-
refactor: initialize text processor from text config
-
feat: add g2p module and updated puncutation handling
-
refactor: large refactor of text processor
-
feat: write string-encoded character and phone token sequences to filelist
-
fix(tests): add empty string join character when decoding
-
refactor: multiple refactors
-
feat(wizard): update handling of input text
-
fix(tests): fix wizard unit tests
-
feat(fs2.cli): add text_type specification
-
feat(wizard): add progress bar for grapheme/phoneme discovery
-
test: add doctests to text suite
-
fix: sort fieldnames to improve filelist readability
-
perf: mind your imports and keep the cli fast
-
chore: update submodules for import refactoring
-
feat: remove duplicates in symbols by default
-
fix: remove lowercase ascii from default symbols
-
fix: missing ascii characters in preprocessing test
-
fix: only encode character or phone strings when they exist
-
fix: pin typer to less than 0.12.0
-
refactor: defining the text representation is unnecessary
-
refactor: change csv to psv
-
refactor: add changes suggested by Sam
-
fix: pin typer to 0.9.0
-
fix: properly process g2p for multi-lingual filelists
-
feat: remove punctuation from automatically guessed characters
-
docs: alternate method
-
feat: message the user about punctuation in character set
-
feat: simpler iteration over fields and skipping punctuation
-
feat: helper functions
-
fix: doctests
-
fix(pfs): phonological features should apply punctuation transformation
-
refactor: save pfs to pfs folder, not text
-
refactor: use dict mappings for symbol and id
-
refactor: set punctuation rule application everywhere
-
refactor: apply fixes and refactors suggested by @joanise
-
fix: filter text data based on target training representation