Event flow dummy

.gitignore

@@ -1 +1,3 @@
-__pycache__
+__pycache__
+.env
+.vscode

docker-compose.yaml

@@ -11,8 +11,9 @@ services:
       # Access the application at http://localhost:8080
       - "8080:80"
     volumes:
-      # Mount the frontend directory
-      - ./frontend/src:/usr/share/nginx/html
+      # Mount the frontend directory (exclusive relabel for this container)
+      - ./frontend/src:/usr/share/nginx/html:Z
+
   orchestrator:
     build:
       # Use the current directory as the build context
@@ -31,10 +32,13 @@ services:
       # Check app.py in the orchestrator directory to see how this is used
       - PYTHONFILE=/app/orchestrator/src/app.py
     volumes:
-      # Mount the utils directory in the current directory to the /app/utils directory in the container
-      - ./utils:/app/utils
-      # Mount the orchestrator/src directory in the current directory to the /app/orchestrator/src directory in the container
-      - ./orchestrator/src:/app/orchestrator/src
+      # Shared utils (allow multiple containers) — use :z
+      - ./utils:/app/utils:z
+      # Orchestrator source (exclusive)
+      - ./orchestrator/src:/app/orchestrator/src:Z
+    depends_on:
+      - fraud_detection
+      - suggestions
   fraud_detection:
     build:
       # Use the current directory as the build context
@@ -53,7 +57,48 @@ services:
       # Check app.py in the fraud_detection directory to see how this is used
       - PYTHONFILE=/app/fraud_detection/src/app.py
     volumes:
-      # Mount the utils directory in the current directory to the /app/utils directory in the container
-      - ./utils:/app/utils
-      # Mount the fraud_detection/src directory in the current directory to the /app/fraud_detection/src directory in the container
-      - ./fraud_detection/src:/app/fraud_detection/src
+      # Shared utils (allow multiple containers) — use :z
+      - ./utils:/app/utils:z
+      # Fraud source (exclusive)
+      - ./fraud_detection/src:/app/fraud_detection/src:Z
+      - ./fraud_detection/ai:/app/fraud_detection/ai:Z
+  suggestions:
+    build:
+      context: ./
+      dockerfile: ./suggestions/Dockerfile
+    ports:
+      - 50053:50053
+    env_file:
+      - ./.env
+    environment:
+      - PYTHONUNBUFFERED=TRUE
+      - PYTHONFILE=/app/suggestions/src/app.py
+      - GEMINI_API_KEY=${GEMINI_API_KEY}
+    volumes:
+      # Shared utils (allow multiple containers) — use :z
+      - ./utils:/app/utils:z
+      # Suggestions source (exclusive)
+      - ./suggestions/src:/app/suggestions/src:Z
+
+  transaction_verification:
+    build:
+      # Use the current directory as the build context
+      # This allows us to access the files in the current directory inside the Dockerfile
+      context: ./
+      # Use the Dockerfile in the transaction_verification directory
+      dockerfile: ./transaction_verification/Dockerfile
+    ports:
+      # Expose port 50052 on the host, and map port 50052 of the container to port 50052 on the host
+      - 50052:50052
+    environment:
+      # Pass the environment variables to the container
+      # The PYTHONUNBUFFERED environment variable ensures that the output from the application is logged to the console
+      - PYTHONUNBUFFERED=TRUE
+      # The PYTHONFILE environment variable specifies the absolute entry point of the application
+      # Check app.py in the transaction_verification directory to see how this is used
+      - PYTHONFILE=/app/transaction_verification/src/app.py
+    volumes:
+      # Shared utils (allow multiple containers) — use :z
+      - ./utils:/app/utils:z
+      # Transaction verification source (exclusive)
+      - ./transaction_verification/src:/app/transaction_verification/src:Z

docs/README.md

@@ -1,3 +1,18 @@
 # Documentation
 
 This folder should contain your documentation, explaining the structure and content of your project. It should also contain your diagrams, explaining the architecture. The recommended writing format is Markdown.
+
+If you want to rund on local computer run:
+pip install grpcio==1.70.0 grpcio-tools==1.70.0 protobuf==5.29.6 watchdog==6.0.0
+
+
+## System Diagram
+
+<img width="8192" height="3581" alt="image" src="https://github.com/user-attachments/assets/0acd86b9-2655-4524-8ffb-06b62690c1c5" />
+
+
+
+## Architecture diagram
+
+<img width="1527" height="347" alt="image" src="https://github.com/user-attachments/assets/8de78f86-4f65-4843-a721-fcdd4b3c092c" />
+

fraud_detection/Dockerfile

@@ -11,5 +11,8 @@ COPY ./fraud_detection/requirements.txt .
 # Install the Python dependencies
 RUN pip install --no-cache-dir -r requirements.txt
 
+# Ensure Python can import from /app (project root)
+ENV PYTHONPATH=/app
+
 # Set the command to run the application
 CMD python utils/other/hotreload.py "fraud_detection/src/app.py"

fraud_detection/ai/data_gen.py

@@ -0,0 +1,39 @@
+import pandas as pd
+import numpy as np
+import random
+
+np.random.seed(42)
+random.seed(42)
+
+n_samples = 1000
+
+data = []
+
+for _ in range(n_samples):
+    price = round(np.random.uniform(10, 2000), 2)
+
+    # Random credit card number (some intentionally start with 999)
+    if np.random.rand() < 0.05:
+        credit_card = "999" + "".join([str(np.random.randint(0, 10)) for _ in range(13)])
+    else:
+        credit_card = "".join([str(np.random.randint(0, 10)) for _ in range(16)])
+
+    # Fraud rule
+    is_fraud = 1 if (price > 1000 or credit_card.startswith("999")) else 0
+
+    data.append([
+        price,
+        credit_card,
+        is_fraud
+    ])
+
+df = pd.DataFrame(data, columns=[
+    "price",
+    "credit_card",
+    "is_fraud"
+])
+
+df.to_csv("./fraud_detection/ai/transactions.csv", index=False)
+
+print("transactions.csv created successfully.")
+print(df.head())

fraud_detection/ai/train.py

@@ -0,0 +1,49 @@
+import pandas as pd
+import numpy as np
+from sklearn.model_selection import train_test_split
+from sklearn.metrics import classification_report, roc_auc_score
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.utils.class_weight import compute_class_weight
+
+import joblib
+
+# Load dataset (example: credit card transactions)
+df = pd.read_csv("./fraud_detection/ai/transactions.csv")
+
+# Features and label
+X = df.drop("is_fraud", axis=1)
+y = df["is_fraud"]
+
+# Train/test split (stratified for imbalance)
+X_train, X_test, y_train, y_test = train_test_split(
+    X, y, test_size=0.2, stratify=y, random_state=42
+)
+
+# Handle class imbalance via class weights
+class_weights = compute_class_weight(
+    class_weight="balanced",
+    classes=np.unique(y_train),
+    y=y_train
+)
+weights = {0: class_weights[0], 1: class_weights[1]}
+
+# Model
+model = RandomForestClassifier(
+    n_estimators=200,
+    max_depth=10,
+    class_weight=weights,
+    random_state=42,
+    n_jobs=-1
+)
+
+model.fit(X_train, y_train)
+
+# Predictions
+y_pred = model.predict(X_test)
+y_proba = model.predict_proba(X_test)[:, 1]
+
+print(classification_report(y_test, y_pred))
+print("ROC-AUC:", roc_auc_score(y_test, y_proba))
+
+joblib.dump(model, "./fraud_detection/ai/fraud_model.joblib")
+