feat: allow simple filenames for synthesis

.coveragerc

@@ -0,0 +1,16 @@
+[run]
+branch = true
+source_pkgs =
+    styletts2
+    everyvoice.model.e2e.StyleTTS2_lightning.styletts2
+omit =
+    *tmp*
+    */tests/*
+    */__main__.py
+
+[report]
+precision = 2
+exclude_lines =
+    pragma: no cover
+    if 0:
+    if __name__ == .__main__.:

.github/workflows/test.yml

@@ -0,0 +1,77 @@
+name: Run Tests
+on:
+  - push
+  - pull_request
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    permissions:
+      contents: read
+    env:
+      # Fix a bug in the ffmpeg installation via awalsh128/cache-apt-pkgs-action
+      LD_LIBRARY_PATH: /usr/lib/x86_64-linux-gnu/lapack:/usr/lib/x86_64-linux-gnu/blas
+    steps:
+      - uses: actions/checkout@v6
+
+      - uses: actions/setup-python@v6
+        with:
+          python-version: "3.10"
+
+      - uses: awalsh128/cache-apt-pkgs-action@681749ae568c81c2037cb9185e38b709b261bd2f # v1.6.1
+        with:
+          packages: sox ffmpeg
+
+      - run: sox --version
+      - run: ffmpeg -version
+
+      - name: Fetch everyvoice
+        run: |
+          git clone https://github.com/EveryVoiceTTS/EveryVoice
+          cd EveryVoice
+          git checkout ${{ github.ref_name }} || git checkout ${{ github.head_ref }} || true
+          git submodule update --init
+          cd everyvoice/model/e2e/StyleTTS2_lightning
+          git fetch origin ${{ github.ref }}
+          git checkout FETCH_HEAD
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@fac544c07dec837d0ccb6301d7b5580bf5edae39 # v8.2.0
+        with:
+          version: "0.11.21"
+
+      - name: Install the project
+        run: |
+          cd EveryVoice/everyvoice/model/e2e/StyleTTS2_lightning/
+          uv sync --locked --all-extras --dev
+          uv pip install --editable ../../../..
+
+      - name: uv pip freeze
+        run: |
+          cd EveryVoice/everyvoice/model/e2e/StyleTTS2_lightning/
+          uv pip freeze
+      - name: uv pip list
+        run: |
+          cd EveryVoice/everyvoice/model/e2e/StyleTTS2_lightning/
+          uv pip list
+
+      - name: Run unit tests
+        run: |
+          cd EveryVoice/everyvoice/model/e2e/StyleTTS2_lightning/
+          uv run coverage run -m pytest
+          uv run coverage xml
+        env:
+          PYTHONPATH: ${{ github.workspace }}/EveryVoice
+
+      - name: Plain text coverage report
+        run: |
+          cd EveryVoice/everyvoice/model/e2e/StyleTTS2_lightning/
+          uv run coverage report
+
+      - uses: codecov/codecov-action@v5
+        with:
+          fail_ci_if_error: false # optional (default = false)
+          token: ${{ secrets.CODECOV_TOKEN }}

