User:Bawolff/GSoC2010/Status
- Code updates: bawolff's checkins
- User:Bawolff/metadata table
- Discussion on wikitech-l about the metadata table
- really ugly blog http://bawolff.blogspot.com
Status
[edit]02:59, 17 August 2010 (UTC)
[edit]Well today is the end of GSOC. I have gotten a lot accomplished over the summer, however at the same time I didn't accomplish all of my goals, due to XMP taking more time than anticipated (however I plan to keep working on this until I do).
Short summary of stuff I've done in my branch:
- Fixup many broken parts of the current Exif support, of particular note the GPS fields. (bugzilla:13172 and many fixes that did not have bugs)
- A jpeg segment extractor to extract the metadata segments of interest.
- Jpeg's can now display file comments (Among other things), which is of interest as mediawiki puts the url of the file there in jpeg thumbnails it serves.
- A class for merging together metadata information from different sources, based on the MWG recommendation.
- Recognize iptc-iim properties. (bugzilla:16956)
- An xmp parser
- Extract XMP segments, and various native metadata from png images
- Extract XMP segments and comment segments from gif
- Support showing multilingual metadata (in formats that support it, specifically XMP and png iTXt chunks)
- Also sort of fixes bugzilla:657 as most forms of embedded copyright data is now shown in the metadata box ( however that bug more sounds like putting a license tag on a page based on copyright data )
Some Stuff I haven't done that I wanted to:
- Re-arrange the way the data is stored in the db. My plan was to do this last since I would know best what type of data i'm dealing with once I've done extracting it, but time ran a little short.
- the xmp support is missing some properties from the iptc4xmp spec.
- I didn't get to doing SVG metadata.
- modify the oggHandler and pdfHandler to display the already extracted, but not displayed metadata.
However I'd definitely like to work on this outside of gsoc, so hopefully I will accomplish my remaining goals. Bawolff 02:59, 17 August 2010 (UTC)
02:42, 10 August 2010 (UTC)
[edit]- Have XMP working for jpg's. However still have to add to the list of properties to extract. Bawolff 02:42, 10 August 2010 (UTC)
00:41, 21 July 2010 (UTC)
[edit]Just a quick update
- Many of the exif tags that were broken (GPS for example) now work (bugzilla:13172)
- Most iptc tags now work. For example, File:Bundesarchiv Bild 183-18379-0003, Erich Hansen, Kirchner, Feist, Oswald Schelter.jpg which on commons has no metadata table, has the following table on my test wiki:
| Image title | Zentralbild Quasch. 16.2.1953 IV. Wintersportmeisterschaften der Deutschen Demokratischen Republik fßr Frieden, Einheit, Demokratie und Sozialismus vom 10.-15.2.1953 in Oberhof UBz: Die Siegermannschaft im Viererbob: Erich Hansen, Kirchner, Feist, Schelter während des Laufes. |
|---|---|
| Author | Quaschinsky, Hans-GĂźnter |
| Date and time of data generation | 16 February 1953 |
| IIM version | 2 |
| Headline | Erich Hansen, Kirchner, Feist, Oswald Schelter |
| Credit/Provider | Bundesarchiv |
| Object name | Bild 183-18379-0003 |
- I'm putting off any potential db changes until after I finish adding extraction for other metadata format.
- I'm mostly done IPTC and plan to start xmp soon.
- My test wiki is behind a dynamic ip. If you're very curious try hunting me down on irc and I'll give you my ip or you could try checking out the img_metadata branch (although there isn't really anything that you can't get from what I describe above, since from the end user point of view its just a table with more entries)
20:26, 7 July 2010
[edit]All code is currently being committed to an img_metadata branch. So far I'm concentrating on fixing up the current Exif support, have added basic IPTC support, and am working on making the metadata changes backwards/forwards compatible. If anyone has any questions about what I'm doing, please don't hesitate to hunt me down on irc (or talk page) and ask. Bawolff 20:26, 7 July 2010 (UTC)