import neo4j RM changes

[Polber]

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 42.18%. Comparing base (e889e89) to head (a74bcb1). Report is 208 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #1529      +/-   ##
============================================
- Coverage     42.19%   42.18%   -0.01%     
+ Complexity     3160     3158       -2     
============================================
  Files           790      790              
  Lines         45981    45981              
  Branches       4920     4920              
============================================
- Hits          19400    19396       -4     
- Misses        25003    25006       +3     
- Partials       1578     1579       +1     

Components	Coverage Δ
spanner-templates	63.32% <ø> (-0.01%)	:arrow_down:
spanner-import-export	64.42% <ø> (-0.03%)	:arrow_down:
spanner-live-forward-migration	74.34% <ø> (ø)
spanner-live-reverse-replication	50.83% <ø> (ø)
spanner-bulk-migration	82.59% <ø> (ø)

see 2 files with indirect coverage changes

[fbiville]

Any updates on this?

[Polber]

@fbiville Running the IT's with the imported resource manager still fails... not sure if there is something I am missing, but it may not be worth investing too much time into this PR since Beam 2.57 will be released in the next couple weeks

[fbiville]

@Polber I realize now that I implement the API for the database wait option completely backwards 🤦 https://github.com/Polber/DataflowTemplates/pull/20 attempts to fix that.

[fbiville]

The CI failure seems unrelated.

[Polber]

@fbiville apologies for the delay. I went ahead and refactored to use the RM from Beam 2.57 since the repo has now been updated to that version. Let me know if the changes look good (tests are passing) and I can get this merged.

[fbiville]

@Polber unfortunately, the fixes I pushed earlier are still needed... the Neo4jResourceManager in Beam uses the wait options for static databases where it should have been for dynamic databases (and this is completely on me, I really have implemented the opposite of what I should have) If you can reinstate the changes in the PR I sent to your fork, that'd be great. I can also recreate the PR from the current state if that's easier for you