Merge pull request #364 from microsoft/retail-sampleData

feat - retail sample data

Author: Fr4nc3

src/Data/.env.sample

@@ -0,0 +1,4 @@
+SEARCH_ENDPOINT=
+SEARCH_API_KEY=
+BLOB_CONNECTION_STRING=
+BLOB_CONTAINER_NAME=

src/Data/create_search_resources.py

@@ -0,0 +1,78 @@
+import os
+from dotenv import load_dotenv
+from azure.core.credentials import AzureKeyCredential
+from azure.search.documents.indexes import SearchIndexClient, SearchIndexerClient
+from azure.search.documents.indexes.models import (
+    SearchIndexerDataSourceConnection,
+    SearchIndexer,
+    SearchIndex,
+    SimpleField,
+    SearchableField,
+    SearchFieldDataType
+)
+
+# Load environment variables from .env file
+load_dotenv()
+
+# Retrieve environment variables
+SEARCH_ENDPOINT = os.getenv("SEARCH_ENDPOINT")
+SEARCH_API_KEY = os.getenv("SEARCH_API_KEY")
+BLOB_CONNECTION_STRING = os.getenv("BLOB_CONNECTION_STRING")
+BLOB_CONTAINER_NAME = os.getenv("BLOB_CONTAINER_NAME")
+
+# Validate required environment variables
+required_vars = [SEARCH_ENDPOINT, SEARCH_API_KEY, BLOB_CONNECTION_STRING, BLOB_CONTAINER_NAME]
+if any(var is None for var in required_vars):
+    raise ValueError("One or more required environment variables are missing.")
+
+def create_search_resources():
+    """Create Azure AI Search resources: data source, index, and indexer."""
+    try:
+        # Initialize SearchIndexClient for index operations
+        index_client = SearchIndexClient(
+            endpoint=SEARCH_ENDPOINT,
+            credential=AzureKeyCredential(SEARCH_API_KEY)
+        )
+        
+        # Initialize SearchIndexerClient for data source and indexer operations
+        indexer_client = SearchIndexerClient(
+            endpoint=SEARCH_ENDPOINT,
+            credential=AzureKeyCredential(SEARCH_API_KEY)
+        )
+
+        # Define data source connection
+        data_source = SearchIndexerDataSourceConnection(
+            name="macae-blob-datasets",
+            type="azureblob",
+            connection_string=BLOB_CONNECTION_STRING,
+            container={"name": BLOB_CONTAINER_NAME}
+        )
+        indexer_client.create_or_update_data_source_connection(data_source)
+        print("Data source 'macae-blob-datasets' created successfully.")
+
+        # Define index schema
+        index = SearchIndex(
+            name="macae-index",
+            fields=[
+                SimpleField(name="id", type=SearchFieldDataType.String, key=True),
+                SearchableField(name="content", type=SearchFieldDataType.String),
+                SearchableField(name="metadata", type=SearchFieldDataType.String)
+            ]
+        )
+        index_client.create_or_update_index(index)
+        print("Index 'macae-index' created successfully.")
+
+        # Define indexer
+        indexer = SearchIndexer(
+            name="macae-indexer",
+            data_source_name="macae-blob-datasets",
+            target_index_name="macae-index"
+        )
+        indexer_client.create_or_update_indexer(indexer)
+        print("Indexer 'macae-indexer' created successfully.")
+
+    except Exception as e:
+        print(f"An error occurred while creating search resources: {e}")
+
+if __name__ == "__main__":
+    create_search_resources()

src/Data/data_upload.py

@@ -0,0 +1,21 @@
+from azure.storage.blob import BlobServiceClient
+import os
+from dotenv import load_dotenv
+
+load_dotenv() 
+
+# Retrieve environment variables
+BLOB_CONNECTION_STRING = os.getenv("BLOB_CONNECTION_STRING")
+BLOB_CONTAINER_NAME = os.getenv("BLOB_CONTAINER_NAME")
+local_folder = "./datasets"
+
+blob_service_client = BlobServiceClient.from_connection_string(BLOB_CONNECTION_STRING)
+container_client = blob_service_client.get_container_client(BLOB_CONTAINER_NAME)
+
+for filename in os.listdir(local_folder):
+    file_path = os.path.join(local_folder, filename)
+    if os.path.isfile(file_path):
+        print(f"Uploading {filename}...")
+        with open(file_path, "rb") as data:
+            container_client.upload_blob(name=filename, data=data, overwrite=True)
+print("Upload complete!")

src/Data/datasets/Competitor_Pricing_Analysis.csv

@@ -0,0 +1,5 @@
+ProductCategory,ContosoAveragePrice,CompetitorAveragePrice
+Dresses,120,100
+Shoes,100,105
+Accessories,60,55
+Sportswear,80,85

src/Data/datasets/Customer_Churn_Analysis.csv

@@ -0,0 +1,6 @@
+ReasonForCancellation,Percentage
+Service Dissatisfaction,40
+Financial Reasons,3
+Competitor Offer,15
+Moving to a Non-Service Area,5
+Other,37

src/Data/datasets/Email_Marketing_Engagement.csv

@@ -0,0 +1,6 @@
+Campaign,Opened,Clicked,Unsubscribed
+Summer Sale,Yes,Yes,No
+New Arrivals,Yes,No,No
+Exclusive Member Offers,No,No,No
+Personal Styling Invite,No,No,No
+Autumn Collection Preview,Yes,Yes,No

src/Data/datasets/Loyalty_Program_Overview.csv

@@ -0,0 +1,2 @@
+TotalPointsEarned,PointsRedeemed,CurrentPointBalance,PointsExpiringNextMonth
+4800,3600,1200,1200

src/Data/datasets/Subscription_benefits_utilization.csv

@@ -0,0 +1,5 @@
+Benefit,UsageFrequency
+Free Shipping,7
+Early Access to Collections,2
+Exclusive Discounts,1
+Personalized Styling Sessions,0

src/Data/datasets/Unauthorized_Access_Attempts.csv

@@ -0,0 +1,4 @@
+Date,IPAddress,Location,SuccessfulLogin
+2023-06-20,192.168.1.1,Home Network,Yes
+2023-07-22,203.0.113.45,Unknown,No
+2023-08-15,198.51.100.23,Office Network,Yes

src/Data/datasets/Warehouse_Incident_Reports.csv

@@ -0,0 +1,4 @@
+Date,IncidentDescription,AffectedOrders
+2023-06-15,Inventory system outage,100
+2023-07-18,Logistics partner strike,250
+2023-08-25,Warehouse flooding due to heavy rain,150