Merge branch 'sap-dev' of github.com:Hestia-Homes/ML into sap-dev-model

.github/workflows/MLPipelinePostMerge.yml

@@ -42,7 +42,14 @@ jobs:
         if [ -z "${latest_version}" ]; then
           increment_version="1.0.0"
         else
-          increment_version=$(echo ${latest_version} | awk -F'.' '{OFS="."; $1+=1; print}')
+          increment_version=$(echo ${latest_version} | awk 'BEGIN {
+              FS="\\."   # Set the field separator to a period
+              OFS="."    # Set the output field separator to a period
+          }
+          {
+              major = $1 + 1   # Increment the major version
+              print major, "0", "0"   # Print the new version
+          }')
         fi
 
         new_tag=${REGISTER_MODEL_NAME}@v${increment_version}
@@ -80,7 +87,14 @@ jobs:
         if [ -z "${latest_version}" ]; then
           increment_version="0.1.0"
         else
-          increment_version=$(echo ${latest_version} | awk 'BEGIN{FS=OFS="."} {$2++; print}')
+          increment_version=$(echo ${latest_version} | awk 'BEGIN {
+              FS="\\."   # Set the field separator to a period
+              OFS="."    # Set the output field separator to a period
+          }
+          {
+              minor = $2 + 1   # Increment the minor version
+              print $1, minor, "0"   # Print the new version
+          }')
         fi
 
         new_tag=${REGISTER_MODEL_NAME}@v${increment_version}
@@ -118,7 +132,14 @@ jobs:
         if [ -z "${latest_version}" ]; then
           increment_version="0.0.1"
         else
-          increment_version=$(echo ${latest_version} | awk 'BEGIN{FS=OFS="."} {$3++; print}')
+          increment_version=$(echo ${latest_version} | awk 'BEGIN {
+              FS="\\."   # Set the field separator to a period
+              OFS="."    # Set the output field separator to a period
+          }
+          {
+              patch = $3 + 1   # Increment the patch version
+              print $1, $2, patch   # Print the new version
+          }')
         fi
 
         new_tag=${REGISTER_MODEL_NAME}@v${increment_version}
@@ -188,7 +209,7 @@ jobs:
         git config user.name "Github-Bot"
         git config user.email "Github-Bot@no-reply.com"
 
-        latest_dev_version=$(gto history ${REGISTER_MODEL_NAME} --asc --plain | awk '{print $NF}' | awk '/dev/')
+        latest_dev_version=$(gto history ${REGISTER_MODEL_NAME} --asc --plain | awk '{print $NF}' | awk '/dev/' | awk 'END {print}')
         if [ -z "${latest_dev_version}" ]; then
           increment_version="1"
         else
@@ -196,7 +217,7 @@ jobs:
         fi
 
         new_tag=${REGISTER_MODEL_NAME}#dev#${increment_version}
-        latest_version=$(gto show model@latest --ref | awk -F"@" '{print $2}')
+        latest_version=$(gto show ${REGISTER_MODEL_NAME}@latest --ref | awk -F"@" '{print $2}')
 
         echo ${new_tag}
 

MODEL_REGISTRY.md

@@ -8,9 +8,17 @@
         "active": true
     },
     "sap": {
-        "version": "v0.0.3",
+        "version": "v0.1.0",
         "stage": {
-            "dev": "v0.0.3"
+            "dev": "v0.1.0"
+        },
+        "registered": true,
+        "active": true
+    },
+    "heat": {
+        "version": "v0.0.1",
+        "stage": {
+            "dev": "v0.0.1"
         },
         "registered": true,
         "active": true

README.md

@@ -10,9 +10,9 @@ tracking and a model registry
 	- A bolt-on service that can implement model monitoring
 
 There are multiple protected branches which adapt the generic pipeline to produce different models:
-- sap_change-**
-- heat_change-**
-- carbon_change-**
+- sap-{dev/staging/prod}-**
+- heat-{dev/staging/prod}-**
+- carbon-{dev/staging/prod}-**
 
 These branches will differ by the configuration files that define the data used and the outputs of the ML-pipeline
 - There can be different additional logic for each branch but the pipeline will be the same.
@@ -31,7 +31,7 @@ In order for this to be set up, some key environment variables needs to be inser
 secrets. Each different model and protected branch has its own set of secrets which allows for flexibility
 between different pipelines.
 
