This Django app provides various API endpoints for users to deploy and manage ML models.
Install Docker and Docker Compose:
- Docker version 26.0.0+ and Docker Compose version 2.28.0+ are required.
- For Ubuntu, refer to the following link: https://docs.docker.com/engine/install/ubuntu/.
- If you're using Debian or Ubuntu, you can use the bash scripts in the
/helpersdirectory to install Docker and Docker Compose:bash helpers/install_docker.sh bash helpers/install_docker_compose.sh
-
Create Configuration Files:
- Create the following files in the root directory:
dev.envprod.envsuperusers.json
- Use the respective
*.distfiles as templates for these new files.
- Create the following files in the root directory:
-
Generate Django Secret Key:
- Visit https://djecrety.ir/ or any other suitable tool.
- Generate a new Django secret key.
-
Set Django Secret Key:
- Open
dev.envandprod.envfiles. - Set the value of
DJANGO_SECRET_KEYto the newly obtained secret key.
- Open
-
Set Admin Email and Site Name:
- Open
dev.envandprod.envfiles. - Set the value of
DJANGO_SITE_NAMEto the desired site name. - Set the value of
DJANGO_SITE_DOMAINto the desired site domain. - Set the value of
ADMIN_EMAILto the desired admin email address.
- Open
-
Set Allowed Hosts (for production environment):
- Open the
prod.envfile. - Set the value of
DJANGO_ALLOWED_HOSTSto the desired host address(es).- Example:
DJANGO_ALLOWED_HOSTS=example.com,www.example.com
- Example:
- Open the
-
Set Database Password:
- Open the
prod.envfile. - Set the value of
DB_PASSWORDto the desired database password.
- Open the
-
Set New Admin Account Credentials:
- Open the
superusers.jsonfile. - Set a new email and password for the admin account in the JSON format.
- Open the
-
Encrypt/ decrypt *.env/ superusers.json files (optional):
bash helpers/gpg_wrapper.sh -f *.env -e encrypt bash helpers/gpg_wrapper.sh -f *.env.gpg -e decrypt
-
Update the included django utilities (optional):
git fetch https://github.com/buswedg/django-utils.git main git subtree pull --prefix=apps/utils https://github.com/buswedg/django-utils.git main --squash
-
Update the included playgrounds (optional):
git fetch https://github.com/buswedg/income_classifier.git main git subtree pull --prefix=playground/income_classifier https://github.com/buswedg/income_classifier.git main --squash
-
Run locally for development:
docker compose -f docker-compose.dev.yml --env-file dev.env up
-
Run for production (without ssl):
docker compose -f docker-compose.prod.yml --env-file prod.env up
-
Run for production (with ssl):
docker compose -f docker-compose.prod.nginx.yml --env-file prod.env up
To inspect emails sent by the app, browse to http://127.0.0.1:8025/
Grab access and refresh tokens:
import requests
url = "http://127.0.0.1:8000/api/token/"
data = {
"username": "buswedg",
"password": "$PASS"
}
response = requests.post(url, data=data)
data = response.json()
access_token = data.get('access')
refresh_token = data.get('refresh')
headers = {'Authorization': f'JWT {access_token}'}Check endpoints:
url = "http://127.0.0.1:8000/api/v1/endpoints/"
response = requests.get(url, headers=headers)
data = response.json()
print(data)
[{'id': 1,
'name': 'income_classifier',
'owner': 'buswedg',
'created_at': '2023-06-08T23:58:53.287663Z'}]Check ml algorithms:
url = "http://127.0.0.1:8000/api/v1/ml-algorithm/"
response = requests.get(url, headers=headers)
data = response.json()
print(data)
[{'id': 1,
'name': 'random forest',
'description': 'Random Forest with simple pre- and post-processing.',
'code': "class RFClassifier(Helpers):\n def __init__(self):\n self.values_fill_missing = joblib.load(os.path.join(ARTIFACTS_PATH, 'train_mode.joblib'))\n self.encoders = joblib.load(os.path.join(ARTIFACTS_PATH, 'encoders.joblib'))\n self.model = joblib.load(os.path.join(ARTIFACTS_PATH, 'rf_classifier.joblib'))\n",
'version': '0.1',
'owner': 'buswedg',
'created_at': '2023-06-08T23:58:53.304885Z',
'parent_endpoint': 1,
'current_status': 'production'},
{'id': 2,
'name': 'extra trees',
'description': 'Extra Trees with simple pre- and post-processing.',
'code': "class ETClassifier(Helpers):\n def __init__(self):\n self.values_fill_missing = joblib.load(os.path.join(ARTIFACTS_PATH, 'train_mode.joblib'))\n self.encoders = joblib.load(os.path.join(ARTIFACTS_PATH, 'encoders.joblib'))\n self.model = joblib.load(os.path.join(ARTIFACTS_PATH, 'et_classifier.joblib'))\n",
'version': '0.1',
'owner': 'buswedg',
'created_at': '2023-06-08T23:58:54.187111Z',
'parent_endpoint': 1,
'current_status': 'testing'}]Check ml algorithm status:
url = "http://127.0.0.1:8000/api/v1/ml-algorithm-status/"
response = requests.get(url, headers=headers)
data = response.json()
print(data)
[{'id': 1,
'active': True,
'status': 'production',
'created_by': 'buswedg',
'created_at': '2023-06-08T23:58:53.309248Z',
'parent_mlalgorithm': 1},
{'id': 2,
'active': True,
'status': 'testing',
'created_by': 'buswedg',
'created_at': '2023-06-08T23:58:54.190485Z',
'parent_mlalgorithm': 2}]Generate a prediction using a ml algorithm:
url = "http://127.0.0.1:8000/api/v1/predict/income_classifier/?status=production&version=0.1"
payload = {
"age": 37,
"workclass": "Private",
"fnlwgt": 34146,
"education": "HS-grad",
"education-num": 9,
"marital-status": "Married-civ-spouse",
"occupation": "Craft-repair",
"relationship": "Husband",
"race": "White",
"sex": "Male",
"capital-gain": 0,
"capital-loss": 0,
"hours-per-week": 68,
"native-country": "United-States"
}
response = requests.post(url, headers=headers, json=payload)
data = response.json()
print(data)
{'probability': 0.07,
'label': '<=50K',
'status': 'OK',
'request_id': 1}