -v
origin	ssh://wmoore@git.openmicroscopy.org/home/git/metadata.git (fetch)
origin	ssh://wmoore@git.openmicroscopy.org/home/git/metadata.git (push)

There are a couple of commits there now

* b2fa0bd (HEAD, origin/master, master) Adding LightPath for Rosier.liff 'Fluo exo/Image007'
* dadc4f1 Corrected Offset and Gain for Rosier.liff. See #7897
* 3072dc3 ome xml generated from BF branch sprint8-bug-fixes

I need to finish going through the Rosier metadata, add LightPath? for other images etc.

Melissa, do you want to clone this repo and see how it looks? See what you get when you test with various branches of BF?

My first impressions of using this private repo are that it's a pain! No PRs, public links etc. I think we're going to have to move to github. We just need to use PFFs for which we have permission.

I'd like to start by asking Martin if he's OK with the ome.xml from his files being public. I'm sure he'd be OK with this. Then we'll have Lei, Liff and Lsm all covered.

[mlinkert]

I had a look at this in the context of #7897, and it seems nice. No, there aren't pull requests, but it's really great to be able to see what the correct output should be (without having to dig through tickets, email, etc. to find what you've told me in the past).

I'll need to fix up the OME-XML graph comparison tests so that it's possible to do something like:

ant -Ddata.directory=/path/to/data/files/ -Domexml.directory=/path/to/metadata.git test-ome-xml

but otherwise this seems usable to me so far.

[jburel]

Moved from sprint 2012-02-28 (9)

[wmoore]

Put publicly available ome.xml here under specification/ on branch: ​https://github.com/will-moore/openmicroscopy/tree/omexml_repo

Closing...