Add Tecan Spark Cyto reader

[melissalinkert]

Backported from a private PR. The new reader adds support for data from this acquisition system: https://lifesciences.tecan.com/plate-reader-live-cell-imaging-cytometry

Plates consist of a sqlite database file, a directory hierarchy containing TIFFs, and optional analysis results in various formats. Only the sqlite database can be accepted by setId. No test data yet, but opening the PR now to verify that no existing tests are affected.

As this is a new reader with a new dependency, a minor release will be necessary.

[sbesson]

Excluding as this failed the nightly repository build. In particular this affects filesets containing Thumbs.db e.g.

   [testng] The factory method class loci.tests.testng.FormatReaderTestFactory.createInstances() threw an exception
   [testng] [2020-11-20 00:43:49,859] [main] setId("/data/karen/Thumbs.db") failed
   [testng] java.io.IOException: Cannot find expected 'Images' directory
   [testng] 	at loci.formats.in.TecanReader.initFile(TecanReader.java:234) ~[formats-gpl-6.6.0-SNAPSHOT.jar:6.6.0-SNAPSHOT]
   [testng] 	at loci.formats.FormatReader.setId(FormatReader.java:1421) ~[formats-api-6.6.0-SNAPSHOT.jar:6.6.0-SNAPSHOT]
   [testng] 	at loci.formats.ImageReader.setId(ImageReader.java:849) ~[formats-api-6.6.0-SNAPSHOT.jar:6.6.0-SNAPSHOT]
...

[melissalinkert]

a730fe7 should fix the build failures, but I'll re-exclude if the next build that includes this PR is still failing.

[melissalinkert]

Two test plates are now in inbox/tecan/, see configuration in https://github.com/openmicroscopy/data_repo_config/pull/480

[sbesson]

Imported the sample plates into a server with this reader included without any issue

All the metadata populated by the reader is present (pixel sizes, channel names, acquisition date, exposure times). And the binary data reads as expected. Only puzzlement from my side is why the image names are named after the DB filename rather than the plate name. I think this is related to the behavior discussed ome/design#57 but we had updated the code to trust the Plate.Name metadata set by reader.

I'll leave David to take a look at the source code.

[dgault]

Source code for the new reader looks good and manual testing of the 2 sample plates also shows no issues and both pixel and metadata look to be read and displayed correctly. All unit tests and repo tests continue to pass with this new reader and the associated config included. Merging for inclusion in 6.6.0