Merge branch 'main' into feat-rename-exec-execute-to-perform

.github/workflows/build_and_deploy.yml

@@ -182,11 +182,11 @@ jobs:
             git fetch private refs/tags/${{ env.PRODUCTION_REPOSITORY_TAG }}
             git -c user.name=dummy -c [email protected] merge FETCH_HEAD
           ) > /dev/null 2>&1
-      - name: Set up Python 3.8
+      - name: Set up Python 3.10
         if: matrix.python_whl
         uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.10"
           architecture: ${{ contains(matrix.artifact_name,'x86') && 'x86' || 'x64' }}
       - name: set up ${{ matrix.target }}
         uses: ./.github/actions/rust-toolchain-from-file
@@ -225,7 +225,9 @@ jobs:
       - name: set cargo version
         run: |
           cargo set-version "$VERSION" --exclude voicevox_core_python_api --exclude downloader --exclude xtask
-          if ${{ matrix.python_whl }}; then cargo set-version "$VERSION" -p voicevox_core_python_api; fi
+          if ${{ matrix.python_whl }}; then
+            sed -i_ 's/version = "\(0\.0\.0\)"/version = "'"$VERSION"'"/' ./crates/voicevox_core_python_api/pyproject.toml
+          fi
       - name: cache target
         uses: Swatinem/rust-cache@v2
         if: ${{ !inputs.is_production }}

.github/workflows/generate_document.yml

@@ -33,7 +33,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.10"
       - name: Setup Java
         uses: actions/setup-java@v4
         with:

.github/workflows/python_lint.yml

@@ -24,14 +24,14 @@ jobs:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.10"
       - name: Install Poetry
         run: |
           python -m pip install --upgrade poetry
           poetry config virtualenvs.create false
       - name: Validate poetry.lock
         run: |
-          poetry lock --no-update
+          poetry lock
           git diff --exit-code
       - name: Install dependencies
         run: poetry install --with test

.github/workflows/test.yml

@@ -39,10 +39,10 @@ jobs:
         run: |
           sudo apt-get update
           sudo apt-get install -y shellcheck
-      - name: Set up Python 3.8
+      - name: Set up Python 3.10
         uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.10"
       - name: Install Pyflakes
         run: pip install 'pyflakes>3,<4'
       - name: actionlint
@@ -67,10 +67,10 @@ jobs:
         uses: ./.github/actions/rust-toolchain-from-file
         with:
           components: clippy,rustfmt
-      - name: Set up Python 3.8
+      - name: Set up Python 3.10
         uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.10"
       - uses: Swatinem/rust-cache@v2
       - run: cargo clippy -v --tests -- -D clippy::all -D warnings --no-deps
       - run: cargo clippy -v -- -D clippy::all -D warnings --no-deps
@@ -125,10 +125,10 @@ jobs:
     runs-on: ${{ matrix.os }}
     steps:
       - uses: actions/checkout@v4
-      - name: Set up Python 3.8
+      - name: Set up Python 3.10
         uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.10"
       - name: Set up Rust
         uses: ./.github/actions/rust-toolchain-from-file
       - uses: Swatinem/rust-cache@v2
@@ -272,10 +272,10 @@ jobs:
         working-directory: ./crates/voicevox_core_python_api
     steps:
       - uses: actions/checkout@v4
-      - name: Set up Python 3.8
+      - name: Set up Python 3.10
         uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.10"
       - name: Set up Rust
         uses: ./.github/actions/rust-toolchain-from-file
       - run: |
@@ -296,6 +296,18 @@ jobs:
           for file in ../../example/python/run{,-asyncio}.py; do
             poetry run python "$file" ../test_util/data/model/sample.vvm --dict-dir ../test_util/data/open_jtalk_dic_utf_8-1.11
           done
+          # https://github.com/VOICEVOX/voicevox_core/issues/873 が再発しないかの確認。
+          #
+          # Python 3.8においては、プロセスの終了までにFatal Python errorに至らなくても`StopIteration`までは達することが多い。
+          # #873の要因は依然として不明であるが、`StopIteration`が出ないことの確認をもって#873の解決とすることにする。
+          #
+          # TODO: 10回の実行におよそ60秒ほどかかるため、状況の経過を見てやめる。
+          for _ in {1..10}; do
+            poetry run python ../../example/python/run-asyncio.py ../test_util/data/model/sample.vvm --dict-dir ../test_util/data/open_jtalk_dic_utf_8-1.11 \
+              2> >(tee ./stderr.txt >&2)
+            # shellcheck disable=SC2059
+            ! grep -q StopIteration ./stderr.txt
+          done
   build-and-test-java-api:
     strategy:
       fail-fast: false

Cargo.toml

@@ -94,12 +94,13 @@ typetag = "0.2.18"
 url = "2.5.4"
 uuid = "1.10.0"
 voicevox_core = { path = "crates/voicevox_core" }
+voicevox_core_macros = { path = "crates/voicevox_core_macros" }
 windows = "0.43.0"
 zip = "0.6.3"
 
 [workspace.dependencies.voicevox-ort]
 git = "https://github.com/VOICEVOX/ort.git"
-rev = "09a9fe1619c1561efafc02f68f0bda4aad879771"
+rev = "cecd844162a1c6188de03b4566c81d9d38a28600"
 
 [workspace.dependencies.open_jtalk]
 git = "https://github.com/VOICEVOX/open_jtalk-rs.git"
@@ -114,7 +115,8 @@ rev = "de226a26e8e18edbdb1d6f986afe37bbbf35fbf4"
 version = "0.0.0"
 edition = "2021"
 publish = false
-rust-version = "1.81.0"
+rust-version = "1.84.0"
+license = "MIT"
 
