import re
import feedparser
import boto3
from datetime import datetime
from dateutil import tz
from dateutil import parser

# DynamoDB
# boto3.resourceは非推奨のためclientを使用
dynamodb = boto3.client('dynamodb')
TABLE_NAME = 'rss_feeds'
# 何日後にDynamoDBから削除するのか設定
day = 3
TTL = 60 * 60 * 24 * day

# Amazon translate
translate = boto3.client('translate')
SRC_LANG = 'en'
TRG_LANG = 'ja'

# timezone
timezone = tz.gettz('Asia/Tokyo')

# RSS URL
RSS_URL = 'https://aws.amazon.com/jp/blogs/aws/feed/'

def lambda_handler(event, context):
    # RSS取得
    data = feedparser.parse(RSS_URL)

    # 1件ずつitemを取得
    for entry in data.entries:
        title = entry.title
        link = entry.link
        # descriptionにはHTMLタグが含まれるので削除
        description = remove_http_tag(entry.description)
        # DynamoDBのTTLで使用するため、unixtimeも算出
        published, unixtime =  get_time_info(entry.published)
        limit_unixtime = unixtime + TTL
        category = entry.category
        
        # 既存判断。既存の場合はスキップ
        if not is_item_exist(link):
            # タイトルと要約はAmazon Translateで翻訳
            ja_title = get_translate_text(title)
            ja_description = get_translate_text(description)
            # DynamoDBへ格納
            put_item(title, link, description, published, category, ja_title, ja_description, limit_unixtime)
            

def remove_http_tag(str):
    return re.sub(re.compile('<.*?>'), '', str)
    

def is_item_exist(link) -> bool:
    options = {
        'TableName': TABLE_NAME,
        'Key': {
            'link': {'S': link},
        }
    }
    
    try:
        response = dynamodb.get_item(**options)
        response['Item']
        print("The Item Already Exist")
        return True
    except KeyError:
        print("New Item : " + link)
        return False

def put_item(title, link, description, published, category, ja_title, ja_description, limit_unixtime):
    options = {
        'TableName': TABLE_NAME,
        'Item': {
            'title': {'S': title},
            'link': {'S': link},
            'description': {'S': description},
            'published': {'S': published},
            'category': {'S': category},
            'ja_title': {'S': ja_title},
            'ja_description': {'S': ja_description},
            'limit_unixtime': {'N': str(limit_unixtime)}
        }
    }
    # DynamoDB格納
    dynamodb.put_item(**options)
    
def get_translate_text(text):
    # 英語→日本語訳
    response = translate.translate_text(
        Text=text,
        SourceLanguageCode=SRC_LANG,
        TargetLanguageCode=TRG_LANG
    )

    return response.get('TranslatedText')
    
def get_time_info(rss_pub_date):
    # datetimeに変換
    datetime = parser.parse(rss_pub_date).astimezone(timezone)
    
    # iso8601に変換
    iso_datetime = datetime.isoformat()
    
    # unixtime取得
    unixtime = datetime.timestamp()
    
    return iso_datetime, unixtime