Run some automatic tests with GitHub Actions (#68)

tests/conftest.py

@@ -0,0 +1,31 @@
+import os
+
+import ctranslate2
+import pytest
+
+
+@pytest.fixture
+def data_dir():
+    return os.path.join(os.path.dirname(os.path.abspath(__file__)), "data")
+
+
+@pytest.fixture
+def jfk_path(data_dir):
+    return os.path.join(data_dir, "jfk.flac")
+
+
+@pytest.fixture(scope="session")
+def tiny_model_dir(tmp_path_factory):
+    model_path = str(tmp_path_factory.mktemp("data") / "model")
+    convert_model("tiny", model_path)
+    return model_path
+
+
+def convert_model(size, output_dir):
+    name = "openai/whisper-%s" % size
+
+    ctranslate2.converters.TransformersConverter(
+        name,
+        copy_files=["tokenizer.json"],
+        load_as_float16=True,
+    ).convert(output_dir, quantization="float16")

tests/test.py

@@ -0,0 +1,25 @@
+from faster_whisper import WhisperModel
+
+
+def test_transcribe(tiny_model_dir, jfk_path):
+    model = WhisperModel(tiny_model_dir)
+    segments, info = model.transcribe(jfk_path, word_timestamps=True)
+
+    assert info.language == "en"
+    assert info.language_probability > 0.9
+    assert info.duration == 11
+
+    segments = list(segments)
+
+    assert len(segments) == 1
+
+    segment = segments[0]
+
+    assert segment.text == (
+        " And so my fellow Americans ask not what your country can do for you, "
+        "ask what you can do for your country."
+    )
+
+    assert segment.text == "".join(word.word for word in segment.words)
+    assert segment.start == segment.words[0].start
+    assert segment.end == segment.words[-1].end