app.py

import os
import json
import random
import datetime
import pandas as pd
import numpy as np
from flask import Flask, render_template, jsonify, request
from faker import Faker

app = Flask(__name__)
fake = Faker()

# Configuration
app.config['JSON_AS_ASCII'] = False
app.config['MAX_CONTENT_LENGTH'] = 512 * 1024 * 1024
# app.jinja_env.variable_start_string = "[["
# app.jinja_env.variable_end_string = "]]"

# Global Cache for simulated data
DATA_CACHE = None

class CostSimulator:
    def __init__(self):
        self.services = ['Amazon EC2', 'Amazon RDS', 'Amazon S3', 'AWS Lambda', 'Amazon CloudFront']
        self.regions = ['us-east-1', 'us-west-2', 'eu-central-1', 'ap-northeast-1']
        self.instance_types = ['t3.micro', 'm5.large', 'c5.xlarge', 'r5.2xlarge']
        
    def generate_data(self, days=90):
        """Generate simulated billing data for the last N days."""
        data = []
        end_date = datetime.date.today()
        start_date = end_date - datetime.timedelta(days=days)
        
        # Base resources (Long running)
        resources = []
        for _ in range(20):
            resources.append({
                'id': fake.uuid4(),
                'service': random.choice(self.services),
                'region': random.choice(self.regions),
                'name': fake.hostname(),
                'tag_env': random.choice(['Production', 'Staging', 'Dev']),
                'base_cost': random.uniform(0.5, 50.0)  # Daily cost
            })
            
        current_date = start_date
        while current_date <= end_date:
            # Add base resource costs
            for res in resources:
                daily_cost = res['base_cost'] * random.uniform(0.9, 1.1)
                
                # Simulate a cost spike in the last 3 days for one specific resource
                if current_date > end_date - datetime.timedelta(days=3) and res['service'] == 'AWS Lambda':
                    daily_cost *= random.uniform(5.0, 10.0) # Anomaly!
                    
                data.append({
                    'date': current_date.strftime('%Y-%m-%d'),
                    'resource_id': res['id'],
                    'resource_name': res['name'],
                    'service': res['service'],
                    'region': res['region'],
                    'environment': res['tag_env'],
                    'cost': round(daily_cost, 4),
                    'usage_amount': round(random.uniform(10, 1000), 2)
                })
            
            # Add some random transient costs (Spot instances, data transfer)
            for _ in range(random.randint(5, 15)):
                svc = random.choice(self.services)
                data.append({
                    'date': current_date.strftime('%Y-%m-%d'),
                    'resource_id': fake.uuid4(),
                    'resource_name': 'Transient-' + fake.word(),
                    'service': svc,
                    'region': random.choice(self.regions),
                    'environment': 'Dev',
                    'cost': round(random.uniform(0.1, 5.0), 4),
                    'usage_amount': round(random.uniform(1, 100), 2)
                })
                
            current_date += datetime.timedelta(days=1)
            
        return pd.DataFrame(data)

def get_data(refresh=False):
    global DATA_CACHE
    if DATA_CACHE is None or refresh:
        sim = CostSimulator()
        DATA_CACHE = sim.generate_data()
    return DATA_CACHE

@app.route('/')
def index():
    return render_template('index.html')

@app.route('/api/summary')
def api_summary():
    df = get_data()
    
    # Total Cost
    total_cost = df['cost'].sum()
    
    # This Month vs Last Month (Simulated logic using 30 day windows)
    today = datetime.date.today()
    last_30_start = (today - datetime.timedelta(days=30)).strftime('%Y-%m-%d')
    prev_30_start = (today - datetime.timedelta(days=60)).strftime('%Y-%m-%d')
    
    current_cost = df[df['date'] >= last_30_start]['cost'].sum()
    prev_cost = df[(df['date'] >= prev_30_start) & (df['date'] < last_30_start)]['cost'].sum()
    
    mom_change = 0
    if prev_cost > 0:
        mom_change = ((current_cost - prev_cost) / prev_cost) * 100
        
    return jsonify({
        'total_cost': round(total_cost, 2),
        'current_month_cost': round(current_cost, 2),
        'mom_change': round(mom_change, 2),
        'forecast': round(current_cost * 1.05, 2) # Simple forecast
    })

@app.route('/api/trend')
def api_trend():
    df = get_data()
    # Group by date and service
    trend = df.groupby(['date', 'service'])['cost'].sum().reset_index()
    
    # Pivot for ECharts
    pivot = trend.pivot(index='date', columns='service', values='cost').fillna(0)
    
    dates = pivot.index.tolist()
    series = []
    for column in pivot.columns:
        series.append({
            'name': column,
            'type': 'bar',
            'stack': 'total',
            'data': pivot[column].round(2).tolist()
        })
        
    return jsonify({
        'dates': dates,
        'series': series
    })

@app.route('/api/anomalies')
def api_anomalies():
    df = get_data()
    # Simple anomaly detection: Daily cost > Mean + 3*STD
    daily_cost = df.groupby(['date', 'service'])['cost'].sum().reset_index()
    anomalies = []
    
    for service in df['service'].unique():
        svc_data = daily_cost[daily_cost['service'] == service]
        mean = svc_data['cost'].mean()
        std = svc_data['cost'].std()
        threshold = mean + 3 * std
        
        outliers = svc_data[svc_data['cost'] > threshold]
        for _, row in outliers.iterrows():
            anomalies.append({
                'date': row['date'],
                'service': service,
                'cost': round(row['cost'], 2),
                'threshold': round(threshold, 2),
                'severity': 'Critical' if row['cost'] > mean * 2 else 'High'
            })
            
    return jsonify(sorted(anomalies, key=lambda x: x['date'], reverse=True))

