2014-11-25 Tuesday Team Meeting
Remember: Agenda must be complete (with estimated times) on the day before the meeting. Any additions after that must go at the bottom (AOCB)
Dundee: Will, Kenny, Petr, Emil, Chris, June, JM, Dominik, Helen, Graeme, Colin, Simon, Mark, Andrew, Jason, Balaji, Remote: Seb B, Josh, Gus, Kelli, Josh B, Ian, Douglas
Agenda - 2:30pm Start
Accepting minutes from last meeting
Project Status (15 minutes max)
Mainline (5 minutes)
5.1.0-m2
units (Petr)
Mostly all done. Bio-Formats issue just fixed by JM.
Other PRs won’t be in m2 (E.g. 5.0.6 rebase).
Blockers all fixed.
Seb: artifacts built and tagged.
Need to decide if we’re happy with tag.
JM: few items remaining on trello board, but these can go to m3.
Josh: m2 tag allows us to get all builds green again.
Hopefully keep everything green during future work.
- tagged build, testing
5.1.0
- Mark: reworking generally working pretty well. Couple of bugs - should be fixed & PR soon.
-
graphs reimplementation Trac #11779 (Mark)
- integration tests pass
- several broken integration tests fixed and a few tests added
- several bug tickets fixed
-
performance improved
- could be better, see Trac #11084 (not graphs, but related)
-
discovered yesterday that new chgrp has bug, Trac #9610
- should be able to fix this week, then open PR for all work
- looking to reproduce Trac #12023, is probably already fixed
- maps: PR for discussion open and in breaking
- conversion of units on DB upgrade (if we decide to do it)
- rendering, timezone, etc. forthcoming
-
Andrew: hope to finish model by end of week.
- Have outline of objects now, just need attributes, compile fixes etc.
- Group sudo (breaking PR open).
- various other internal efforts (necromancer, docs, hiring etc.)
- Josh: probably about 4 weeks behind where we wanted to be wrt model & DB changes.
Glencoe Update (5 minutes)
- Chris: series of PRs to bring back to OME. All based on 5.0
- Import PR opening soon. Can port forward to develop after.
- Other BF work mostly based on develop.
- Jason: no plans to backport everything - but Glencoe work already tested etc.
Consortium Update
- Ian: currently bugfixing FLIMFIT. Not working on 5.1 yet.
- Douglas: jsTree testing feedback - most issues fixed. Waiting on Hibernate fixes for perms projections.
- Not much imaging covered in conference
- Should OMERO be installable by anyone in the lab? Is this a good idea? How to make it easier?
- Mark H. Fiji fixes in progress. Been busy paper writing etc.
- Seb S. been away
-
Dallas update from Sebastien:
- Very cold!
- Lab: biologists, instrumentation team, ‘dry lab’ image processing & analysis pipeline
- dry lab: git training. uTrack bug fixes coming soon.
- 3D tracking, processing etc is next big thing, using their own cluster.
- Not currently using OMERO, but BF is highly used for analysis on cluster etc.
- Fileformat discussions: ND2. 200-300 GB files. Reader not very performant on large files
- Memoiser fixes effectively solves these issues. Open in 1 second.
- Want to use it in ImageJ/Fiji.
- OME-tiff generated by custom LightSheet microscope.
- Subcontract acquisition to Labview software -> OME-tif.
- 1 tiff per z-stack. Bugs in headers.
- Our sample files are hard to find.
- Needs to be easier to find specific examples.
- Will soon hit very large files (TB)
Bio-Formats - next steps, development planning
- Trying to understand how much BF effort is needed
- Tricky to see the status, quite cumbersome etc.
- Decoupling: independent B-F and OMERO releases possible?
- 5.1 timeline (early spring) - need to limit coding accordingly
- Need to improve internal work-flow, docs etc. E.g. for new hires.
-
Chris - Glencoe work has been increasingly B-F focussed.
- Highly inefficient at the moment and difficult for open-source team to get involved.
- Need to share knowledge, workflow use-cases etc with OME team.
- Complex use-cases from digital pathology etc. Metadata critical. E.g. Zoom
- Real world use requires better performance than we typically test with.
-
Jason: - larger datasets, performance critical
- E.g. ND2 or czi are complex formats with heterogenous internal structure
- Sheer scale of BF usage, including OMERO and other applications
- All of this is too much for Melissa - needs more of team involved.
- Data Model doesn’t cover a lot of metadata from various formats, E.g. DICOM.
-
Melissa - big tasks - can’t do all at once. Preference:
- Docs & samples
- Decoupling BF & OMERO releases
- IO / memoizer
-
Chris: 1 - 2 months per component (3 - 4 people).
- Remove submodule, update build system, testing fixes, model decoupled.
- Possibly stop adding file formats between now and 5.1 release?
- JM: scope docs work and gather info needed by 17th Dec workshop (after meeting).
- Melissa: this may not leave us enough time before 5.1 release (3-4 wks in Jan).
- Chris: fixing BF bugs is still the priority.
- Chris will add more detail to trello card on Decoupling
AOB (5 mins max - technical discussions should be highlighted to relevant people and rescheduled)
-
Can we manage a docs afternoon before xmas holidays? If so, possible dates? Friday afternoon with drinks and snacks?
- Helen: maybe too busy before Christmas?
- Josh: Not necessarily! Depends on how many people we need.
- JM: Maybe after Tues 17th meeting?
-
Can we manage a docs afternoon before xmas holidays? If so, possible dates? Friday afternoon with drinks and snacks?