/Music-Meteorologist-ds

Primary LanguageTSQLMIT LicenseMIT

Production site: https://www.sound-drip.com/

There's a lot going on here, so we'll let you know where things are.

Top Level

  • data_collection is a directory where we aggregated all the spotify song data
  • model_notebooks is a directory where all current and former knn models are found. There is also a notebook where we tried DBSCAN
  • model_validation is a directory used to test out song outputs from our model for listening as part of the model validation step
  • Flask_Notebooks is our directory for notebooks used for staging production code prior to deploying changes to flask
  • spotipy is our directory where we explored the Spotify API extensively in preparation for API calls
  • RDS-postgres/SOUNDDRIP is our sandbox for testing out new updates to our RDS-hosted postgrSQL tables
  • Flask_AWS/SOUNDDRIP is our production flask app

More information can be found on our notion doc: https://www.notion.so/For-Current-DS-145f105cd7a1459bbc955a97624ecc60

API INPUTS/OUTPUTS

Root endpoints utilized on local flask deployments

  • Root are the local APIs
  • SD DS Prod refers to the main one being consumed by the application

Information regarding the API calls made between front-end and DS can be found here: https://documenter.getpostman.com/view/10161796/SzRuYroD?version=latest#intro

Local Deployment

Environment manager documentation can be found here: https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/

  1. Install pip's preferred virtual environment manager
python3 -m pip install --user virtualenv
  1. Access Flask_AWS/SOUNDDRIP
  2. Create environment, replacing env with your preferred environment name
python3 -m venv env
  1. Activate environment
source env/bin/activate
  1. Install all dependencies from the requirements.txt file located in the current directory
pip3 install -r requirements.txt
  1. Download all compressed file object dependencies from S3 in this directory

https://s3.console.aws.amazon.com/s3/buckets/sound-drip/Flask-Prod-Files/?region=us-east-1&tab=overview

  1. Copy following files to ./models:

    • genres_array_2.pkl
    • model5.joblib
    • scalar3.joblib
    • slider_model6.joblib

  2. Place song_id_array3.pkl in a new data folder after creating it under the SOUNDDRIP directory

  3. Create env_vars.py with the necessary database credentials under ./misc/

  4. Run the application

python3 application.py

  • NOTE: If you get an error that a particular package is missing even though you have confirmed it is installed, you will need to pip install those packages utilizing an optional target argument:

    • You can find your virtual enviroment's target directory by accessing python3 shell within your virtual environment and running sys.path
pip install [package name] --target [target directory]