 # min-sized-rustを元にrelease buildのサイズが小さくなるようにした
 # https://github.com/johnthagen/min-sized-rust

crates/downloader/Cargo.toml

@@ -2,6 +2,7 @@
 name = "downloader"
 edition.workspace = true
 rust-version.workspace = true
+license.workspace = true
 
 [[bin]]
 name = "download"

crates/test_util/Cargo.toml

@@ -2,6 +2,7 @@
 name = "test_util"
 edition.workspace = true
 rust-version.workspace = true
+license.workspace = true
 
 [dependencies]
 libloading.workspace = true

crates/voicevox_core/Cargo.toml

@@ -52,7 +52,7 @@ thiserror.workspace = true
 tracing.workspace = true
 uuid = { workspace = true, features = ["v4", "serde"] }
 voicevox-ort = { workspace = true, features = ["download-binaries", "__init-for-voicevox"] }
-voicevox_core_macros = { path = "../voicevox_core_macros" }
+voicevox_core_macros.workspace = true
 
 [dev-dependencies]
 heck.workspace = true

crates/voicevox_core/src/core.rs

@@ -0,0 +1,3 @@
+mod adjust;
+
+pub(crate) use self::adjust::{ensure_minimum_phoneme_length, pad_decoder_feature};

crates/voicevox_core/src/core/adjust.rs

@@ -0,0 +1,6 @@
+//! 推論操作の前処理と後処理。
+
+mod post;
+mod pre;
+
+pub(crate) use self::{post::ensure_minimum_phoneme_length, pre::pad_decoder_feature};

crates/voicevox_core/src/core/adjust/post.rs

@@ -0,0 +1,12 @@
+//! 推論の出力の後処理。
+
+pub(crate) fn ensure_minimum_phoneme_length(mut output: Vec<f32>) -> Vec<f32> {
+    const PHONEME_LENGTH_MINIMAL: f32 = 0.01;
+
+    for output_item in output.iter_mut() {
+        if *output_item < PHONEME_LENGTH_MINIMAL {
+            *output_item = PHONEME_LENGTH_MINIMAL;
+        }
+    }
+    output
+}

crates/voicevox_core/src/core/adjust/pre.rs

@@ -0,0 +1,37 @@
+//! 推論の入力の前処理。
+
+/// 音が途切れてしまうのを避けるworkaround処理。
+// TODO: 改善したらここのpadding処理を取り除く
+pub(crate) fn pad_decoder_feature<const PADDING_FRAME_LENGTH: usize>(
+    f0: ndarray::Array1<f32>,
+    phoneme: ndarray::Array2<f32>,
+) -> (usize, ndarray::Array1<f32>, ndarray::Array2<f32>) {
+    let start_and_end_padding_size = 2 * PADDING_FRAME_LENGTH;
+    let length_with_padding = f0.len() + start_and_end_padding_size;
+    let f0_with_padding = make_f0_with_padding(f0, PADDING_FRAME_LENGTH);
+    let phoneme_with_padding = make_phoneme_with_padding(phoneme, PADDING_FRAME_LENGTH);
+    return (length_with_padding, f0_with_padding, phoneme_with_padding);
+
+    fn make_f0_with_padding(
+        f0_slice: ndarray::Array1<f32>,
+        padding_size: usize,
+    ) -> ndarray::Array1<f32> {
+        // 音が途切れてしまうのを避けるworkaround処理
+        // 改善したらこの関数を削除する
+        let padding = ndarray::Array1::<f32>::zeros(padding_size);
+        ndarray::concatenate![ndarray::Axis(0), padding, f0_slice, padding]
+    }
+
+    fn make_phoneme_with_padding(
+        phoneme_slice: ndarray::Array2<f32>,
+        padding_size: usize,
+    ) -> ndarray::Array2<f32> {
+        // 音が途切れてしまうのを避けるworkaround処理
+        // 改善したらこの関数を削除する
+        let mut padding = ndarray::Array2::<f32>::zeros((padding_size, phoneme_slice.ncols()));
+        padding
+            .slice_mut(ndarray::s![.., 0])
+            .assign(&ndarray::arr0(1.0));
+        ndarray::concatenate![ndarray::Axis(0), padding, phoneme_slice, padding]
+    }
+}

crates/voicevox_core/src/engine/audio_file.rs

@@ -1,5 +1,33 @@
 use std::io::{Cursor, Write as _};
 
+use super::AudioQuery;
+
+pub(crate) fn to_s16le_pcm<const BASE_SAMPLING_RATE: u32>(
+    wave: &[f32],
+    &AudioQuery {
+        volume_scale,
+        output_sampling_rate,
+        output_stereo,
+        ..
+    }: &AudioQuery,
+) -> Vec<u8> {
+    let num_channels: u16 = if output_stereo { 2 } else { 1 };
+    let repeat_count: u32 = (output_sampling_rate / BASE_SAMPLING_RATE) * num_channels as u32;
+    let bytes_size = wave.len() as u32 * repeat_count * 2;
+    let buf: Vec<u8> = Vec::with_capacity(bytes_size as usize);
+    let mut cur = Cursor::new(buf);
+
+    for value in wave {
+        let v = (value * volume_scale).clamp(-1., 1.);
+        let data = (v * 0x7fff as f32) as i16;
+        for _ in 0..repeat_count {
+            cur.write_all(&data.to_le_bytes()).unwrap();
+        }
+    }
+
+    cur.into_inner()
+}
+
 /// 16bit PCMにヘッダを付加しWAVフォーマットのバイナリを生成する。
 pub fn wav_from_s16le(pcm: &[u8], sampling_rate: u32, is_stereo: bool) -> Vec<u8> {
     let num_channels: u16 = if is_stereo { 2 } else { 1 };