From 00d5c4f5c98f6bef882cc55a5ed4fe0d71d0f97b Mon Sep 17 00:00:00 2001
From: Ayushman Rai <hiayushman25@gmail.com>
Date: Mon, 13 Oct 2025 14:30:30 +0530
Subject: [PATCH] Added FastAPI server for model predictions

---
 api/server.py       | 59 +++++++++++++++++++++++++++++++++++++++++++++
 api/server_flask.py | 38 +++++++++++++++++++++++++++++
 2 files changed, 97 insertions(+)
 create mode 100644 api/server.py
 create mode 100644 api/server_flask.py

diff --git a/api/server.py b/api/server.py
new file mode 100644
index 0000000..e785787
--- /dev/null
+++ b/api/server.py
@@ -0,0 +1,59 @@
+from fastapi import FastAPI
+from pydantic import BaseModel
+from models.decision_tree import DecisionTreeModel
+from utils.data_helpers import generate_sample_regression_data
+import pandas as pd
+import uvicorn
+
+app = FastAPI(title="ML Simulator API", version="1.0")
+
+# ----- Request Schema -----
+class ModelRequest(BaseModel):
+    model: str
+    task_type: str = "classification"  # or regression
+    features: list
+
+# ----- Endpoints -----
+@app.get("/")
+def home():
+    return {"message": "Welcome to the ML Simulator API 🎯"}
+
+@app.post("/predict")
+def predict(req: ModelRequest):
+    """
+    Example Input:
+    {
+        "model": "decision_tree",
+        "task_type": "classification",
+        "features": [[5.1, 3.5, 1.4, 0.2]]
+    }
+    """
+    model_name = req.model.lower()
+    task_type = req.task_type.lower()
+
+    # Load or initialize model
+    if model_name == "decision_tree":
+        model = DecisionTreeModel(task_type=task_type)
+        # Generate training data just for demo
+        if task_type == "classification":
+            from sklearn.datasets import make_classification
+            X, y = make_classification(n_samples=100, n_features=len(req.features[0]), random_state=42)
+        else:
+            df = generate_sample_regression_data(n_samples=100, n_features=len(req.features[0]))
+            X, y = df.iloc[:, :-1], df.iloc[:, -1]
+
+        # Train model
+        model.train(X, y)
+
+        # Predict
+        prediction = model.predict(pd.DataFrame(req.features))
+        return {
+            "model_used": f"Decision Tree ({task_type})",
+            "prediction": prediction.tolist()
+        }
+    else:
+        return {"error": "Model not implemented yet."}
+
+
+if __name__ == "__main__":
+    uvicorn.run("api.server:app", host="0.0.0.0", port=8000, reload=True)
diff --git a/api/server_flask.py b/api/server_flask.py
new file mode 100644
index 0000000..8bad11e
--- /dev/null
+++ b/api/server_flask.py
@@ -0,0 +1,38 @@
+from flask import Flask, request, jsonify
+from models.decision_tree import DecisionTreeModel
+from utils.data_helpers import generate_sample_regression_data
+import pandas as pd
+from sklearn.datasets import make_classification
+
+app = Flask(__name__)
+
+@app.route('/')
+def home():
+    return jsonify({"message": "Welcome to ML Simulator Flask API"})
+
+@app.route('/predict', methods=['POST'])
+def predict():
+    data = request.get_json()
+    model_name = data.get('model', 'decision_tree')
+    task_type = data.get('task_type', 'classification')
+    features = data.get('features', [])
+
+    if model_name == "decision_tree":
+        model = DecisionTreeModel(task_type=task_type)
+        if task_type == "classification":
+            X, y = make_classification(n_samples=100, n_features=len(features[0]), random_state=42)
+        else:
+            df = generate_sample_regression_data(n_samples=100, n_features=len(features[0]))
+            X, y = df.iloc[:, :-1], df.iloc[:, -1]
+
+        model.train(X, y)
+        prediction = model.predict(pd.DataFrame(features))
+        return jsonify({
+            "model_used": f"Decision Tree ({task_type})",
+            "prediction": prediction.tolist()
+        })
+    else:
+        return jsonify({"error": "Model not implemented"}), 400
+
+if __name__ == '__main__':
+    app.run(debug=True)