Med Cabinet Strain Recommender API

Small back - end Flask API model providing marijuana strain recommendations for the Med Cabinet project, based on desired effects and ailments to be treated.

Deployment

Running Flask app locally:

# Unix:
FLASK_APP = app flask run

# Windows:
export FLASK_APP = app  # one-time thing, to set the env var
flask run

Migrate the db:

FLASK_APP = app flask db init
FLASK_APP = app flask db migrate
FLASK_APP = app flask db upgrade

Local Environment Access:

http: // 127.0.0.1: 5000/

Running Flask app on heroku:

heroku login

Creating a new application server(MUST BE DONE FROM WITHIN THE REPOSITORY'S ROOT DIRECTORY):

git remote - v
heroku create  # optionally provide a name... "heroku create medi-cabinet"
git remote - v

Deploying to production:

git push heroku master
# or... git push heroku my_branch:master

Viewing production app in browser:

heroku open

Checking production server logs:

heroku logs - -tail

Provisioning production database:

```sh
heroku config
heroku addons: create heroku - postgresql: hobby - dev
# > provisions a new DATABASE_URL
heroku config

Migrating the production database:

# first login to the server, then run the migration commands there:
heroku run bash
# ... FLASK_APP=app flask db init
# ... FLASK_APP=app flask db migrate
# ... FLASK_APP=app flask db upgrade

# that should work, but alternatively you might be able to run these
# detached commands (if you didn't ignore your migrations dir):
heroku run "FLASK_APP=app flask db init"
heroku run "FLASK_APP=app flask db stamp head"
heroku run "FLASK_APP=app flask db migrate"
heroku run "FLASK_APP=app flask db upgrade"

Usage

Raw data output

Endpoint returning raw tables from the postgreSQL DB:

/cabinet  # raw output from cabinet table

**Parameters: ** None

**Returns: ** JSON array containing available strain information

Example: https://medi-cabinet.herokuapp.com/cabinet

Recommended Strains

Endpoint to return a list of recommendations.

/recommend

**Parameters: ** Passing a POST request to the endpoint with an JSON object that looks like:

 {
        "effects": ["happy", "euphoric", "creative"],
        "ailments": ["anxiety", "depression", "pain"],
        "negatives": ["dry mouth", "paranoid", "dizzy"]
    }

**Returns: ** JSON array containing strain id and n recommendations.

Example:

[
    {
        "id": "72"
    },
    {
        "id": "0"
    },
    {
        "id": "33"
    },
    {
        "id": "169"
    },
    {
        "id": "988"
    },
    {
        "id": "403"
    },
    {
        "id": "55"
    },
    {
        "id": "390"
    },
    {
        "id": "881"
    },
    {
        "id": "683"
    }
]

Model

Machine Learning model to recommend cannabis strains based on user input.

Full documentation and data and source files on the model can be found here: [ml - engineering](https: // github.com / MediCabinet / ml - engineering)

DATA

Sources:

• [Kushy API](https: // raw.githubusercontent.com / kushyapp / cannabis - dataset / master / Dataset / Strains / strains - kushy_api.2017 - 11 - 14.csv)
  • Provides chemical composition of strains
• [Kaggle / Leafly](https: // www.kaggle.com / kingburrito666 / cannabis - strains)
  • Provides strain name, type, rating, effects, taste, and description
• Data Scraped from Leafly
  • Provides a rating for each strain regarding specific ailments, negative side effects, and postive effects a user may want to take into account

MACHINE LEARNING MODEL

K - Nearest - Neighbor model takes a pandas series holding user input regarding their cannabis strain preferences and what is most important to them, and outputs a list of its nearest neighbors - most similar strains.

Inputs:

• Type of strain a user is looking for (hybrid, indica, sativa)
• Desired effects(creative, energetic, euphoric, focused, happy, hungry)
• Ailments they may be looking for relief from (anxiety, depression, fatigue, headaches, lack of appetite, pain, stress)
• Negative side effects they are trying to avoid(anxious, dizzy, dry eyes, dry mouth, headache, paranoid)

Testing

Flask API functionality was verified using Postman.

Project Information

[Product Vision Document](https: // docs.google.com / document / d / 1PNvyYa1qH1uxq - YKAhYnAPhT5jSBBE3XgYDzgQpFIUE / edit # heading=h.p0mtiic9v46n)

[Med Cabinet Project Pitch and Rubrics](https: // www.notion.so / Med - Cabinet - 7960b90bb485430483bb266f7b738308)

Links:

• [MediCabinet](https: // github.com / MediCabinet)

• [Marketing](https: // github.com / MediCabinet / marketing)

• [ML - Engineering](https: // github.com / MediCabinet / ml - engineering)

• [Data - Engineering](https: // github.com / MediCabinet / data - engineering)

• [Front - End](https: // github.com / MediCabinet / front - end)

• [Back - End](https: // github.com / MediCabinet / backend)