AVAA Changelog

0.69 (upcoming release)

• improved Tests feedbacks in nested documents
• improved doc-import feedbacks with error logs
• added new toolkit cli arguments test-all and test-or-die
• build process now validates all tests before packaging
• documentation generation also runs mkdocs after building
• migrated non-standard XML OP element structure to OPERATION
• new provider.sppas to handle XRA files
• new experimental sppas processors (img_to_cartoon, annotate)

0.68

• improved Python integration with venv and python install manager
• new proc.create-tiers
• new view.tiers
• added CSV provider (experimental)
• proc.export-to-tei: proper encoding, role-map attribute
• Server Mode: admin panel, user management, home, profile
• Server Mode: shared document, document as corpus, user files
• improved MDC logs
• improved feedback payload (annotations preview)

0.67

• added NoScribe provider (HTML version)
• added "lambda" attribute to the select primitive (custom js function filter)
• improved FFmpeg installer for Windows
• YouTube provider can now load comments as annotations with metadata
• view.table: export displayed content to CSV
• proc.export-to-tei: export array of objects to tei table format
• new proc.merge-annotation-files
• new proc.load-image-annotations
• Server Mode: added htpasswd login method
• Remote Corpus: added WebDav downloads

0.66

• new "replace" and "clear" combining modes for stack/unstack/filter/select
• improved linux desktop compatibility with xdg-open and dbus for browsing files
• fixed "Open in VLC" and "Export Clip" for linux in annotations preview
• auto download filters for video-anonymizer and frei0r processors
• python wrapper improvements
• default blank assets are now embedded as resources
• renamed op.set-global-variable to op.save-variable
• proc.hardsub can now work with corpus files directly
• added experimental YouTube provider
• new cache system for generic providers
• new tab session recovers convert-progress after page reload or reconnection

0.65

• editor: "add file to corpus" allows multiple file selection
• editor: fixed random display order of files when loading an AVAA Corpus
• fixed "contains" not working for inherited Styles
• fixed line endings in avaa-toolkit.sh
• new op.load-data-from-variables

0.64

• EAF format now looks first for RELATIVE_MEDIA_URL
• new op.replace-with-annotations-in-sequence

0.63

• new view.controlled-vocabulary
• new proc.divide-tiers

0.62

• using MDC for better logs in server threads
• improved loops with headings and interactive mode
• new op.control-vocabulary
• new op.convert-to-slices
• new proc.export-to-tei
• op.set-tier: added template for scripted mapping
• proc.rename-tiers: added template for scripted mapping
• changed default port to 41744

0.61

• proc.audio-anonymizer: fixed timeOrigin bug
• new proc.reduce-corpus-media
• new proc.export-corpus-standalone
• fixed quotes in linux cli
• fixed api.exec() threading
• new avaa server mode
• added --ssl cli arg
• .exe not limited in RAM usage anymore

0.60

• added "difference" mode for select primitive
• added combining modes to stack/unstack/filter primitives
• fixed selecting annotations with a parent
• new mod.table-of-contents to customize table of content
• new mod.pdf-options to customize pdf export
• new proc.speech-to-text-faster-whisper
• new view.transcript
• new op.load-data-from-script
• new op.improve-transcript
• new op.combine-same-tier-consecutive-annotations
• new op.combine-overlapping-annotations
• new op.duration-of-pause
• op.set-global-variable: will set current data if no op value
• op.replace-with-same-timecode-annotations-from-tier: fixed range selection
• op.replace-with-annotations-in-sequence-from-tier: preserve combined annotations
• op.speaker-diarization-pyannote: fixed bug when too many annotations are passed
• op.speaker-diarization-pyannote: fixed bug when processing multiple corpus files
• op.duration-by: added modes
• op.group-by: added deep grouping (arrays in input object)
• op.sort: fixed natural sort
• view.table: separate display of combined annotations, automatic default columns
• view.timeline: fixed printable backgrounds
• proc.ffmpeg-denoise: added NLM and RNN modes
• proc.rename-tiers: added file attribute
• removed TeiCorpo dependency, migrated to providers for external handling
• editor: improved opening documents (valid links)
• editor: improved selection for tier/select inputs
• editor: added clipboard paste in empty containers
• editor: show error when opening invalid xml file
• editor: new button to make PDF
• editor: new AVAA Config screen
• avaa-config.xml: PDF settings

