2008-10-24 Omero Review Meeting
Attending: Jason, Brian, Chris, Jean-Marie, Donald, Ola, Colin, Melissa, Will, Andrew
Agenda
-
OMERO iterations
- Metadata
- Blitz (also: sessions #1114)
- Upcoming...
- Editor strategy for beta 4 (standalone or integrated)
Minutes
-
Meetings
- Every Friday at 2 o'clock Dundee time
-
Iteration
-
Metadata
- We've got it. More in there than can be parsed out
- String "...100x/1.3 uplan apo ix70"
- Mag / aperature / correction / model number
- Should be in all DV log files
- plan apo is a standard
- uplan apo is a derivative (Olypmus lense)
- field: "other"?
-
Data transfer
- Re-architecting for performance
- Previous problem: standard 4 colors --> fixed
-
Current problem: gray scale showing up with three colors
- Leica software interprets that somehow to come up with grayscale
- Data acquired is stored as "color", even if grayscale.
- If channel was "green", stored as green. (For use in Photoshop)
- Bioformats sees "RGB", i.e. multicolor
- Should be one color, and that color should be set to "green"
- Problem is the channel filler? ... should only be getting one channel if its an "indexed color tiff"
- Could be that we're getting wrong indexes?
-
Original files
- Were swamping the file. Really large ones are crashing server
- It's the number of objects which are swamping things.
- There aren't many options while running under JBoss
- Moving forward it will be solved
- Arrange a time to talk to Imperial (using file format which is a pain)
- Tar'ing the files together?
- Note: The issue will rearise with OMERO.fs
-
Creation dates
- Not storing creation dates with images
- Strategy was that we were storing them with the original file
- but with the tar we will only have a single timestamp: first, last, average?
- Ola: Creation date in events? No. But server-side service. * Concept of a fileset / logical file
-
Summary: Do both for Beta3.1
- Zip/tar tons of files in importer
- Discuss the creation date (database upgrade)
- How critical is this?
-
Remaining in metadata iterations
- Currently not reading from database
-
Need to get model loosened up to get things dumped in (next big step)
- objects, detectors, illumination ... mid next week
-
Remaining performance issues
- Interleaving in the bioformats setters
- Everything has to be stored on the client before moving forward
- But to ultimately get the performance, will it always just come down to ftp'ing files?
-
Annotations et al
- Been changing how users interact with the annotations to make it more natural
- Also looking into more natural data entry * TextAnnotation --> Url, Xml, * Moving away from UrlAnnotation * Initial intent was to embed URL. Working. * Making links, i.e. parsing wiki texts (re: editor) * OMERO.wiki, Java parser. * Service to translate links into HTTP:// * Need to define the language
- Moving annotations into the server
- Making tag non-annotation
- Need to standardize the little language early on (for editor & web)
-
Other
-
Graz : wiki linked with OMERO
- Looking into which terms are equal, synonymous
- How to we work toward interoperability?
-
Graz : wiki linked with OMERO
-
Metadata