import redis

# 實(shí)例化一個redis連接池
pool = redis.ConnectionPool(host='127.0.0.1', port=6379, decode_responses=True, db=4, password='')

r = redis.Redis(connection_pool=pool)
# 爬蟲實(shí)例限制為4 即只允許4個scrapy實(shí)例在運(yùn)行
limited = 4

# 聲明redis的樂觀鎖
lock = r.Lock()

# lock.acquire中有while循環(huán)，即它會線程阻塞，直到當(dāng)前線程獲得redis的lock，才會繼續(xù)往下執(zhí)行代碼
if lock.acquire():
	# 1、從reids中取一條爬蟲信息
	info = redis.get() 
	
	# 2、while循環(huán)監(jiān)聽爬蟲運(yùn)行的數(shù)量
	while True:
		req = requests.get('http://127.0.0.1:6800/daemonstatus.json').json()
		# 統(tǒng)計當(dāng)前有多少個爬蟲在運(yùn)行
		running = req.get('running') + req.get('pending')
		
		# 3、判斷是否等待還是要增加爬蟲數(shù)量
		# 3.1 如果在運(yùn)行的數(shù)量大于等于設(shè)置到量 則繼續(xù)等待
		if running >= limited:
			continue
		
		# 3.2 如果小于 則啟動爬蟲
		start_scrapy(info)
		# 3.3 將info從redis中刪除
		redis.delete(info)
		# 3.4 釋放鎖
		lock.release()
		break
		

@shared_task
def scrapy_control(key_uuid):

    r = redis.Redis(connection_pool=pool)
    db = MysqlDB()
    speed_limited = db.fetch_config('REPTILE_SPEED')
    speed_limited = int(speed_limited[0])

    keywords_num = MysqlDB().fetch_config('SEARCH_RANDOM')
    keywords_num = int(keywords_num[0])


    # while True:
    lock = r.lock('lock')
    with open('log/celery/info.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + ' 進(jìn)入處理環(huán)節(jié)' +  '\n')
    try:
        # acquire默認(rèn)阻塞 如果獲取不到鎖時 會一直阻塞在這個函數(shù)的while循環(huán)中
        if lock.acquire():
            with open('log/celery/info.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + ' 獲得鎖' +  '\n')
            # 1 從redis中獲取信息
            redis_obj = json.loads(r.get(key_uuid))
            user_id = redis_obj.get('user_id')
            contents = redis_obj.get('contents')
            
            # 2 使用while循環(huán)處理核心邏輯          
            is_hold_print = True
            while True:
                req = requests.get('http://127.0.0.1:6800/daemonstatus.json').json()
                running = req.get('running') + req.get('pending')
                # 3 如果仍然有足夠的爬蟲在運(yùn)行 則hold住redis鎖，等待有空余的爬蟲位置讓出
                if running >= speed_limited:
                    if is_hold_print:
                        with open('log/celery/info.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + ' 爬蟲在運(yùn)行，線程等待中' +  '\n')
                        is_hold_print = False
                    time.sleep(1)
                    continue
                
                # 4 有空余的爬蟲位置 則往下走
                # 4.1 處理完所有的內(nèi)容后 釋放鎖
                if len(contents) == 0:
                    r.delete(key_uuid)
                    with open('log/celery/info.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + ' 任務(wù)已完成，從redis中刪除' +  '\n')
                    lock.release()
                    with open('log/celery/info.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + ' 釋放鎖' +  '\n')
                    break

                # 4.2 創(chuàng)建task任務(wù)
                task_uuid = str(uuid.uuid4())
                article_obj = contents.pop()
                article_id = article_obj.get('article_id')
                article = article_obj.get('content')
                try:
                    Task.objects.create(
                        task_uuid = task_uuid,
                        user_id = user_id,
                        article_id = article_id,
                        content = article
                    )
                except Exception as e:
                    with open('log/celery/error.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + '->' + str(task_uuid) + ' 創(chuàng)建Task出錯: ' + str(e) +  '\n')
                # finally:
                # 4.3 啟動爬蟲任務(wù) 即便創(chuàng)建task失敗也會啟動
                try:
                    task_chain(user_id, article, task_uuid, keywords_num)
                except Exception as e:
                    with open('log/celery/error.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + ' 啟動任務(wù)鏈?zhǔn)? ' + str(e) +  '\n')
                
                # 加入sleep 防止代碼執(zhí)行速度快于爬蟲啟動速度而導(dǎo)致當(dāng)前線程啟動額外的爬蟲
                time.sleep(5)

    except Exception as e:
        with open('log/celery/error.log', 'a') as f: f.write(str(datetime.datetime.now()) + '--' + str(key_uuid) + ' 獲得鎖之后的操作出錯: ' + str(e) +  '\n')
        lock.release()