chore: bump version to 1.0.16

mlModelSaver/__init__.py

@@ -77,16 +77,16 @@ class MlModelSaver:
         self.modelsFolder = f'{self.baseRelativePath}/{config.get('modelsFolder', '~~modelsFolder')}'
         ensure_directory_exists(self.modelsFolder)
 
-    def listOfPickels(self):
+    def listOfPickles(self):
         files = os.listdir(self.modelsFolder)
-        pickelsList = [file for file in files if file.endswith('.pkl')]
+        picklesList = [file for file in files if file.endswith('.pkl')]
-        return pickelsList
+        return picklesList
 
     def listOfModels(self):
-        pickelsList = self.listOfPickels()
+        picklesList = self.listOfPickles()
         modelsList = []
-        for pickekFileName in pickelsList:
+        for pickleFileName in picklesList:
-            modelsList.append(pickekFileName.split(".pkl")[0])
+            modelsList.append(pickleFileName.split(".pkl")[0])
         return modelsList
 
 
@@ -95,6 +95,12 @@ class MlModelSaver:
         supported_keys = [key for key, value in supportedModels.items() if value.get('supported')]
         return supported_keys
 
+    def loadModelByName(self, modelName):
+        filename = f'{self.modelsFolder}/{modelName}.pkl'
+        loaded_model = pickle.load(open(filename, 'rb'))
+        self.cachedModels[loaded_model.mlModelSaverConfig.get("modelName")] = loaded_model
+        return loaded_model
+
     def exportModel(self, model, config):
         transformer = config.get("transformer", default_transformer)
         model.mlModelSaverTransformer = transformer
@@ -111,8 +117,12 @@ class MlModelSaver:
         model.mlModelSavePredict = partial(mlModelSavePredict, model)
         filename = f'{self.modelsFolder}/{modelName}.pkl'
         pickle.dump(model, open(filename, 'wb'))
-        loaded_model = pickle.load(open(filename, 'rb'))
+        return self.loadModelByName(modelName)
-        self.cachedModels[loaded_model.mlModelSaverConfig.get("modelName")] = loaded_model
+
-        return loaded_model
+    def getModel(self, modelName):
+        model = self.cachedModels.get(modelName, None)
+        if model != None:
+            return model
+        return self.loadModelByName(modelName)
 
 

package.json

@@ -1,6 +1,6 @@
 {
   "name": "mlModelSaver",
-  "version": "1.0.15",
+  "version": "1.0.16",
   "description": "Make life easier for save and serving ml models",
   "main": "index.js",
   "repository": "git@github.com:smartdev-ca/mlModelSaver.git",

pytests/test_mlModelSaver.py

@@ -1,128 +1,82 @@
-import pickle
+# test_mlModelSaver.py
-import json
 
+import sys
 import os
 
-from functools import partial
+sys.path.insert(
-
+    0,
-def ensure_directory_exists(directory_path):
+    os.path.abspath(
-    """
+        os.path.join(
-    Ensure that the specified directory exists. If it doesn't, create it.
+            os.path.dirname(__file__),
-
+            '..'
-    Parameters:
+        )
-    directory_path (str): The path of the directory to ensure exists.
+    )
-    """
+)
-    os.makedirs(directory_path, exist_ok=True)
 
 
-def check_file_exists(file_path):
+def test_ensureCLassInstance():
-    """
+    from mlModelSaver import MlModelSaver
-    Check if the specified file exists.
+    mlModelSaverInstance1 = MlModelSaver({
-
+        "baseRelativePath": "test_baseRelativePath",
-    Parameters:
+        "modelsFolder": "test_modelsFolder"
-    file_path (str): The path of the file to check.
+    })
-
+    assert mlModelSaverInstance1.baseRelativePath == "test_baseRelativePath"
-    Returns:
+    assert mlModelSaverInstance1.modelsFolder == "test_baseRelativePath/test_modelsFolder"
-    bool: True if the file exists, False otherwise.
+    tesSupportedModels = mlModelSaverInstance1.showSupportedModels()
-    """
+    assert tesSupportedModels == ['sm.OLS']
-    if os.path.isfile(file_path):
-        print(f"File '{file_path}' exists.")
-        return True
-    else:
-        print(f"File '{file_path}' does not exist.")
-        return False
 
