0.26.0 regression-floating-error & flake

0.26.0 add regression floating error and abs error in integrated algo

.github/pull_request_template.md

@@ -7,4 +7,4 @@
 ## Checklist before requesting a review
 - [ ] I have performed a self-review of my code
 - [ ] I have increased the version number in the `frontend/package.json` and `backend/swagger.yaml` files
-- [ ] I have checked that Black, Prettier and Cspell have been run on my code
+- [ ] I have checked that Black, Flake8, Prettier and Cspell are passing

.github/workflows/pull-request-check.yml

@@ -24,6 +24,22 @@ jobs:
           pip install black
           black --check --diff .
 
+  flake8-check: # Check that the backend codebase does not contain linting errors
+    name: Flake8 check
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v2
+      - name: Set up Python 3.8
+        uses: actions/setup-python@v1
+        with:
+          python-version: 3.8
+      - name: Install dependencies and check flake8 format
+        run: |
+          cd backend
+          python -m pip install --upgrade pip
+          pip install flake8
+          flake8 .
+
   prettier-check: # Check that the frontend codebase is formatted with prettier
     name: Prettier format check
     runs-on: ubuntu-latest
@@ -84,6 +100,12 @@ jobs:
           echo "PREVIOUS_VERSION=$PREVIOUS_VERSION"
           echo "BRANCH_VERSION=$BRANCH_VERSION"
 
+          # If pervious version is beta, ignore test
+          if [[ $PREVIOUS_VERSION == *"beta"* ]]; then
+            echo "Beta version detected, skipping version upgrade check."
+            exit 0
+          fi
+
           if [ "$BRANCH_VERSION" == "" ]; then
             echo "No version found in current branch."
             exit 1

backend/config/init_config.py

@@ -141,7 +141,7 @@ def get_config_values(section, config_parser):
     for key in os.environ.keys():
         if key.startswith(ENV_VAR):
             # Get the key name without the env var prefix
-            key_name = key[len(ENV_VAR) :]
+            key_name = key[len(ENV_VAR) :]  # noqa
             values[key_name] = str.lower(os.environ[key])
 
     return values
@@ -184,7 +184,7 @@ def init_config():
                 continue
 
             # Deal with booleans
-            if type(config[section][key]) == bool:
+            if type(config[section][key]) is bool:
                 if value == "false":
                     set_config_value(section, key, False)
                 elif value == "true":
@@ -199,7 +199,7 @@ def init_config():
                     continue
 
             # Deal with integers
-            elif type(config[section][key]) == int:
+            elif type(config[section][key]) is int:
                 try:
                     set_config_value(section, key, int(value))
                 except ValueError:
@@ -212,7 +212,7 @@ def init_config():
                     continue
 
             # Deal with strings
-            elif type(config[section][key]) == str:
+            elif type(config[section][key]) is str:
                 set_config_value(section, key, str(value))
 
         # Deal with list based config elements

backend/controller/projects.py

@@ -10,7 +10,7 @@
 
 
 def ping():
-    return "Online v0.15.1", 200
+    return "Online", 200
 
 
 def get_data_providers_project(dataProviderId):

backend/modules/algoProviders/AlgoProvider.py

@@ -1,5 +1,6 @@
 # Class for AlgoProvider
-import requests, json
+import requests
+import json
 from modules.algoProviders.AlgoProviderException import AlgoProviderException
 
 

backend/modules/algoProviders/integratedAlgoProvider/algorithms/classificationErrorMetric.py

@@ -12,7 +12,8 @@
 # Technical details (must respect the algo-api format):
 algorithm_description = {
     "name": "Classification Metric",
-    "description": """Calculates the classification error according to the ground truth and the predictions""",
+    "description": """Calculates the classification error according \
+to the ground truth and the predictions""",
     "author": "DebiAI",
     "version": "1.0.0",
     "creationDate": "2023-10-30",
@@ -26,21 +27,24 @@
         },
         {
             "name": "Predictions",
-            "description": "List of predictions, must have the same length as the ground truth list",
+            "description": "List of predictions, must have the same \
+length as the ground truth list",
             "type": "array",
             "arrayType": "text",
         },
     ],
     "outputs": [
         {
             "name": "Binary error",
-            "description": "Classification metric of the input list, False if GDT == PRED, True otherwise",
+            "description": "Classification metric of the input list, \
+False if GDT == PRED, True otherwise",
             "type": "array",
             "arrayType": "boolean",
         },
         {
             "name": "Binary success",
-            "description": "Classification metric of the input list, True if GDT == PRED, False otherwise",
+            "description": "Classification metric of the input list, \
+True if GDT == PRED, False otherwise",
             "type": "array",
             "arrayType": "boolean",
         },

backend/modules/algoProviders/integratedAlgoProvider/algorithms/regressionErrorMetric.py

