Support pre-FS original_metadata.txt

[jamoore]

See #10349

The implementation of the OriginalMetadataRequest? currently silently ignores pre-FS data, since there is not yet a database upgrade for properly testing that part of the scenario. Post-demo 3, the implementation for parsing the INI-like file should be added to handle the file annotations of the form:

[GlobalMetadata]
IlluminationChannel #2 Acquire=true
IlluminationChannel #1 Acquire=true
DataChannel #3 Acquire=true
DataChannel #1 Acquire=true
...
[SeriesMetadata]
Track #3 Sampling Method=1
DetectionChannel #1 SPI Wavelength Start=415.0
Track #3 Bleach Scan Number=0
...

At the same time, the OriginalMetadataRequestTest? should be reviewed, since getting it working during demo 3 was also difficult. (Note: it might be necessary to re-add the generation of the original_metadata.txt file to the code base, even if just under test/, since it was removed from OMEROMetadataStoreClient on develop).

[jmoore <josh@…>]

(In [ccf7b8b1ef871d54febb5f04dd696c9d25348937/ome.git] on branch develop) Outline for handling pre-FS data (See #10703)

[jburel]

Client might have to put the code back to handle pre-FS data.
This has to be decided soon

[jburel]

Moving out of Demo 4.x since we will not handle a mix of data for now

[jamoore]

Mark, would you be ok to look at this ASAP? (i.e. before other DB changes)

[mtbcarroll]

Looks like the FS version comes out with a Hashtable<String, Object> for each of global and series. OriginalMetadataRequestI.loadFileAnnotation gets a file annotation ID but how to parse the corresponding INI-style file into such I've yet to determine; for instance, given the String table key, do we thus already know which datatype we expect after the "=" for filling in the Object value, so we can tell a false Boolean from a "false" String? And, how to fit the INI "[…]" section names into this? No doubt it'll all become clear next week.

[mtbcarroll]

Oh, it looks like the [Global metadata] or [Series metadata] INI section names are about deciding on global or series, makes sense. (Have to double-check the spacing, capitalization.)

[mtbcarroll]

Apparently the Map<String, RType> maps may always receive RString values.

Given how INI-like the format looks, probably best to parse it as exactly INI.

[mtbcarroll]

Okay, so the description has [GlobalMetadata], OriginalMetadataRequestI has [Global metadata], webclient/views.py has [Global Metadata]. If I try downloading from howe I get [GlobalMetadata].

[jamoore]

I would vote for using the value from the pre-FS property files across the board. Melissa, Jean-Marie, Will?

[wmoore]

OK - let me know if you need any changes in web etc.

[jburel]

Yes, same value.

[mtbcarroll]

From OriginalMetadataRequestI I wonder what the easiest way is to get access to AbstractFileSystem.getFilesPath so that it can actually read the original metadata.

[jamoore]

You'll need to also inject the OriginalFilesService along with the PixelsService.

[mtbcarroll]

Fixed by ​https://github.com/openmicroscopy/openmicroscopy/pull/1717.

(I'm not sure what's going on with that test class, but after some poking it looks like quite a separate effort.)

[jamoore]

Which test class?

[mtbcarroll]

OriginalMetadataRequestTest

[jamoore]

If you mean how it works, happy to discuss. If you mean how it is supposed to have OMERO4 data present, that's a big(gish) issue. Most likely we'll have to have MockFactories for producing both types.

[mtbcarroll]

Well, it doesn't seem to work right now. (-: For attaching OMERO4 data, I guess that we need a general test framework capability to get non-FS data into an FS server (I don't know to what extent that is more than just "not in a fileset"), after that it's just a matter of attaching a file annotation that's in the right namespace. Probably worth some kind of a ticket anyway, or adding to an existing one.

[Josh Moore <josh@…>]

(In [a6df36dc09a77cbe8c336559585372285fb1f309/ome.git] on branch develop) Merge pull request #1717 from mtbc/trac-10703-original-metadata

fix #10703: download pre-FS original metadata in FS