Bug: MIF chgrp GraphConstraintExceptions

[wmoore]

On today's gretzky build, chgrp on a Dataset containing a complete 3-image MIF gives stack trace below for /home/omero/fresh_import/test_images_good/hamamatsu-vms/openslide/CMU-1/CMU-1-40x - 2010-01-12 13.24.05.vms

STEP ERR stacktrace: ome.services.graphs.GraphConstraintException(message=FileAnnotation:47 improperly linked by 1 objects
at ome.services.graphs.GraphStep.graphValidation(GraphStep.java:407)
at ome.services.chgrp.ChgrpStep.action(ChgrpStep.java:98)
at ome.services.graphs.GraphState.execute(GraphState.java:351)
at omero.cmd.graphs.ChgrpI.step(ChgrpI.java:157)
at omero.cmd.basic.DoAllI$X.step(DoAllI.java:107)
at omero.cmd.basic.DoAllI.step(DoAllI.java:319)
...
, message: , GraphConstraintException: true, step: 36, id: 3

Moving 2 images from a MIF: test_images_good/zeiss-lsm-martin/051215-j-tf.mdb [01-01]

STEP ERR stacktrace: ome.services.graphs.GraphConstraintException(message=Image:1153 improperly links to 1 objects
at ome.services.chgrp.ChgrpValidation.action(ChgrpValidation.java:106)
at ome.services.graphs.GraphState.execute(GraphState.java:351)
at omero.cmd.graphs.ChgrpI.step(ChgrpI.java:157)
at omero.cmd.basic.DoAllI$X.step(DoAllI.java:107)
at omero.cmd.basic.DoAllI.step(DoAllI.java:319)
at omero.cmd.HandleI.steps(HandleI.java:435)
at omero.cmd.HandleI$1.doWork(HandleI.java:365)
at omero.cmd.HandleI$1.doWork(HandleI.java:363)
at sun.reflect.GeneratedMethodAccessor238.invoke(Unknown Source)
...
, message: , GraphConstraintException: true, step: 694, id: 284

[jamoore]

PR opened : ​https://github.com/openmicroscopy/openmicroscopy/pull/1223

[jmoore <josh@…>]

(In [a762f7043aabb16a80c301e649852e6f69327473/ome.git] on branch develop) Activate REAP logic on all tables (Fix #11000)

Since REAP was not being properly passed down the
tree of GraphSteps?, this commit considers any row
that is to be re-checked (i.e. is present in the
reapTableIds map) to be SOFT. In other words, a
failure for a row that will be processed again
later on does not constitute a hard failure.

Note: this is not necessarily strictly valid in
all cases. A row is in 2 separate subtrees, then
really it should be checked if the entire subtree
is to be re-checked. That logic may have to wait
for a revamp of the ome.services.graph package.

Tested via CMU-1 and openmicroscopy/bioformats#536

[jmoore <josh@…>]

(In [79b90c07085dc9fe5066ed20a56d8020b9871e66/ome.git] on branch develop) Passing test in chgrp.py (See #11000)