-For example, for the branch sap_change-dev, the prefix=SAP_CHANGE_DEV, and the following secrets are:
+For example, for the branch sap-dev, the prefix=SAP_DEV, and the following secrets are:
 
 - {prefix}_ECR_URI, which is the URI of the ECR repository to push to. For example, for the
    sap change model this is the lambda-sap-prediction-dev repository.
@@ -58,7 +58,7 @@ First, navigate to the root directory of the repository. Open a terminal and exe
 2. command to build the Docker image:
 
 ```bash
-docker build -t sap_change -f deployment/Dockerfile.prediction.lambda .
+docker build -t sap -f deployment/Dockerfile.prediction.lambda .
 ```
 
 This will build a Docker image tagged as sap_change using the Dockerfile.prediction.lambda located
@@ -68,7 +68,7 @@ in the deployment directory.
 Once the image is built, you can run it using the following command:
 
 ```bash
-docker run -p 9000:8080 -v ~/.aws/credentials:/root/.aws/credentials:ro -e RUNTIME_ENVIRONMENT=dev sap_change
+docker run -p 9000:8080 -v ~/.aws/credentials:/root/.aws/credentials:ro -e RUNTIME_ENVIRONMENT=dev -e PREDICTIONS_BUCKET=retrofit-sap-predictions-dev sap
 ```
 This command does the following:
 
@@ -79,6 +79,7 @@ Sets the RUNTIME_ENVIRONMENT variable to dev.
 To test the Lambda function, use the following curl command:
 
 ```json
-curl -XPOST "http://localhost:9000/2015-03-31/functions/function/invocations" -d '{"body": "{\"file_location\": \"s3://retrofit-data-dev/model_build_data/change_data/rdsap_full/test_data_with_id.parquet\", \"property_id\": 1, \"portfolio_id\": 4, \"created_at\": \"now\"}"'
+curl -XPOST "http://localhost:9000/2015-03-31/functions/function/invocations" -d '{"body": "{\"file_location\": \"s3://retrofit-data-dev/sap_change_model/one_sample_test_dataset.parquet\", \"property_id\": 1, \"portfolio_id\": 4, \"created_at\": \"now\"}"}'
 ```
+
 This will send a POST request to the running Lambda function and pass in the required data as JSON.

deployment/handlers/prediction_app.py

@@ -107,6 +107,7 @@ def handler(event, context):
             predictions_column_name=generate_predictions_params[
                 "predictions_column_name"
             ],
+            identifier_column=generate_predictions_params["identifier_column"],
         )
 
         return {

modules/ml-pipeline/src/pipeline/configs/settings.yaml

@@ -43,6 +43,7 @@ default:
     test_data_filepath: ./data/prepared_data/test.parquet
     predictions_output_filepath: ./data/predictions/predictions.parquet
     predictions_column_name: predictions
+    identifier_column: id
 
   generate_metrics:
     dataclient_type: local

modules/ml-pipeline/src/pipeline/core/DataClient.py

@@ -142,9 +142,15 @@ class AWSS3Client:
         buffer = BytesIO()
         obj.to_parquet(buffer, index=False)
 
+        # Reset the buffer position to the beginning
+        buffer.seek(0)
+
         bucket, key = location.strip("s3://").split("/", 1)
         self.client.upload_fileobj(buffer, bucket, key)
 
+        # Close the buffer
+        buffer.close()
+
     def _load_parquet(self, location: str, load_config: dict) -> pd.DataFrame:
         """
         Load a parquet file

modules/ml-pipeline/src/pipeline/core/MLMetrics.py

@@ -4,6 +4,7 @@ Implementation of MLMetrics, all of which will have two methods:
 - Generate Plot Suite
 """
 
+import numpy as np
 import pandas as pd
 from typing import Union
 from sklearn.metrics import (
@@ -14,6 +15,18 @@ from sklearn.metrics import (
 )
 from core.interface.InterfaceMetrics import MLMetrics
 
+# Define the function to return the SMAPE value
+def symmetric_mape(actual, predicted) -> float:
+
+    # Convert actual and predicted to numpy
+    # array data type if not already
+    if not all([isinstance(actual, np.ndarray), isinstance(predicted, np.ndarray)]):
+        actual, predicted = np.array(actual), np.array(predicted)
+
+    return np.mean(
+        np.abs(predicted - actual) / ((np.abs(predicted) + np.abs(actual)) / 2)
+    )
+
 
 def metrics_factory(metrics_type: str) -> MLMetrics:
     metrics = {
@@ -34,7 +47,7 @@ class RegressionMetrics:
         median_absolute_error,
         mean_squared_error,
         mean_absolute_percentage_error,
-        # max_error
+        symmetric_mape,
     ]
 
     def generate_metrics(

modules/ml-pipeline/src/pipeline/generate_predictions.py

@@ -14,6 +14,7 @@ def generate_predictions(
     test_data_filepath: str,
     predictions_output_filepath: str,
     predictions_column_name: str,
+    identifier_column: str = "id",
 ):
     """
     For a given model, we generate prediction and evaluate this against the true target
@@ -52,6 +53,12 @@ def generate_predictions(
     predictions_df = pd.DataFrame(predictions)
     predictions_df.columns = [predictions_column_name]
 
-    output_dataclient.save_data(
-        obj=predictions_df, location=predictions_output_filepath, save_config=None
+    output_df = (
+        pd.concat([test_data[identifier_column], predictions_df], axis=1)
+        if identifier_column in test_data.columns
+        else predictions_df
+    )
+
+    output_dataclient.save_data(
+        obj=output_df, location=predictions_output_filepath, save_config=None
     )