styletts2/cli/synthesize.py

@@ -6,11 +6,13 @@
 
 from __future__ import annotations
 
+import sys
 from pathlib import Path
 
 import typer
 from everyvoice import logger
 from everyvoice.base_cli import command, default_typer_args
+from everyvoice.base_cli.interfaces import typer_file_option
 from everyvoice.model.feature_prediction.FastSpeech2_lightning.fs2.type_definitions import (
     SynthesizeOutputFormats,
 )
@@ -138,30 +140,54 @@ def synthesize(
         file_okay=True,
         dir_okay=False,
     ),
-    reference: Path = typer.Option(
-        ...,
+    reference: Path | None = typer_file_option(
+        None,
         "--reference",
         "-r",
-        help="Reference audio file used to extract speaker style.",
-        exists=True,
+        help="Reference audio file used to extract speaker style. Required"
+        " unless every row of --filelist provides its own 'reference' or"
+        " 'reference_path' column.",
     ),
     text: list[str] = typer.Option(
-        ...,
+        [],
         "--text",
         "-t",
-        help="Text string(s) to synthesize. Repeat the flag for multiple utterances.",
+        help="Text string(s) to synthesize. Repeat the flag for multiple utterances."
+        " Use --filelist instead if you want to synthesize a lot of sentences or"
+        " have different speaker/language/reference per sentence.",
+    ),
+    filelist: Path | None = typer_file_option(
+        None,
+        "--filelist",
+        "-f",
+        help="The path to a file containing a list of utterances (a.k.a filelist)."
+        " Expected columns: 'basename', 'characters' (or 'phones'), 'speaker',"
+        " 'language', and optionally 'reference' (or 'reference_path'). Any column"
+        " that is absent falls back to the corresponding --speaker/--language/"
+        "--reference CLI option. Use --text if you want to just synthesize one sample.",
     ),
     output_dir: Path = typer.Option(
         Path("synthesis_output"),
         "--output-dir",
         "-o",
-        help="Directory where synthesized files will be written.",
+        help="Directory where synthesized files will be written."
+        " By default, filenames include the basename, speaker, language, and"
+        " other metadata, e.g. 'LJ050-0269--LJ--eng--ckpt=100000--pred.wav'."
+        " Use --simple-filenames to write just 'LJ050-0269.wav' instead.",
     ),
     output_type: list[SynthesizeOutputFormats] = typer.Option(
         [SynthesizeOutputFormats.wav],
         "--output-type",
         help="Output format(s) to produce.",
     ),
+    simple_filenames: bool = typer.Option(
+        False,
+        "--simple-filenames",
+        help="Write output filenames as just the basename and extension"
+        " (e.g. 'LJ050-0269.wav') instead of the default."
+        " Only use this if your basenames are unique across speakers and"
+        " languages, otherwise outputs can overwrite each other.",
+    ),
     accelerator: str = typer.Option(
         "auto",
         "--accelerator",
@@ -204,21 +230,42 @@ def synthesize(
 ):
     """Synthesize audio from text using a trained StyleTTS2 model.
 
-    Example:
+    Examples:
 
     **everyvoice synthesize text-to-wav logs_and_checkpoints/.../stage-2-last.ckpt \\
         --reference path/to/reference.wav \\
         --text "Hello world" --text "How are you?"**
+
+    Or, for batch synthesis from a filelist:
+
+    **everyvoice synthesize text-to-wav logs_and_checkpoints/.../stage-2-last.ckpt \\
+        --reference path/to/reference.wav --filelist my_filelist.psv --simple-filenames**
     """
+    # Do argument error checking before doing expensive imports
+    if text and filelist:
+        print(
+            "Got arguments for both --text and --filelist."
+            " You can only synthesize using one of these options",
+            file=sys.stderr,
+        )
+        sys.exit(1)
+    if not text and not filelist:
+        print("You must define either --text or --filelist", file=sys.stderr)
+        sys.exit(1)
+    if text and reference is None:
+        print(
+            "Missing --reference option, which is required when using --text.",
+            file=sys.stderr,
+        )
+        sys.exit(1)
+
     import lightning as L
     import torch
-    from everyvoice.model.feature_prediction.FastSpeech2_lightning.fs2.utils import (
-        truncate_basename,
-    )
-    from everyvoice.utils import slugify
 
     from .utils_heavy import (
         StyleTTS2SynthesisDataModule,
+        build_filelist_entries,
+        build_text_entries,
         get_styletts2_synthesis_output_callbacks,
     )
 
@@ -238,23 +285,41 @@ def synthesize(
     state = torch.load(model_path, map_location="cpu", weights_only=True)
     global_step = int(state.get("global_step", 0))
 
-    entries = [
-        {
-            "raw_text": t,
-            "basename": truncate_basename(slugify(t)),
-            "speaker": speaker,
-            "language": language,
-            "reference_path": str(reference),
-            "diffusion_steps": diffusion_steps,
-            "embedding_scale": embedding_scale,
-            "acoustic_blend": acoustic_blend,
-            "prosody_blend": prosody_blend,
-        }
-        for t in text
-    ]
+    try:
+        if text:
+            entries = build_text_entries(
+                text,
+                str(reference),
+                speaker,
+                language,
+                diffusion_steps,
+                embedding_scale,
+                acoustic_blend,
+                prosody_blend,
+            )
+        else:
+            assert filelist is not None
+            filelist_loader = module.config["ev_config"].training.filelist_loader
+            entries = build_filelist_entries(
+                filelist_loader(filelist),
+                str(reference) if reference else None,
+                speaker,
+                language,
+                diffusion_steps,
+                embedding_scale,
+                acoustic_blend,
+                prosody_blend,
+            )
+    except ValueError as e:
+        logger.error(str(e))
+        sys.exit(1)
 
     callbacks = get_styletts2_synthesis_output_callbacks(
-        output_type, output_dir, global_step, module.sr
+        output_type,
+        output_dir,
+        global_step,
+        module.sr,
+        simple_filenames=simple_filenames,
     )
     if not callbacks:
         logger.warning("No output format requested; nothing to do.")