0.59

• embed jre from adoptium
• added avaa-toolkit.exe windows launcher
• added avaa-config.xml

0.58

• added concordancer attributes (show-tier..)
• better density-timeline absolute-start display
• fixed document import missing styles
• fixed snapshots time-origin bug
• fixed progress % when stacking clips with multiple video-inputs
• fixed clip when multiple video-inputs but no video available at a specific time
• FTP can now deploy multiple zips into same directory
• editor: added color-picker for color typed fields

0.57

• allow custom subdirectories in FTP deployer
• FTP improvements (UTF-8 and absolute paths)
• handle FTP shortcut connection link

0.56

• cancellable conversion + improved progress feedback
• density-timeline (relate-to-corpus & time-between)
• deploy via FTP

0.55

• now using logback
• fixed vad-silero

0.54

• new oTranscribe provider for handling OTR format
• added launcher shell script for linux/macOS
• added examples folder

0.53

• new: custom tags can be associated to corpus files
• fix core issues related to reduce-corpus
• fix core issues with pipeline input modes
• python modules auto installed with pip if possible
• view.wordcloud: added attributes + auto size-factor
• proc.video-anonymizer: added more deface options

0.52

• new document description field
• new advene provider for handling AZP format
• new bulk provider for batch processing many files
• new reduce-corpus processor
• new export-corpus-media processor
• new video-anonymizer processor
• new r-script processor draft
• moved deface processor to video-anonymizer
• updated media-converter processor
• new mongo database operations
• editor: inserting blocks anywhere fixed
• editor: charts sample image when adding charts

0.51

• handles multiple media inputs correctly (+timeOrigin fix)
• new global settings: video-input / video-inputs / video-stack-mode
• new hardsub processor to burn subtitles into video clips
• migrated SRT handling to subrib provider
• migrated MKV handling to matroska provider
• editor: insert new blocks at any node

0.50

• chart legend adapted to AVAA_STYLE
• save live chart preview to PNG/JPG/WEBP (with custom background/scale)

0.49

• added equals/contains id match for global styles colors/labels
• adapted charts to use new AVAA_STYLE object

0.48

• disk usage ui (assets/temp)
• restricting server connections to allowed origins
• location.hash shortcut links
• new rename-tiers processor

0.47

• charts live preview in editor
• fix global labelling in horizontal-bar chart
• fix donut chart

0.46

• annotations provider system
• new export-to-eaf processor
• new export-to-srt processor
• new sequences-to-domain processor
• new remove-sequences-from-domain processor
• new speaker-diarization-pyannote processor
• ui: improved domain and realtime feedbacks

0.45

• fix set-tier reusing same tier for different annotation files
• fix frei0r effects rendering invalid image format for firefox player
• view.form: added cotext and extend-duration attributes
• improved load-annotations-from-forms
• new custom document CSS via CSS tags in STYLE tags

0.44

• fix processor temp path

0.43

• introduction of Processors
• new file-match attribute for select and filter tags
• editor realtime log feedback and various improvements

0.42

• Remote Corpus (auto download corpus if needed)
• introduction of <? return 'str' ?> syntax in HTML block to insert custom strings
• introduction of variables which can be used in HTML blocks as {{varname}}
• introduction of Modifiers, scripts that can hook anywhere during generation
• new IF tags for conditional branching
• new set-global-variable operation to define variables
• new slideshow modifier to transform page into a PowerPoint like presentation

0.41

• new scrape operation to fetch and extract data from webpages

0.40

• fix html escaping in legend.swatches

0.39

• fix handling arrays of annotations containing null values in extra branch

0.38

• editor selectable nodes with copy/paste
• scope for Style entries (tier/value)

0.37

• annotation.clone() was not copying tags

0.36

• view.intercoding coincidence and weight matrix + krippendorff

0.35

• added --lang argument
• editor now supports undo

0.34

• text encoding fixed in density view and set-tier operation

0.33

• text encoding fallback in getTier()
• renamed "select-" operations to "replace-with-"
• multi-language initial implementation

0.32

• deployment facility (unpack zip and host)

0.31

• file dialog improvements when adding file to domain
• added file in domain will try to use relative path if possible
• removed version from release folder name, easing updates
• attributes deprecation warning in editor and documentation
• a click on the top-left green rectangle brings back home tab
• new files can be created
• fixed "browse" for operations