Errors in running unit test after development installation

[ghost]

Hello, I installed lhotse from github (development installation) as follows:

https://lhotse.readthedocs.io/en/latest/getting-started.html#installation

but I got 3 errors from pytest test as follows:

======== test session starts ==============[79/279]
platform linux -- Python 3.6.9, pytest-5.4.3, py-1.10.0, pluggy-0.13.1
rootdir: /scl_nas/lhotse
plugins: hypothesis-5.41.2
collected 779 items / 3 errors / 776 selected

========= ERRORS========
_________________ ERROR collecting test/test_feature_set.py ____________________________
test/test_feature_set.py:7: in
import torchaudio
/usr/local/lib/python3.6/dist-packages/torchaudio/init.py:14: in
from torchaudio.backend import (
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/init.py:1: in
from . import utils
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/utils.py:7: in
from . import (
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/soundfile_backend.py:15: in
import soundfile
/usr/local/lib/python3.6/dist-packages/soundfile.py:142: in
raise OSError('sndfile library not found')
E OSError: sndfile library not found
____________ ERROR collecting test/augmentation/test_torchaudio.py ________________________________________
test/augmentation/test_torchaudio.py:9: in
torchaudio = pytest.importorskip('torchaudio', minversion='0.6')
/usr/local/lib/python3.6/dist-packages/torchaudio/init.py:14: in
from torchaudio.backend import (
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/init.py:1: in
from . import utils
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/utils.py:7: in
from . import (
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/soundfile_backend.py:15: in
import soundfile
/usr/local/lib/python3.6/dist-packages/soundfile.py:142: in
raise OSError('sndfile library not found')
E OSError: sndfile library not found
___________________ ERROR collecting test/known_issues/test_augment_with_executor.py ___________________
test/known_issues/test_augment_with_executor.py:12: in
torchaudio = pytest.importorskip('torchaudio', minversion='0.7.1')
/usr/local/lib/python3.6/dist-packages/torchaudio/init.py:14: in
from torchaudio.backend import (
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/init.py:1: in
from . import utils
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/utils.py:7: in
from . import (
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/soundfile_backend.py:15: in
import soundfile
/usr/local/lib/python3.6/dist-packages/soundfile.py:142: in
raise OSError('sndfile library not found')
E OSError: sndfile library not found
=========== warnings summary ============
lhotse/recipes/heroico.py:81
/scl_nas/lhotse/lhotse/recipes/heroico.py:81: DeprecationWarning: invalid escape sequence \d
answers_line_pattern = re.compile("\d+/\d+\t.+")

lhotse/recipes/heroico.py:83
/scl_nas/lhotse/lhotse/recipes/heroico.py:83: DeprecationWarning: invalid escape sequence \d
heroico_recitations_line_pattern = re.compile("\d+\t.+")

lhotse/recipes/heroico.py:86
/scl_nas/lhotse/lhotse/recipes/heroico.py:86: DeprecationWarning: invalid escape sequence \d
usma_line_pattern = re.compile("s\d+\t.+")

lhotse/recipes/heroico.py:87
/scl_nas/lhotse/lhotse/recipes/heroico.py:87: DeprecationWarning: invalid escape sequence -
usma_native_demo_pattern = re.compile("usma/native-[fm]-\w+-\S+-\S+-\S+-\S+-\w+\d+")

lhotse/recipes/heroico.py:89
/scl_nas/lhotse/lhotse/recipes/heroico.py:89: DeprecationWarning: invalid escape sequence \d
usma_native_prompt_id_pattern = re.compile('s\d+')

lhotse/recipes/heroico.py:91
/scl_nas/lhotse/lhotse/recipes/heroico.py:91: DeprecationWarning: invalid escape sequence -
"nonnative-[fm]-[a-zA-Z]+\d*-[a-zA-Z]+-[a-zA-Z]+-[a-zA-Z]+-[a-zA-Z]+-[a-zA-Z]+\d+")

lhotse/recipes/heroico.py:92
/scl_nas/lhotse/lhotse/recipes/heroico.py:92: DeprecationWarning: invalid escape sequence .
usma_nonnative_path_pattern = re.compile('nonnative.+.wav')

