def test_determine_labels(tmpdir): #fs is the fake filesystem fixture
"""test the function that determines what labels exist in a directory"""
from pathlib import Path
base_dir = tmpdir
label_dir = base_dir.mkdir("label")
test_1_phonemes = 'ɖ ɯ ɕ i k v̩'
test_1_phonemes_and_tones = 'ɖ ɯ ˧ ɕ i ˧ k v̩ ˧˥'
test_2_phonemes = 'g v̩ tsʰ i g v̩ k v̩'
test_2_phonemes_and_tones = 'g v̩ ˧ tsʰ i ˩ g v̩ ˩ k v̩ ˩'
label_dir.join("test1.phonemes").write(test_1_phonemes)
label_dir.join("test1.phonemes_and_tones").write(test_1_phonemes_and_tones)
label_dir.join("test2.phonemes").write(test_2_phonemes)
label_dir.join("test2.phonemes_and_tones").write(test_2_phonemes_and_tones)
all_phonemes = set(test_1_phonemes.split(' ')) | set(
test_2_phonemes.split(' '))
from persephone.corpus import determine_labels
phoneme_labels = determine_labels(Path(str(base_dir)), "phonemes")
assert phoneme_labels
assert phoneme_labels == all_phonemes
all_phonemes_and_tones = set(test_1_phonemes_and_tones.split(' ')) | set(
test_2_phonemes_and_tones.split(' '))
phoneme_and_tones_labels = determine_labels(Path(str(base_dir)),
"phonemes_and_tones")
assert phoneme_and_tones_labels
assert phoneme_and_tones_labels == all_phonemes_and_tones
def test_determine_labels_throws():
"""Test that a non existant directory will throw"""
import pathlib
from persephone.corpus import determine_labels
non_existent_path = pathlib.Path("thispathdoesntexist")
with pytest.raises(FileNotFoundError):
determine_labels(non_existent_path, "phonemes")
def test_fast():
"""
A fast integration test that runs 1 training epoch over a tiny
dataset. Note that this does not run ffmpeg to normalize the WAVs since
Travis doesn't have that installed. So the normalized wavs are included in
the feat/ directory so that the normalization isn't run.
"""
# 4 utterance toy set
TINY_EXAMPLE_LINK = "https://cloudstor.aarnet.edu.au/plus/s/g2GreDNlDKUq9rz/download"
tiny_example_dir = join(DATA_BASE_DIR, "tiny_example/")
rm_dir(Path(tiny_example_dir))
download_example_data(TINY_EXAMPLE_LINK)
labels = corpus.determine_labels(Path(tiny_example_dir), "phonemes")
corp = corpus.Corpus("fbank",
"phonemes",
Path(tiny_example_dir),
labels=labels)
exp_dir = experiment.prep_exp_dir(directory=EXP_BASE_DIR)
model = experiment.get_simple_model(exp_dir, corp)
model.train(min_epochs=2, max_epochs=5)
# Assert the convergence of the model at the end by reading the test scores
ler = get_test_ler(exp_dir)
# Can't expect a decent test score but just check that there's something.
assert ler < 2.0
def test_tutorial():
""" Tests running the example described in the tutorial in README.md """
# 1024 utterance sample set.
NA_EXAMPLE_LINK = "https://cloudstor.aarnet.edu.au/plus/s/YJXTLHkYvpG85kX/download"
na_example_dir = join(DATA_BASE_DIR, "na_example/")
rm_dir(Path(na_example_dir))
download_example_data(NA_EXAMPLE_LINK)
# Test the first setup encouraged in the tutorial
labels = corpus.determine_labels(Path(na_example_dir), "phonemes")
corp = corpus.Corpus("fbank", "phonemes", Path(na_example_dir), labels)
exp_dir = experiment.train_ready(corp, directory=EXP_BASE_DIR)
# Assert the convergence of the model at the end by reading the test scores
ler = get_test_ler(exp_dir)
assert ler < 0.3