Activity for Mark Greenwood
-
Mark Greenwood modified ticket #9
Warnings mechanism
-
Many of the PRs now have proper logging in them. It's not common across all PRs or in their interface etc. but I think it means this issue is essentially covered, so to clean up the list here (new issues are on github) I'm going to close this and mark it as done
-
Importong/Exporting of Corpus as a zip file
-
Moved to GitHub see https://github.com/GateNLP/gate-core/issues/48
-
Moved to github https://github.com/GateNLP/gate-core/issues/47
-
Make corpus "save as XML" interruptible
-
Indicate document name in exception output
-
Indicate document name in exception output
-
Moved to gihub https://github.com/GateNLP/gate-core/issues/46
-
debug symbols missing in gate-core
-
Make document viewing and navigation easier
-
True I suppose. You could certainly imagine having a new VR for corpus that is basically the document viewer with the next previous buttons (that might be fairly trivial to implement) but we certainly don't want to group documents under corpora in the resources tree. I've reopened and will add the corpus viewer as a feature request when I move them to github
-
confirmation before removing resources from GUI
-
I can't see how we could possibly implement the second part of this, and if the user has asked to delete a resource they have already done multiple actions, so a confirm dialog would be very annyoing.
-
Better/more detailed (developer) documentation
-
Documentation has been updated since this issue was opened, and also we no longer recommend adding any information inside creole.xml (exception attributes on CREOLE-DIRECTORY and REQUIRES elements). Also full details are available in the developer module slides from FIG if you need more details than are in the userguide.
-
Some of this has been fixed and the rest can't be fixed so I'm closing the issue. The "can't be fixed" relates to the fact that documents aren't children of a corpus. Instead a corpus can reference a document, and a document can be referenced by more than one corpus. As such it's impossible to have a next document button as that could be undefined or have multiple options depending on which corpus you look in. The same for the resource tree, displaying a document under a corpus isn't possible as...
-
Make document viewing and navigation easier
-
mysql and remote datastores
-
It's unlikely we'll add support for this to datastores (we used to have some DB database support but it never worked well), but it might be possible to populate a corpus from a DB, in the same way we do for CSV files etc. As for"PHP syntax for CMS files" not quite sure what this means.
-
you could claim the same for almost any PR, which is one of the reasons we have the section by section processing. I'm still inclined to say "wont-fix" but I'll leave it open for now and move it across to GitHub when I've finished cleaning up the full list of features so thay we can think about it further.
-
Add option to delete user files in uninstaller
-
Not sure it's possible given the way the installer works, as it's not something that's installed or within the installed folder.
-
Annot Transfer to copy annots from within multiple types
-
So as well as not wanting to break the param types, there are already two othe ways you could achieve the same behaviour. You could have multiple AST PRs or you could run the AST inside seciton by section processing instead which I think would give the required flexibility.
-
GATE Integration with Open Semantic Framework & RDF generate using any Unstructured Data & any Social website
-
This is a quesiton rather than a feature request so closing
-
Add logging times to Messages
-
Difficult to do this sensibly with things printed to System.out and System.err. Those places which already use a propper logger do do this correctly though.
-
Reserved/Excluded file names in SerialDataStore
-
This is essentially fixed now, although not in as flexible way as originally envisaged, but fixed enough that I'm closing this to clean up the list of things we might want to work on
-
Changes to Machine Learning PR and Weka Wrapper engine
-
We no longer support the ML plugin, please use the LearningFramework instead
-
RASP parser plugin problems
-
We've dropped support for RASP now as the currently available version won't compile out of the box on modern systems. For wrapping command line tools please see the Tagger Framework plugin
-
Learning plugin: allow to specify nr of maximum nlp features
-
We no longer support the Learning plugin, please use the LearningFramework instead ;)
-
Need a parameter to the ranked label list and their scores, not just a single label (Max label)
-
We've now replaced the Learning plugin with the LearningFramework and don't intended to make any further changes to the old plugin.
-
AnnotationSchema define a single annotation type and are not related to annotation sets in any sense. As they are essentially XML Schema adding extra GATE specific info seems a bad idea.