-- Docs: https://docs.pytest.org/en/latest/warnings.html
====== short test summary info =======
ERROR test/test_feature_set.py - OSError: sndfile library not found
ERROR test/augmentation/test_torchaudio.py - OSError: sndfile library not found
ERROR test/known_issues/test_augment_with_executor.py - OSError: sndfile library not found
!!!!!!!!!!!!!!!!!!!!!!!!!! Interrupted: 3 errors during collection !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
=========== 7 warnings, 3 errors in 4.34s ==============

To resolve this errors, we consider re-installation of Torchaudio with higher version than 0.7.2?
because version of Torchaudio is 0.7.2.

Other environment is as follows:

Ubuntu OS version: 18.04.5 LTS
Python version: 3.6.9
Cuda version: 11.0
Torch version: 1.7.1+cu110
Torchvision version: torchvision-0.8.2+cu110

[ghost]

Oh, I got it. I search "OSError: sndfile library not found" and execute the following command:
sudo apt-get install libsndfile1

and it works well with some warnings as follows:

========== test session starts ==========
platform linux -- Python 3.6.9, pytest-5.4.3, py-1.10.0, pluggy-0.13.1
rootdir: /scl_nas/lhotse
plugins: hypothesis-5.41.2collected 833 items

test/test_audio_url.py s [ 0%]
test/test_feature_set.py .............................. [ 3%]
test/test_manipulation.py ...................................................... [ 10%]
test/test_qa.py .... [ 10%]
test/test_recording_set.py .....................................................ssss [ 17%]
test/test_resample_randomized.py . [ 17%]
test/test_serialization.py ...................................................................................... [ 27%]
test/test_supervision_set.py ........................ [ 30%]
test/test_utils.py ......................................... [ 35%]
test/augmentation/test_torchaudio.py ...................... [ 38%]
test/cut/test_cut.py ...................... [ 41%]
test/cut/test_cut_augmentation.py ............................ [ 44%]
test/cut/test_cut_drop_attributes.py ............ [ 45%]test/cut/test_cut_mixing.py ........... [ 47%]
test/cut/test_cut_ops_preserve_id.py ........................................ [ 51%]
test/cut/test_cut_set.py .......................... [ 55%]
test/cut/test_cut_set_creation.py .... [ 55%]
test/cut/test_cut_set_mix.py ......... [ 56%]
test/cut/test_cut_trim_to_supervisions.py ..... [ 57%]
test/cut/test_cut_truncate.py ................................... [ 61%]
test/cut/test_feature_extraction.py ........ssssssss [ 63%]
test/cut/test_invariants_randomized.py .. [ 63%]
test/cut/test_masks.py ................ [ 65%]test/cut/test_padding_cut.py ................................................. [ 71%]
test/dataset/test_collation.py ...... [ 72%]
test/dataset/test_cut_transforms.py ................ [ 74%]
test/dataset/test_dataloading.py s [ 74%]
test/dataset/test_diarization.py . [ 74%]
test/dataset/test_sampling.py .........x...........x..........................................................................................x............. [ 89%]
test/dataset/test_signal_transforms.py ..................... [ 91%]test/dataset/test_speech_recognition_dataset.py ............. [ 93%]
test/dataset/test_speech_recognition_dataset_randomized.py . [ 93%]
test/dataset/test_speech_synthesis_dataset.py ... [ 93%]
test/dataset/test_unsupervised_dataset.py ... [ 94%]
test/dataset/test_vad_dataset.py . [ 94%]
test/features/test_kaldi_features.py .... [ 94%]
test/features/test_librosa_fbank.py ..........ssssss [ 96%]
test/known_issues/test_augment_with_executor.py .s [ 97%]
test/known_issues/test_cut_consistency.py ...................... [ 99%]
test/known_issues/test_mixed_cut_num_frames.py .. [100%]

