📚 Documentation issue reports and fix offers

[mthrok]

This issue is for reporting the error on documentation and offering the update.
If you find an error on documentation or contents hard to understand, feel free to add a comment. (Adding a suggestion would make it easier for the contributor to work on)

If you would like to work on the fix, feel free to open a PR and attach the screenshot of the documentation.
See examples: #972 #1001

We are currently working on bringing up the automated master build for documentation at pytorch.org/audio, but meanwhile the documentation needs to be built manually. The following is the steps to build documentation.

git clone https://github.com/pytorch/audio
cd docs
pip install -r requirements.txt
make html

# open build/html/index.html with your browser.

[mthrok]

Issue at

audio/torchaudio/sox_effects/sox_effects.py

Lines 222 to 226 in b7c17f8

	... def __init__(self, flist: List[str], sample_rate: int):
	... super().__init__()
	... self.flist = flist
	... self.sample_rate = sample_rate
	... self.rng = None

self.rng should be a random generator.

[mthrok]

✅ Resolved in #1022

---

Issue in datasets.rst:

All datasets are subclasses of :class:torch.utils.data.Dataset i.e, they have __getitem__ and __len__ methods implemented.

__len__ is not the requirement of subclassing torch.utils.data.Datset. It should be something like,

All datasets are subclasses of :class:torch.utils.data.Dataset, and they have __getitem__ and __len__ methods implemented.

</deitals

[mthrok]

✅ Resolved in #1022

---

Issue at

audio/torchaudio/functional.py

Line 1242 in b7c17f8

Scott Lehman, Effects Explained, http://harmony-central.com/Effects/effects-explained.html

The link is invalid. Remove it or maybe replace it with web archive?

[mthrok]

Issues in functional.py

audio/torchaudio/functional.py

Lines 723 to 725 in b7c17f8

	References:
	http://sox.sourceforge.net/sox.html
	https://www.w3.org/2011/audio/audio-eq-cookbook.html#APF

The syntax of references are not correct, they should be something like

audio/torchaudio/models/conv_tasnet.py

Lines 25 to 30 in b7c17f8

	Reference:
	- Conv-TasNet: Surpassing Ideal Time--Frequency Magnitude Masking for Speech Separation
	Luo, Yi and Mesgarani, Nima
	https://arxiv.org/abs/1809.07454

[mthrok]

✅ Resolved in #1022

---

Issue in functional.rst.

DB_to_amplitude is missing

[mthrok]

✅ Resolved in #1022

---

Issue:

audio/torchaudio/utils/sox_utils.py

Lines 27 to 32 in b7c17f8

	Args:
	verbosity (int): Set verbosity level of libsox.
	1: failure messages
	2: warnings
	3: details of processing
	4-6: increasing levels of debug messages

The list is not properly rendered https://pytorch.org/audio/0.7/utils.html#torchaudio.utils.sox_utils.set_verbosity

[mthrok]

✅ Resolved in #1022

---

Issue: griffinlim is missing from functional.rst

[mthrok]

Issues

audio/torchaudio/backend/sox_io_backend.py

Lines 97 to 101 in d25a4dd

	Returns:
	torch.Tensor:
	If the input file has integer wav format and normalization is off, then it has
	integer type, else ``float32`` type. If ``channels_first=True``, it has
	``[channel, time]`` else ``[time, channel]``.

The return types of load functions should be tuple of [Tensor, int]. The legacy backends have the accurate description.

[mthrok]

Issue:

audio/torchaudio/backend/_soundfile_backend.py

Lines 103 to 107 in d25a4dd

	Returns:
	torch.Tensor:
	If the input file has integer wav format and normalization is off, then it has
	integer type, else ``float32`` type. If ``channels_first=True``, it has
	``[channel, time]`` else ``[time, channel]``.

The return types of load functions should be tuple of [Tensor, int]. The legacy backends have the accurate description.