درس ۵ سطح پیشرفته: (Machine Learning) یادگیری ماشین در پایتون – FastAPI + Scikit-learn + استقرار مدل با Docker

pip install scikit-learn pandas numpy fastapi uvicorn python-multipart joblib

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestRegressor
import joblib

# داده نمونه (می‌تونی از Kaggle دانلود کنی)
df = pd.read_csv("house_prices.csv")

X = df[['area', 'bedrooms', 'age']]
y = df['price']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

model = RandomForestRegressor(n_estimators=100, random_state=42)
model.fit(X_train, y_train)

print(f"دقت مدل: {model.score(X_test, y_test):.3f}")

# ذخیره مدل
joblib.dump(model, "house_price_model.pkl")
print("مدل ذخیره شد!")

# main.py
from fastapi import FastAPI
from pydantic import BaseModel
import joblib
import numpy as np

app = FastAPI(title="پیش‌بینی قیمت خانه", version="1.0")

# بارگذاری مدل
model = joblib.load("house_price_model.pkl")

class HouseInput(BaseModel):
    area: float
    bedrooms: int
    age: int

@app.post("/predict")
async def predict_price(house: HouseInput):
    features = np.array([[house.area, house.bedrooms, house.age]])
    prediction = model.predict(features)[0]
    return {
        "predicted_price": round(float(prediction), 2),
        "unit": "میلیون تومان",
        "input": house.dict()
    }

@app.get("/")
async def home():
    return {"message": "مدل پیش‌بینی قیمت خانه آماده است! به /docs برو"}

# Dockerfile
FROM python:3.11-slim

WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .

EXPOSE 8000
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]

from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline

pipeline = Pipeline([
    ('scaler', StandardScaler()),
    ('model', RandomForestRegressor())
])

pipeline.fit(X_train, y_train)
joblib.dump(pipeline, "house_pipeline.pkl")

from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB

emails = ["خرید ویاگرا", "جلسه فردا", "فاکتور خرید", "تبریک برنده شدی!"]
labels = [1, 0, 0, 1]  # 1=اسپم

vectorizer = TfidfVectorizer()
X = vectorizer.fit_transform(emails)

model = MultinomialNB()
model.fit(X, labels)

def is_spam(text):
    vec = vectorizer.transform([text])
    return "اسپم" if model.predict(vec)[0] == 1 else "عادی"

print(is_spam("تبریک! شما برنده ۱ میلیون شدید"))

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
import joblib
import numpy as np

app = FastAPI(title="سیستم تشخیص تقلب بانکی")

model = joblib.load("fraud_model.pkl")

class Transaction(BaseModel):
    amount: float
    time: float
    v1: float
    v2: float
    # ... تا V28
    class Config:
        schema_extra = {"example": {...}}

@app.post("/detect-fraud")
async def detect_fraud(tx: Transaction):
    features = np.array([[tx.amount, tx.time, tx.v1, tx.v2]])  # تا V28
    prediction = model.predict(features)[0]
    probability = model.predict_proba(features)[0]
    
    return {
        "is_fraud": bool(prediction),
        "fraud_probability": float(probability[1]),
        "risk_level": "HIGH" if probability[1] > 0.8 else "MEDIUM" if probability[1] > 0.5 else "LOW"
    }