Readable Code for Audio Signal Visualization

[lucifertrj]

The following code takes to many variables, and I have simplified the following code into more readable format.

rows = 3
cols = 3
n = rows * cols
fig, axes = plt.subplots(rows, cols, figsize=(16, 9))

for i in range(n):
  if i>=n:
    break
  r = i // cols
  c = i % cols
  ax = axes[r][c]
  ax.plot(example_audio[i].numpy())
  ax.set_yticks(np.arange(-1.2, 1.2, 0.2))
  label = label_names[example_labels[i]]
  ax.set_title(label)
  ax.set_ylim([-1.1,1.1])

to

plt.figure(figsize=(16,10))
for i in range(9):
  plt.subplot(3,3,i+1)
  audio_signal = example_audio.numpy()[i]
  plt.plot(audio_signal)
  plt.title(label_names[example_labels[i]])
  plt.yticks(np.arange(-1.2, 1.2, 0.2))
  plt.ylim([-1.1,1.1])

[github-actions]

Preview

Preview and run these notebook edits with Google Colab:

• site/en/tutorials/audio/simple_audio.ipynb

Rendered notebook diffs available on ReviewNB.com.

Format and style

Use the TensorFlow docs notebook tools to format for consistent source diffs and lint for style:

$ python3 -m pip install -U --user git+https://github.com/tensorflow/docs

$ python3 -m tensorflow_docs.tools.nbfmt notebook.ipynb

$ python3 -m tensorflow_docs.tools.nblint --arg=repo:tensorflow/docs notebook.ipynb

If commits are added to the pull request, synchronize your local branch: git pull origin audio_change

[markmcd]

The following code takes to many variables and I have simplified the following code into more readable format.

Thanks - I like your suggestion, but some of these variables are there to improve readability, so I have some requests:

• Leave the rows, cols and n assignments, and make sure they work. It's consistent with the rest of the document, is clearer, and allows readers to tweak to fit their screen.
• Leave example_audio[i].numpy() (or just example_audio[i] should work in this context)
• Add spaces after every comma

[lucifertrj]

Hi @markmcd,

Thank you for the review. By code readable, I mainly meant simplifying r = i // cols and c = i % cols with plt.subplot(). I have kept n, rows and cols. And I have made sure that it works. Here is the result:

[lucifertrj]

Also, no need to mention numpy(). This is included in both the code i.e., official docs and my previous commit. I have fixed it in latest commit.

[MarkDaoust]

Suggested change

	" plt.subplot(3, 3, i+1)\n",
	" plt.subplot(rows, cols, i+1)\n",

[lucifertrj]

sorry to miss this out. I shall update it

[lucifertrj]

@MarkDaoust, I updated (3, 3) to (rows, cols). Sorry, I didn't notice that you already reviewed the PR

[8bitmp3]

@lucifertrj I formatted the notebook for you again, given the large diff because of the meta data.

TODO for you:

1. Please remove the cell outputs, and then
2. Format the notebook again.

This guide here explains how apply the tensorflow-docs nbfmt tool to any IPYNB file to make sure the JSON is cleaner, and passes the tests: https://www.tensorflow.org/community/contribute/docs#notebook_formatting

TL;DR suggested steps:

1. In a separate Python environment (with venv/virtualenv), or in the same one where you have TensorFlow installed, install the tensorflow-docs package: python3 -m pip install git+https://github.com/tensorflow/docs.
2. Run: python3 -m tensorflow_docs.tools.nbfmt simple_audio.ipynb

The output should then say:

Format notebook: simple_audio.ipynb

Then you can git add, commit, and push to your branch.

If you have questions, feel free to ask 👍

[lucifertrj]

@lucifertrj I formatted the notebook for you again, given the large diff because of the meta data.

TODO for you:

1. Please remove the cell outputs, and then
2. Format the notebook again.

This guide here explains how apply the tensorflow-docs nbfmt tool to any IPYNB file to make sure the JSON is cleaner, and passes the tests: https://www.tensorflow.org/community/contribute/docs#notebook_formatting

TL;DR suggested steps:

1. In a separate Python environment (with venv/virtualenv), or in the same one where you have TensorFlow installed, install the tensorflow-docs package: python3 -m pip install git+https://github.com/tensorflow/docs.
2. Run: python3 -m tensorflow_docs.tools.nbfmt simple_audio.ipynb

The output should then say:

Format notebook: simple_audio.ipynb

Then you can git add, commit, and push to your branch.

If you have questions, feel free to ask +1

Hi @8bitmp3, I apologise for taking a two days gap. I was busy with my internship work. I will complete the TODO assigned to this PR tomorrow without fail.

[lucifertrj]

Done @8bitmp3

[lucifertrj]

@markmcd @MarkDaoust

The following code does nothing in the simple audio documentation, and train_ds.take(1) doesn't require a break. Because this will return just a single batch/

for example_spectrograms, example_spect_labels in train_spectrogram_ds.take(1):
  break

The above code can be modified into the following:

for example_spectrograms, example_spect_labels in train_spectrogram_ds.take(1):
  print(example_spectrograms.shape)
  print(example_spect_labels.shape)

Should I make the changes, let me know what you think.

[markmcd]

The following code does nothing in the simple audio documentation, and train_ds.take(1) doesn't require a break. Because this will return just a single batch/

for example_spectrograms, example_spect_labels in train_spectrogram_ds.take(1):
  break

This code does do something - try removing it, restarting and running the notebook to see what happens. .take(1) returns a dataset, and the for-loop calls its __iter__ method, which materialises rows from the dataset.

You can see in the next cell that example_spectrograms is used. Without this loop, that variable would not exist.

Another way this could be written that might avoid this confusion is:

example_spectrograms, example_spect_labels = next(train_spectrogram_ds.take(1))

This is pretty explicit about what's happening, but next() may be a little harder to grok than a for-loop for someone newer to Python.

[lucifertrj]

to grok than a for-loop for someone newer to Python.

Got it thanks

[lucifertrj]

@8bitmp3 , the PR is still showing Code Owner Review required. I have made the necessary changes.

[MarkDaoust]

next(train_spectrogram_ds.take(1))

I think that also needs an iter in there.

[MarkDaoust]

Thanks. LGTM.