-
Support specifying AnnotationSet name for new annotations using AnnotationSchema
-
gate-core.jar is now down at just 3.4MB (although it does have a lot of dependencies pulled in via Maven) and the JAPE parser etc. is now in the ANNIE plugin. You still can't use JAPE without gate-core though as JAPE does pattern matching over GATE annotations and features all of which are defined in gate-core
-
Make gate-core.jar modular
-
Use the section by section processing PR instead
-
Allow to restrict JAPE phase to containing annotations
-
This isn't a bug in the PR. If you read the MetaMap documentation (https://metamap.nlm.nih.gov/Docs/README_javaapi.html) you'll see that the two options are vaild. X means truncate candidate mappings and y enables word sense disambiguation. I'm guessing the problem is that y only works if you have also run the WSD MetaMap server as well as the base server.
-
MetaMap default for metaMapOptions runtime parameter is not valid on (at least) metamap 2016
-
Measurement tagger enters endless loop when tries to parse specific text
-
It's definately a bug (although the example isn't a valid measurement it shouldn't cause an infinite loop) which is now being tracked on github https://github.com/GateNLP/gateplugin-Tagger_Measurements/issues/3
-
I've checked and this is still an bug. I've marked as closed-accepted to tidy up the bug tracker as bugs are now tracked alongside the latest dev code on github. The new issue can be found at https://github.com/GateNLP/gateplugin-Tagger_Measurements/issues/2
-
Measurement tagger creates incorrect annnotations
-
TermRaider Error Message
-
I think that the underlying bug (if there was one) has been fixed since the issue was first raised as I can't reproduce the error given the attached document.
-
Mimir results page does not update when no results
-
Not sure when exactly but this was fixed (probably by accident) as you now get a sensible meesage when there are no results.
-
Change JAPE_Plus FSM optimisation routine to use equals()
-
BUG: annotation display incorrect when annotating file with newlines and whitespace
-
Possibly a bug now being tracked at https://github.com/GateNLP/gate-core/issues/38
-
Exception on saving xgapp makes the file empty.
-
Not sure what we could do about this even if I could reproduce it. Essentially saving has started which wipes out the old file and then the unrecoverable exception occurs.
-
Runtime-parameter value field does not re-adjust properly to GUI size
-
bug is now being tracked on github https://github.com/GateNLP/gate-core/issues/25
-
Duplicate plugin paths in persisted application files
-
with the whole plugin mechanism having changed this is probably irrelevant and I believe even with directory plugins the changes means this is no longer an issue
-
Help->Search In Mailing List does not work
-
Bug is still open bu has been moved to github https://github.com/GateNLP/gate-core/issues/37
-
Deadlock in gate application
-
Batch Learning: Review of postProcessingChunk
-
I'm closing this bug as it's old and in the interim we've switched to using the LearningFrameowrk plugin instead
-
NullPointerException in Batch Learning (evaluation mode)
-
I'm closing this bug as it's old and in the interim we've switched to using the LearningFrameowrk plugin instead
-
Space is automatically added in Inline xml files.
-
This is expected behavour and can be controlled through the add white space on unpack option.
-
removing random lib folder
-
revert the pom here so that I don't break the omtd build :(
-
fix the parent pom
-
a romanian flag instead of the ANNIE icon
-
mavenized
-
tieded up both gapp files
-
fixed the main gapp file
-
started to convert to maven
-
removing this for now
-
use the right creole.xml file
-
mavenized ready to move to git
-
moved the ontology editor into the plugin.... I think
-
moved the ontology editor into the plugin.... I think
-
Rethink creation of $relpath$ and related paths
-
changing the artifact name before moving to github
-
changed the artifact id prior to moving code to...
-
saving options and settings now works again whe...
-
bumped jackson to the latest version to avoid a...
-
turns out we need compile and runtime not just ...
-
ensure full runtime dependency resolution is av...
-
force the base-plugin to always be resolved fro...
-
more notes
-
empty plugin generated from this now compiles
-
some cleaning up
-
initial stab at an archetype to make creating n...
-
done some of the metadata conversion
-
some reordering and adding an extra plugin
1 >
