Sparrowhawk is an open-source implementation of Google's Kestrel text-to-speech text normalization system. It follows the discussion of the Kestrel system as described in:
Ebden, Peter and Sproat, Richard. 2015. The Kestrel TTS text normalization system. Natural Language Engineering, Issue 03, pp 333-353.
After sentence segmentation (sentence_boundary.h), the individual sentences are
first tokenized with each token being classified, and then passed to the
normalizer. The system can output as an unannotated string of words, and richer
annotation with links between input tokens, their input string positions, and
the output words is also available.
This version is a fork of the original
google/sparrowhawk package, archived
in 2022, with blessings from @rwsproat to keep using the name.
Instead of using autotools for building, this version uses Bazel for dependency management and building, and uses C++17 for building. Tested to be working with the text normalization rules in https://github.com/NVIDIA/NeMo-text-processing, without having to pull in the Python dependencies at runtime.
To depend on Sparrowhawk, add the following to your MODULE.bazel:
bazel_dep(name = "sparrowhawk", version = "2.0.0")
And add a dependency on @sparrowhawk//:normalizer_lib in your C++ files where
you #include "sparrowhawk/normalizer.h".