ASSESS (Automated SignalS Evaluation Service System)¶
ASSESS allows you to assess the signal in your dataset, understand the important variables and deploy your model in seconds.
Command line interface¶
Training a model¶
python train.py
--label "label"# the name of the target variable.
--path ../data/iris2.csv # path to the dataset csv file.
--explainmodel # optional argument, to store plots explaining the model' decisions during the predictive process.
example:
python train.py --target "label" --path "../data/iris2.csv" --explainmodel
Monitoring the model via the MLflow UI¶
Place yourself in the ASSESS/src directory then:
mlflow ui
Generate predictions¶
python predict.py --path ../data/iris2.csv
Deploy the model as a local REST API¶
from the ASSESS directory:
mlflow models serve -m "src\mlruns\0\\ :raw-html-m2r:`<model id>`\ \artifacts\sk_model" -p 1234
HTTPs query from a Python script¶
# ... define X as the pandas dataframe containing the observations
http_data = X.to_json(orient='split')
host = '127.0.0.1'
port = '1234'
url = f'http://{host}:{port}/invocations'
headers = {'Content-Type': 'application/json'}
r = requests.post(url=url, headers=headers, data=http_data)
return r.text