uniovi-simur-wearablepermed-utils 1.27.0

Uniovi Simur WearablePerMed Utils

Description

Uniovi Simur WearablePerMed Utils

Schaffolding

Execute PyScaffold command to create the project:

$ putup --markdown uniovi-simur-wearablepermed-utils -p wearablepermed_utils \
     -d "Uniovi Simur WearablePerMed Utils." \
     -u https://github.com/SiMuR-UO/uniovi-simur-wearablepermed-utils.git 

Create a virtual environment inside for your project and active it:

$ python -m venv .venv
$ source .venv/bin/activate

Install and upgrade tox automation project manager:

$ pip install --upgrade tox

Install and upgrade project modules:

$ pip install -U numpy pandas scipy openpyxl matplotlib

Code and Debugging

Install library modules:

$ pip install -r requirements.txt

Install module locally for debugg

$ pip install -e .

Save project requirements:

$ pip freeze > requirements.txt

Project management

Project commands for: test, clean, build, generate documentation or publish your library in pypi repository Don't forget update the version library from setup.cfg project build file:

$ tox
$ tox -e clean
$ tox -e build
$ tox -e docs
$ tox -e publish -- --repository pypi

Testing

• Convert binary file to csv:

  $ sensor_bin_to_csv \
  --bin-file /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_C.BIN \
  --csv-file /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_C.csv
  

• Segment all sensor csv files for each participant and correct deviation if exists: For participant csv files without some deviation error

  $ csv_to_segmented_activity \
  --csv-file /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_M.csv \
  --excel-activity-log /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_RegistroActividades.xlsx \
  --body-segment M \
  --output /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_seg_M.npz
  

  For participant csv files with some deviation error

  $ csv_to_segmented_activity \
  --csv-file /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_M.csv \
  --excel-activity-log /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_RegistroActividades.xlsx \
  --body-segment M \
  --output /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_seg_M.npz \
  --sample-init 16088120 \
  --start-time 17:40:00
  

• Window segmented files for convolution model likes The argument include-not-estructure-data can be included if you want add not estructure data

  $ segmented_activity_to_stack \
  --npz-file /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_seg_M.npz \
  --crop-columns 1:7 \
  --window-size 250 \
  --window-overlapping-percent 50 \
  --output /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_tot_M.npz
  

• Extract features from windowed segmented files for random forest model likes

  $ stack_to_features \
  --stack-file /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_tot_M.npz \
  --output /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1053/PMP1053_W1_tot_M_features.npz∫
  

• Partial aggregation for each participant datasets

  $ aggregate_windows_features \
  --dataset-folder /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1003 \
  --ml-models ESANN,RandomForest \
  --ml-sensors thigh,hip,wrist \
  --output-folder /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input/PMP1003
  

• Total aggregation for all participant datasets to train models

  $ model_aggregation \
  --dataset-folder /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/input \
  --output-folder /home/miguel/git/uniovi/simur/uniovi-simur-wearablepermed-utils/data/output \
  --case-id case_sample
  

Pipeline:

Note

This project has been set up using PyScaffold 4.6. For details and usage information on PyScaffold see https://pyscaffold.org/.