implement testing using pytest framework

tests/test_main.py

@@ -1,16 +1,15 @@
-# data is properly loading from source or not
-# check build_features working or not    ---- done
-# check model training is working fine or not and model return kr rha he ki nahi
-# model tunning performance improve kr rha he ki nahi
-# check whether visualization is working fine or not, or plot return kr rha he ki nahi
-# check api working fine or not, response kya de rha he api glt or sahi input pr
-# check streamlit working fine or not,
-
 import pathlib
 import pytest
 import yaml
+import pandas as pd
+from fastapi.testclient import TestClient
+from sklearn.base import BaseEstimator
+from prod.api import app
 from src.features.build_features import feat_eng
 from src.data.make_dataset import load_data
+from src.data.load_dataset import extract_data
+from src.models.train_model import train_model
+from src.visualization.visualize import conf_matrix
 
 curr_dir = pathlib.Path(__file__)
 home_dir = curr_dir.parent.parent
@@ -20,7 +19,20 @@
 processed_data = params['make_dataset']['processed_data']
 data_source = f'{home_dir.as_posix()}{processed_data}/train.csv'
 
+drive_ = params['load_dataset']['drive_link']
+
+# test data loading functionality 
+@pytest.fixture(params = [drive_])
+def load_dataset(request) :
+     return extract_data(request.param)
+
+def test_load_data(load_dataset) : 
+     if isinstance(load_dataset, pd.DataFrame) : 
+          assert isinstance(load_dataset, pd.DataFrame)
+     else : 
+          pytest.fail('Unable to fetch data from remote server')
 
+# test feature generation functionality
 # 'request' is a fixture that provides information about the currently executing test.
 # It allows us to access various attributes and methods to get information about the test
 # environment, access fixtures, and more.
@@ -30,15 +42,138 @@ def load(request) :
      return feat_eng(load_data(request.param))
 
 def test_build_features(load) : 
-     assert load.shape[1] == 20
+     if load.shape[1] == 20 :
+          assert load.shape[1] == 20
+     else :
+          pytest.fail('build_feature is not working properly')
+
+parameters = params['train_model']
+TARGET = params['base']['target']
+
+train_dir = f"{home_dir.as_posix()}{params['build_features']['extended_data']}/extended_train.csv"
+test_dir = f"{home_dir.as_posix()}{params['build_features']['extended_data']}/extended_test.csv"
+model_dir = f"{home_dir.as_posix()}{parameters['model_dir']}"
+pathlib.Path(model_dir).mkdir(parents = True, exist_ok = True)
+
+train_data = load_data(train_dir)
+test_data = load_data(test_dir)
+X_train = train_data.drop(columns = [TARGET]).values
+Y = train_data[TARGET]
+
+X_test = test_data.drop(columns = [TARGET]).values
+y = test_data[TARGET]
+
+# test model training functionality
+@pytest.fixture
+def get_model() :
+     return train_model(X_train, Y, parameters['n_estimators'], parameters['criterion'], parameters['max_depth'], min_samples_leaf = parameters['min_samples_leaf'],
+                 min_samples_split = parameters['min_samples_split'], random_state = parameters['random_state'], yaml_file_obj = params)
+
+def test_train_model(get_model) :
+     if isinstance(get_model['model'], BaseEstimator) :
+          assert isinstance(get_model['model'], BaseEstimator)
+     else :
+          pytest.fail('not able to train the model')
 
+# test ploting functionality
+def test_conf_metrix(get_model) :
+     filename = conf_matrix(y_test = y, y_pred = get_model['model'].predict(X_test), labels = get_model['model'].classes_, path = './tests/trash', params_obj = params)
+     if pathlib.Path(filename).exists() :
+          assert pathlib.Path(filename).exists()
+     else :
+          pytest.fail('unable to generate confusion matrix')
 
+# test api endpoint
+# valid input
+user_input1 = {  
+                'fixed_acidity': 7.6,
+                'volatile_acidity': 0.715,
+                'citric_acid': 0,
+                'residual_sugar': 15.1,
+                'chlorides': 0.068,
+                'free_sulfur_dioxide': 30,
+                'total_sulfur_dioxide': 35,
+                'density': 0.99533,
+                'pH': 3.48,
+                'sulphates': 0.65,
+                'alcohol': '11.4'}
+
+# invalid input - remove few columns
+user_input2 = {  
+                'fixed_acidity': 9.9384,
+                'volatile_acidity': 0.6412,
+                'citric_acid': 0.4372,
+                'residual_sugar': 5.4216,
+                'chlorides': 0.0719,
+                'free_sulfur_dioxide': 7.2126,
+                'total_sulfur_dioxide': 12.7873,
+                'density': 0.997,
+                'pH': 3.255,
+                'sulphates': 0.7923 }
+
+# invalid input - values out-of-range
+user_input3 = {  
+                'fixed_acidity': 700.6,
+                'volatile_acidity': 0.715,
+                'citric_acid': 780,
+                'residual_sugar': 15.1,
+                'chlorides': 0.068,
+                'free_sulfur_dioxide': 3000,
+                'total_sulfur_dioxide': 335,
+                'density': 10.99533,
+                'pH': 33.48,
+                'sulphates': 0.65,
+                'alcohol': '11.4'}
+
+@pytest.fixture
+def client() :
+     return TestClient(app)
+
+# test response of api
+def test_response(client) :
+     response = client.get("/")
+     if response.status_code == 200 : 
+          assert response.status_code == 200
+          assert response.json() == {'api status': 'up & running'}
+     else : 
+          pytest.fail('api is not responding')
+
+# test predict endpoint with valid input
+def test_validIp(client) :
+     response = client.post("/predict", json = user_input1)
+     if response.status_code == 200 : 
+          assert response.status_code == 200
+          assert 'predictions' in response.json()
+          assert 'probability' in response.json()
+     else : 
+          pytest.fail('predict endpoint failed on valid input')
+
+# test predict with invalid inputs
+def test_invalidIp(client) :
+     response1 = client.post("/predict", json = user_input2)
+     response2 = client.post("/predict", json = user_input3)
+     if response1.status_code == 422 and response2.status_code == 422 :  
+          assert response1.status_code == 422
+          assert response2.status_code == 422
+     else :
+          pytest.fail('predict endpoint failed on invalid input')
+
+def test_details(client) :
+     response = client.get("/details")
+     if response.status_code == 200 : 
+          assert 'model name' in response.json()
+          assert 'model run_id' in response.json()
+          assert 'model version' in response.json()
+          assert 'model aliases' in response.json()
+     else :
+          pytest.fails('details endpoint not responding')
 
 '''
-tox.ini file initially khud se create karo or configure karo
-but ek cheez ka dhyan de ki agr test_files kisi dir me hoto kese tox.ini me mention krna he
-or khas baat requirements.txt dhyan se dena, local package bhi mention krna
-bss fix requirements.txt reload kr k testing kran hoto
-cmd: tox -r
-else cmd: tox
+create & configure tox.ini file
+Commands : 
+     cmd: tox 
+     cmd: tox -r (reload the packages)
 '''
+
+# pytest cmd to test single function
+# cmd: "pytest tests/test_main.py::test_build_features --verbose --disable-warnings"