Activity for Robert Fromont
-
Robert Fromont committed [r4036] on Code
Use new Task.getTasksArray() method instead of Task.getTaskThreadGroup().
-
FrequencyLayerManager 4.69: Ensure 'all scopes' option works for CSV export.
-
nzilbb.annotator.mfa 1.0.0: Add support extract only left/right channel for alignment.
-
Use same shortcut icon specification for the error page as for valid pages.
-
Ensure that Express page titles are XML-escaped.
-
nzilbb.annotator.mor 1.0.3: Ensure that repeated compounds are correctely merged back into the original graph.
-
Don't echo invalid speaker_number parameter back to the user.
-
For Annotator layer generation, selectively include only layers necessary for validation, to decrease memory requirement.
-
Ensure role_permission columns that are compared with other have the correct collation.
-
Ensure role_permission columns that are compared with other have the correct collation.
-
nzilbb.annotator.mor 1.0.2: Avoid conucurrent modification error when moving tags to a neighboring word.
-
Remove protection for skin stylesheet and shiftclick script, so that css/js requests are not the previous request before successful login.
-
nzilbb.annotator.mediapipe 0.2.0: Add suppport for jpg images instead of png.
-
nzilbb.annotator.mediapipe 0.1.9: Use mp4 instead of webm for annotated video, if ffmpeg is available.
-
nzilbb.annotator.mediapipe 0.1.8: Instantaneous annotations for blendshape data and frame images are created using the graph store directly, if available, to reduce memory footprint of annotator.
-
nzilbb.annotator.partitioner 1.1.0: add support for labels to be a serial number rather than size.
-
Update CHANGLOG
-
Fix blank turn/utterance annotations too.
-
Esnure there are no annotations with blank labels.
-
Ensure 'word' column isn't selected for the 'Match' setting by default, instead of / as well as 'MatchId'
-
Remove Digit Span as an option for elicitation tasks.
-
Fix typo in help.
-
Don't open documentation in a new window.
-
Don't overwrite files that were possibly locally updated.
-
nzilbb.annotator.stanfordner 0.1.2: Ensure destroyed automatic annotations don't prevent creation of new annotations.
-
Ensure there are no unaligned phrase layers
-
FrequencyLayerManager 4.68: Ensure predictabilities are always expressed as decimals (not with E-notation).
-
Make default layers on the transcript page configurable with a system attribute.
-
FrequencyLayerManager 4.65: Add support for a maximum inter-word pause before the pause counts as a boundary for word-pair predictabilities/informativities.
-
FrequencyLayerManager 4.53: Tag turn-initial tokens with previous-predictability and turn-final tokens with following-predictability - i.e. include boundaries as a context for predictability. Also ensure word-pair stats are not generated for all corpora if a single corpus is selected.
-
FrequencyLayerManager 4.47: If a specific corpus is selected, only gather stats for that corpus, and don't tag predictabilities/informativities of tokens outside that corpus.
-
FrequencyLayerManager 4.46: ensure progress bar progressess during predictability/informativity computations.
-
FrequencyLayerManager 4.46: Add support for tagging tokens with their previous/following predictability/informativity.
-
FrequencyLayerManager 4.4: Compute previous/following informativity for word-pair layers.
-
FrequencyLayerManager 4.35: add support for computation and export of Previous and Following Predictability - i.e. the conditional probability of one word in a bigram given the other.
-
When generating layers, delete only automatic annotations, leave manual annotations in place.
-
Ensure episode and corpus labels have same limits as other labels.
-
nzilbb.annotator.stanfordner 0.1.1: Improve installation of classifiers at installation time, and doocumentation.
-
When clicking clicking through the license, ensure that all URL parameters of the 'next' page are effectively passed through, not just the first one.
-
Fix name of module.
-
Remove WEB-INF/lib/stanford-parser.jar if it's there.
-
Fix typo
-
nzilbb.annotator.stanfordner 0.1.0: Initial version.