@app.route('/api/recommendations')
def api_recommendations():
    # Simulate logic: Find resources that are consistent but "expensive" -> Suggest RI
    # Or find Dev resources running on weekends
    
    recommendations = [
        {
            'id': 1,
            'title': '购买 EC2 预留实例 (RI)',
            'description': '检测到 5 台 m5.large 实例长期运行，购买 1 年期全预付 RI 可节省 35%。',
            'potential_savings': 450.00,
            'effort': 'Medium',
            'category': 'Rate Optimization'
        },
        {
            'id': 2,
            'title': '清理未关联的弹性 IP',
            'description': '发现 3 个 EIP 未绑定到运行中的实例。',
            'potential_savings': 15.00,
            'effort': 'Low',
            'category': 'Waste'
        },
        {
            'id': 3,
            'title': 'S3 生命周期策略优化',
            'description': '2TB 的日志数据超过 90 天未访问，建议归档至 Glacier。',
            'potential_savings': 120.50,
            'effort': 'Medium',
            'category': 'Storage'
        },
        {
            'id': 4,
            'title': 'RDS 实例空闲检测',
            'description': 'db-staging-01 在过去 7 天 CPU 使用率低于 2%。',
            'potential_savings': 85.20,
            'effort': 'High',
            'category': 'Rightsizing'
        }
    ]
    return jsonify(recommendations)

@app.route('/api/breakdown')
def api_breakdown():
    df = get_data()
    # Breakdown by Service
    by_service = df.groupby('service')['cost'].sum().sort_values(ascending=False)
    # Breakdown by Environment
    by_env = df.groupby('environment')['cost'].sum().sort_values(ascending=False)
    
    return jsonify({
        'service_breakdown': [{'name': k, 'value': round(v, 2)} for k, v in by_service.items()],
        'env_breakdown': [{'name': k, 'value': round(v, 2)} for k, v in by_env.items()]
    })

def _ensure_schema(df: pd.DataFrame) -> pd.DataFrame:
    required_cols = ['date', 'resource_id', 'resource_name', 'service', 'region', 'environment', 'cost', 'usage_amount']
    for col in required_cols:
        if col not in df.columns:
            df[col] = None
    df['date'] = df['date'].fillna(datetime.date.today().strftime('%Y-%m-%d')).astype(str)
    df['resource_id'] = df['resource_id'].fillna(df['resource_name'].fillna('').astype(str) + '-' + pd.Series(range(len(df))).astype(str)).astype(str)
    df['resource_name'] = df['resource_name'].fillna('Imported-' + pd.Series(range(len(df))).astype(str)).astype(str)
    df['service'] = df['service'].fillna('Unknown Service').astype(str)
    df['region'] = df['region'].fillna('us-east-1').astype(str)
    df['environment'] = df['environment'].fillna('Dev').astype(str)
    df['cost'] = pd.to_numeric(df['cost'], errors='coerce').fillna(0.0)
    df['usage_amount'] = pd.to_numeric(df['usage_amount'], errors='coerce').fillna(0.0)
    return df[required_cols]

@app.route('/api/upload', methods=['POST'])
def api_upload():
    global DATA_CACHE
    file = request.files.get('file')
    if not file:
        return jsonify({'status': 'error', 'message': '未收到文件'}), 400
    filename = file.filename or 'uploaded'
    try:
        if filename.lower().endswith('.csv'):
            chunks = pd.read_csv(file.stream, chunksize=100000, encoding='utf-8', on_bad_lines='skip')
            df = pd.concat(list(chunks), ignore_index=True)
        elif filename.lower().endswith('.json'):
            payload = json.load(file.stream)
            df = pd.DataFrame(payload if isinstance(payload, list) else [payload])
        else:
            tmp_path = os.path.join('/tmp', filename)
            file.save(tmp_path)
            return jsonify({'status': 'success', 'message': '二进制文件已保存', 'path': tmp_path})
        df = _ensure_schema(df)
        DATA_CACHE = df
        return jsonify({'status': 'success', 'rows': int(len(df))})
    except Exception as e:
        return jsonify({'status': 'error', 'message': f'导入失败: {str(e)}'}), 500

def _is_api_request():
    return request.path.startswith('/api/')

@app.errorhandler(404)
def handle_404(e):
    if _is_api_request():
        return jsonify({'status': 'error', 'message': '接口不存在'}), 404
    return ('页面不存在 (404)', 404)

@app.errorhandler(500)
def handle_500(e):
    if _is_api_request():
        return jsonify({'status': 'error', 'message': '服务器内部错误'}), 500
    return ('内部服务器错误，请稍后再试', 500)

@app.route('/api/refresh', methods=['POST'])
def api_refresh():
    get_data(refresh=True)
    return jsonify({'status': 'success'})

if __name__ == '__main__':
    port = int(os.environ.get('PORT', 7860))
    app.run(host='0.0.0.0', port=port)