============= warnings summary =========
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/utils.py:54
/usr/local/lib/python3.6/dist-packages/torchaudio/backend/utils.py:54: UserWarning: "sox" backend is being deprecated. The default backend will be changed to "sox_io" backend in 0.8.0 and "sox" backend will be removed in 0.9.0. Please migrate to "sox_io" backend. Please refer to pytorch/audio#903 for the detail.
'"sox" backend is being deprecated. '

test/test_feature_set.py::test_feature_extractor[mfcc-exception_expectation0]
/usr/local/lib/python3.6/dist-packages/torchaudio/compliance/kaldi.py:574: UserWarning: The function torch.rfft is deprecated and will be removed in a future PyTorch release. Use the new torch.fft module functions, instead, by importing torch.fft and calling torch.fft.fft or torch.fft.rfft. (Triggered internally at /pytorch/aten/src/ATen/native/SpectralOps.cpp:590.)
fft = torch.rfft(strided_input, 1, normalized=False, onesided=True)

test/test_serialization.py: 1 test with warning
test/cut/test_cut.py: 8 tests with warnings
test/cut/test_cut_set.py: 2 tests with warnings
test/dataset/test_collation.py: 2 tests with warnings
test/dataset/test_diarization.py: 1 test with warning
test/dataset/test_sampling.py: 1 test with warning
test/dataset/test_signal_transforms.py: 16 tests with warnings
test/dataset/test_speech_recognition_dataset.py: 12 tests with warnings
test/dataset/test_speech_synthesis_dataset.py: 3 tests with warnings
test/dataset/test_unsupervised_dataset.py: 3 tests with warnings
test/dataset/test_vad_dataset.py: 1 test with warning
/scl_nas/lhotse/lhotse/cut.py:2208: UserWarning: Your manifest was created with Lhotse version earlier than v0.8, when MonoCut was called Cut. Please re-generate it with Lhotse v0.8 as it might stop working in a future version (usingmanifest.from_file() and then manifest.to_file() should be sufficient).
'Your manifest was created with Lhotse version earlier than v0.8, when MonoCut was called Cut. '

test/cut/test_cut_set.py::test_cut_set_describe_runs
/usr/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: numpy.ufunc size changed, may indicate binary incompatibility. Expected 192 from C header, got 216 from PyObject
return f(*args, **kwds)

test/dataset/test_sampling.py::test_single_cut_sampler_low_max_frames
test/dataset/test_speech_recognition_dataset.py::test_k2_speech_recognition_iterable_dataset_low_max_frames
/scl_nas/lhotse/lhotse/dataset/sampling/single_cut.py:171: UserWarning: The first cut drawn in batch collection violates the max_frames, max_cuts, or max_duration constraints - we'll return it anyway. Consider increasing max_frames/max_cuts/max_duration.
"The first cut drawn in batch collection violates "

test/dataset/test_sampling.py::test_bucketing_sampler_single_cuts_equal_duration
test/dataset/test_sampling.py::test_bucketing_sampler_cut_pairs_equal_duration[False]
test/dataset/test_sampling.py::test_bucketing_sampler_cut_pairs_equal_duration[True]
/scl_nas/lhotse/lhotse/dataset/sampling/bucketing.py:328: DeprecationWarning: Calling np.sum(generator) is deprecated, and in the future will give a different result. Use np.sum(np.fromiter(generator)) or the python sum builtin instead.
total_duration = np.sum(c.duration for c in cuts)

test/dataset/test_sampling.py::test_single_cut_sampler_with_lazy_cuts[BucketingSampler]
test/dataset/test_sampling.py::test_single_cut_sampler_lazy_shuffle[BucketingSampler]
test/dataset/test_sampling.py::test_cut_pairs_sampler_lazy_shuffle[BucketingSampler]
/scl_nas/lhotse/lhotse/dataset/sampling/bucketing.py:96: UserWarning: Lazy CutSet detected in BucketingSampler: this is not well supported yet, and you might experience a potentially long lag while the buckets are being created.
"Lazy CutSet detected in BucketingSampler: this is not well supported yet, "

-- Docs: https://docs.pytest.org/en/latest/warnings.html
================ 809 passed, 21 skipped, 3 xfailed, 61 warnings in 94.33s (0:01:34) ========================

[pzelasko]

The warnings are expected. I think you’re good to go.

[ghost]

@pzelasko Thank you, I'll close this issue.