Extracting features gait

[Erikpostt]

PR contains

• Updates on imu_preprocessing.py with added functionality for windowing and feature engineering;
• Directory of .ipynb files for documentation purposes

Waiting with reviewing (and merging) until

• Currently waiting for merging of PR Refactoring #2 to resolve changes in imu_preprocessing.py;
• Notebooks are still missing expected output

[vedran-kasalica]

I recommend merging the main branch into this branch after PR #6 receives approval. This step will ensure accurate testing of the changes.

[Erikpostt]

@vedran-kasalica what do you recommend here now?

[vedran-kasalica]

I recommend merging the main branch into this branch after PR #6 receives approval. This step will ensure accurate testing of the changes.

@Erikpostt I would recommend this, but we can also do that together on Thursday.

[Erikpostt]

@vedran-kasalica what do you think of the proposed changes?

[vedran-kasalica]

@vedran-kasalica what do you think of the proposed changes?

The changes look good!
I tested it locally and it all runs well (although I do generate binaries that are different on the byte level). We should probably save the outputs of the Notebooks in a separate folder (and add it to .gitignore so we don't track them) to be able to compare the results to the expected values. When I run it now I override the test data used for comparison.

I will finish reviewing the code tomorrow. We will probably merge the PR and then set up some tests based on your Notebooks and start adding more modules to the toolbox. I have some suggestions (e.g., we probably don't need to store the field meta_filename in our _meta.json files), but I will make a separate PR for that later, so we can distinguish refactoring from this PR.

[vedran-kasalica]

@Erikpostt it was not clear to me what is the content of rotation_meta file. We don't have an intermediary file specified in the pipeline under that name, right?

[Erikpostt]

@Erikpostt it was not clear to me what is the content of rotation_meta file. We don't have an intermediary file specified in the pipeline under that name, right?

You are right Vedran, this is the gyro pre-processed (grey) block at the 2.preprocessed_data level. As to your comment, it would probably be a good idea to be consistent in the naming of the files.

[vedran-kasalica]

The changes look good. I tested the Notebooks locally and they run without issues.

Instead of requesting changes to the PR, I opened an issue (see #8). It would be good to address it soon after merging this PR.

I propose to merge the PR and create new PRs to address the follow-up refactoring.

@kretep once you review the PR, feel free to merge it.