 
-supportedModels = {
+def test_OLS_LinearRegression():
-    "sm.OLS": {
+    from mlModelSaver import MlModelSaver
-        "supported": True
+    import numpy as np
-    }
+    import pandas as pd
-}
+    import statsmodels.api as sm
-
+    from helpers import add_constant_column
-supportedDataType = {
+    salaryMisDf = pd.read_excel("./datasets/Salary_MIS.xlsx")
-    "int": {
+    salaryBasedOnGpaMisStatistics = sm.OLS(
-        "supported": True
+        salaryMisDf["Salary"],
-    },
+        add_constant_column(salaryMisDf[["GPA", "MIS", "Statistics"]])
-    "float": {
+    )
-        "supported": True
+    salaryBasedOnGpaMisStatisticsFit = salaryBasedOnGpaMisStatistics.fit()
-    },
+    mlModelSaverInstance2 = MlModelSaver({
-    "binary":{
+        "baseRelativePath": ".",
-        "supported": True
+        "modelsFolder": "~~tmp/testModels"
-    }
+    })
-}
-
-def default_transformer(x):
-    return x
-
-
-def mlModelSavePredict(self, df, typeOfPredict = 'normal'):
-    dfAfterTransformation = self.mlModelSaverTransformer(df)
-    output = []
-    outputsName = self.mlModelSaverConfig.get("outputs", [{"name": "result"}])
-    outputsName = [item["name"] for item in outputsName]
-    if typeOfPredict == 'normal':
-        results = self.predict(dfAfterTransformation)
-        for  value in results:
-            output.append({
-                outputsName[0]: value,
-            })
-        return output
-
-class MlModelSaver:
-
-    cachedModels = {}
-
-    def __init__(self, config):
-        self.baseRelativePath = config.get('baseRelativePath', '.')
-        self.modelsFolder = f'{self.baseRelativePath}/{config.get('modelsFolder', '~~modelsFolder')}'
-        ensure_directory_exists(self.modelsFolder)
-
-    def listOfPickles(self):
-        files = os.listdir(self.modelsFolder)
-        picklesList = [file for file in files if file.endswith('.pkl')]
-        return picklesList
-
-    def listOfModels(self):
-        picklesList = self.listOfPickles()
-        modelsList = []
-        for pickleFileName in picklesList:
-            modelsList.append(pickleFileName.split(".pkl")[0])
-        return modelsList
 
 
 
-    def showSupportedModels(self):
+    loadedModel = mlModelSaverInstance2.exportModel(
-        supported_keys = [key for key, value in supportedModels.items() if value.get('supported')]
+        salaryBasedOnGpaMisStatisticsFit,
-        return supported_keys
+        {
-
+            "modelName": "salaryBasedOnGpaMisStatistics",
-    def loadModelByName(self, modelName):
+            "description": "Predict Salary based on GPA MIS Statistics for salaryMisDf",
-        filename = f'{self.modelsFolder}/{modelName}.pkl'
+            "modelType": "sm.OLS",
-        loaded_model = pickle.load(open(filename, 'rb'))
+            "inputs": [
-        self.cachedModels[loaded_model.mlModelSaverConfig.get("modelName")] = loaded_model
+                {
-        return loaded_model
+                    "name": "GPA",
-
+                    "type": "float",
-    def exportModel(self, model, config):
+                },
-        transformer = config.get("transformer", default_transformer)
+                {
-        model.mlModelSaverTransformer = transformer
+                    "name": "MIS",
-        if "transformer" in config:
+                    "type": "binary"
-            del config["transformer"]
+                },
-        model.mlModelSaverConfig = config
+                {
-        isModelSupporter = supportedModels.get(
+                    "name": "Statistics",
-            config.get("modelType", ''),
+                    "type": "binary"
-            {}
+                }
-        ).get("supported", False)
+            ],
-        if not isModelSupporter:
+            "transformer": add_constant_column,
-            raise ValueError(f'only {self.showSupportedModels()} are supported and {config.get("modelType", '')} is not supported')
+            "outputs": [
-        modelName = model.mlModelSaverConfig['modelName']
+                {
-        model.mlModelSavePredict = partial(mlModelSavePredict, model)
+                    "name": "Salary",
-        filename = f'{self.modelsFolder}/{modelName}.pkl'
+                    "type": "int"
-        pickle.dump(model, open(filename, 'wb'))
+                }
-        return self.loadModelByName(modelName)
+            ]
-
+        }
-    def getModel(self, modelName):
+    )
-        model = self.cachedModels.get(modelName, None)
+    from mlModelSaver import check_file_exists
-        if model != None:
+    assert check_file_exists("./~~tmp/testModels/salaryBasedOnGpaMisStatistics.pkl") == True
-            return model
+    testData = salaryMisDf[["GPA", "MIS", "Statistics"]].iloc[0:2]
-        return self.loadModelByName(modelName)
+    predictedValueWithLoadedModel = loadedModel.mlModelSavePredict(testData, 'normal')
-
+    assert  predictedValueWithLoadedModel == [{'Salary': 73.9924679451542}, {'Salary': 69.55525482441558}]
-
+    assert list(mlModelSaverInstance2.cachedModels.keys()) == ['salaryBasedOnGpaMisStatistics']

setup.py

@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
 
 setup(
     name='mlModelSaver',
-    version='1.0.15',
+    version='1.0.16',
     packages=find_packages(),
     description='Make life easier for saving and serving ML models',
     long_description=open('DOCS.md').read(),  # Assumes you have a README.md file