@@ -12,7 +12,8 @@
 # Technical details (must respect the algo-api format):
 algorithm_description = {
     "name": "Regression Metric",
-    "description": """Calculates the regression error according to the ground truth, the predictions and a ceil value""",
+    "description": """Calculates the regression error according to the ground truth, \
+the predictions and a ceil value""",
     "author": "DebiAI",
     "version": "1.0.0",
     "creationDate": "2023-05-23",
@@ -26,7 +27,8 @@
         },
         {
             "name": "Predictions",
-            "description": "List of predictions, must have the same length as the ground truth list",
+            "description": "List of predictions, must have the same length as the  \
+ground truth list",
             "type": "array",
             "arrayType": "number",
         },
@@ -40,16 +42,38 @@
         },
     ],
     "outputs": [
+        {
+            "name": "Error",
+            "description": "Difference between the ground truth and the predictions",
+            "type": "array",
+            "arrayType": "number",
+        },
+        {
+            "name": "Absolute error",
+            "description": "Absolute value of the error",
+            "type": "array",
+            "arrayType": "number",
+        },
         {
             "name": "Binary error",
-            "description": "Regression metric of the input list, False if abs(GDT - PRED) < ceil, True otherwise",
+            "description": "True if Absolute error > ceil, False otherwise",
             "type": "array",
             "arrayType": "boolean",
         },
         {
             "name": "Error percentage",
             "type": "number",
         },
+        {
+            "name": "Binary success",
+            "description": "True if Absolute error <= ceil, False otherwise",
+            "type": "array",
+            "arrayType": "boolean",
+        },
+        {
+            "name": "Success percentage",
+            "type": "number",
+        },
     ],
 }
 
@@ -72,26 +96,54 @@ def use_algorithm(inputs):
         raise TypeError("Ground truth and predictions must have the same length")
 
     # Calculate regression metric
-    regressionMetric = []
-    for i in range(len(gdt)):
-        error = gdt[i] - predictions[i]
+    nb_values = len(gdt)
+    error = [None] * nb_values
+    absolute_error = [None] * nb_values
+    binary_error = [None] * nb_values
+    binary_success = [None] * nb_values
 
-        if abs(error) < ceil:
-            regressionMetric.append(False)
+    for i in range(nb_values):
+        error_value = gdt[i] - predictions[i]
+        error[i] = error_value
+        absolute_error[i] = abs(error_value)
+
+        if abs(error_value) > ceil:
+            binary_error[i] = True
+            binary_success[i] = False
         else:
-            regressionMetric.append(True)
+            binary_error[i] = False
+            binary_success[i] = True
 
-    percentage = regressionMetric.count(True) / len(regressionMetric)
-    percentage = round(percentage * 100, 2)
+    # Calculate percentages
+    error_percentage = binary_error.count(True) / nb_values
+    error_percentage = round(error_percentage * 100, 2)
+    success_percentage = binary_success.count(True) / nb_values
+    success_percentage = round(success_percentage * 100, 2)
 
     # Return outputs
     return [
+        {
+            "name": "Error",
+            "value": error,
+        },
+        {
+            "name": "Absolute error",
+            "value": absolute_error,
+        },
         {
             "name": "Binary error",
-            "value": regressionMetric,
+            "value": binary_error,
         },
         {
             "name": "Error percentage",
-            "value": percentage,
+            "value": error_percentage,
+        },
+        {
+            "name": "Binary success",
+            "value": binary_success,
+        },
+        {
+            "name": "Success percentage",
+            "value": success_percentage,
         },
     ]

backend/modules/algoProviders/integratedAlgoProvider/integratedAlgoProvider.py

@@ -37,7 +37,7 @@ def _get_algorithm_python(algorithm_name):
 
 
 class IntegratedAlgoProvider(AlgoProvider):
-    ### Integrated AlgoProvider
+    # Integrated AlgoProvider
     # Used to expose the algorithms that are integrated
     # directly in DebiAI
     def __init__(self):

backend/modules/dataProviders/pythonDataProvider/dataUtils/projects.py

@@ -97,7 +97,6 @@ def create_project(projectId, projectName):
     os.mkdir(DATA_PATH + projectId)
     os.mkdir(DATA_PATH + projectId + "/blocks")
     os.mkdir(DATA_PATH + projectId + "/models")
-    os.mkdir(DATA_PATH + projectId + "/requests")
     os.mkdir(DATA_PATH + projectId + "/selections")
 
     now = pythonModuleUtils.timeNow()

backend/modules/dataProviders/pythonDataProvider/dataUtils/pythonModuleUtils.py

@@ -40,7 +40,7 @@ def clean_filename(filename):
 
 
 def is_filename_clean(filename):
-    cleanFilename = "".join(i for i in filename if i not in "\/:*?<>|")
+    cleanFilename = "".join(i for i in filename if i not in "\/:*?<>|")  # noqa
     return filename == cleanFilename