-
nzilbb.annotator.partitioner 1.0.1: Migrate Partition Manager from legacy API
-
nzilbb.annotator.partitioner 1.0.1: Migrate Partition Manager from legacy API
-
nzilbb.annotator.mor 1.0.1: Update English grammar URL, improve matching of MOR output tokens to original input tokens, add support for input tokens that include spaces (so that tokens like "American_Samoa" can be converted to "American Samoa" in an intermediate word layer that is the input for MorTagger).
-
nzilbb.annotator.mor 0.3.2: Add support for top-level span layers for utterance chunking (in addition to phrase layers) - NB only for cases where there are tokens for only one participant per transcript.
-
nzilbb.annotator.orthography 0.2.3: Improve validation and user documentation.
-
nzilbb.annotator.orthography 0.2.2: Implement filter out (i.e. empty pattern) using all("layer").length == 0 during insert, instead of initially including them and then excluding them post-hoc.
-
nzilbb.annotator.orthography 0.1.6: Ensure unticking 'lowerCase' checkbox in task app works.
-
nzilbb.annotator.jython 1.0.6: Upgrade to jython 2.7.4 by default, improve memory management by only creating one engine per script.
-
Improve security of Frequency/LIWC extension pages.
-
Revert to old Praat icon.
-
nzilbb.annotator.flatlexicon 3.1.12: Do no save annotations with blank labels.
-
nzilbb.annotator.flatlexicon 3.1.11: add support for tagging aligned word layers, e.g. MOR stem.
-
nzilbb.annotator.flatlexicon 3.1.10: Ensure `\` can be used as a field delimiter (for CELEX files)
-
Final version of nzilbb.labbcat.server
-
add instructions for retro-fitting form-based auth.
-
jsendpraat 3.3
-
Update CHANGELOG
-
Latest nzilbb... libraries.
-
Ensure the edit/layers/regenerate endpoint works with JDK calls.
-
Remove JNLP utilities links, they no longer work.
-
Don't redirect to task page if it's an DSK request.
-
Use SVG icon for express.
-
Disable unused utilities.
-
Fix home page help link.
-
Fix wordFormColumn parameter processing.
-
Ensure UI loads dash stats from new endpoint.
-
Validate corpus passed into express/browse page.
-
Docker image to use form-based authentication instead of basic authentication.
-
Update Firefos jsendpraat extension to 3.2, to ensure users upgrade native-messaging that supports cookie-based authentication.
-
Remove edit option from help pages.
-
Update automated tests for SDK changes.
-
Remove '(Imported from Transcriber)' from layer descriptions.
-
Add missing ElicitSpeech manifest
-
Add missing ElicitSpeech files.
-
Ensure export includes installed layermanagers.
-
Migrate Praat integration token endpoint to nzilbb.labbcat.server
-
Use /api/keepalive instead of /keepalive
-
Migrate keepalive to API.
-
Migrate license agreement page to Angular.
-
Update jsendpraat to support form-based user auth.
-
Migrate tree to Angular.
-
Migrate seriesFiles to Angular: episode/documents
-
Migrate threads/thread pages to Angular, as tasks/task.
-
Migrate threads/thread pages to Angular, as tasks/task.
-
Migrate dictionary lookup pages to Angular.
-
Deprecate old view endpoints, ready for deletion.
-
Migrate corpus info page to Angular.
-
Use session cookie for Praat integration access.
-
Add option to export only first dictionary entry.
-
Move dictionary lookup endpoints to api.
-
Add API endpoints for utterance suggestion/correction.
-
Replace 'edit word' option with in-situ tag editing and variant listing.
-
Migrate extractIntervals to Angular (annotation/intervals)
-
Mark deprecation correctly.
-
Add dataVersion system attribute.
-
Rename CsvSearchResults as CsvResults.
-
Move generateLayerUtterances to edit/generateLayerUtterances where it belongs.
1 >
