GPS trace analysis for nautical activities

analyze speed performance, extracts doppler speed information out of .gpx session files.

Features:

• n * vmax over x seconds
• n * vmax over x meter
• vmin of n * best jibes
• vmoy > vmin (average cruising speed)
• total distance > vmin (planning distance)
• % distance > vmin on distance (true planning ratio)
• % distance > vmin on time (planning ratio on time)

command line use

Install dependencies

pip3 install -r requirements.txt

Analyse a local gpx file

python3 src/core/gps_analysis.py -f author_filename.gpx

author_filename.gpx is the gps session file that you want to analyse.

Options

• -f loop over several gpx files:

  python3 src/core/gps_analysis.py -f file_1.gpx file_i.gpx ...

  make sure that the different files have different author_... prefix

• -rd loop recursively over all gpx files of a given directory and subdir:

  python3 src/core/gps_analysis.py -rd directory_name

• -p plot speed, distance and course (orientation) results

• -c crunch data with matplotlib graphs. Use history results from csv_results/all_results.csv and can be run without a gpx file

The results of all the gpx files are ranked and aggregated in the same ranking_results.csv file (see output)

REST API

Install dependencies and run the server (Flask based):

pip3 install -r requirements.txt

pip3 install -r requirements_flask.txt

python3 src/core/flask_restplus_server.py

Go to http://127.0.0.1:9999/ to get the REST API documentation from Swagger:

2 endpoints are currently available:

• /fetch_gpx_file/<path:file_url>
  to fetch and analyse file @ file_url
  you can test it with the swagger, or

  curl -X POST http://127.0.0.1:9999/gpsana/fetch_gpx_file/<file_url>?support=windsurf -H 'accept: application/json'

  in the url text, use '%3A' for ':' and '%2F' for '/'
  
  

• /upload_gpx_file
  to upload file directly on the server
  

  curl -X POST http://localhost:9999/gpsana/upload_gpx_file?support=windsurf -H "accept: application/json" -F "file=@<file_path>"

Running REST API server in a Docker container

• Build image:

  docker build -t gpsana_web .

• run the image:

  docker run -it --rm -p 8080:8080 gpsana_web

Access the server:

    http://localhost:8080/

Configuration

Analysis parameters can be modified in /config/config.yaml file

Output

The results to report are defined in the config.yaml file and are fully parametrizable:

• list of functions to call
• each function can be called several times with different args
• description to put in the final report (result.csv)
• ranking_group to calculate overall ranking based on groups
• if the yaml config file is modified, the all time ranking csv reports are reset (a all_time_results.old copy is saved)

Leveraging pandas DataFrame:

• format html so that doppler speed data is recognized
• import to pandas df all gps fields, including doppler speed
• reindex and resample data
• filter against glitches and orientation 360° turns

• plot and save results to csv files in csv_results/ directory:

  • debug.csv: full DataFrame of the gpx file
    => deep debug only => erased after each gpx file processing (not for use with a list of gpx file to process)

  • filename_result_debug.csv: DataFrame zoom on the high scores of the session
    => debug only (not for presentation)
    

  • filename_result.csv: individual result summary for the submitted gpx file
    => sessions summary
    

  • all_time_results.csv: swapping history file to record all time sessions history
    => program internal use (not for presentation)
    => updated at each run

  • ranking_results.csv: all time sessions or users history with ranking
    => overall results presentation
    => erased at each run
    

  • execution.log: full logging of the run with setlevel=INFO