title_with_video/applications/schedule/search_schedule.py

"""
@author: luojunhui
调用接口在微信内搜索视频
"""
import json
import time

from applications.match_algorithm.rank import title_similarity_rank
from applications.search import *
from static.config import gh_id_dict, db_article
from applications.functions.log import logging
from applications.functions.video_item import VideoProducer
from applications.functions.kimi import KimiServer
from applications.functions.common import request_etl

DEFAULT_ACCOUNT_ID_DICT = {
        "uid": 69637498,
        "nick_name": "望长安"
    }


class SearchABTest(object):
    """
    搜索策略实验方案
    """

    ori_title = None
    article_summary = None
    article_keys = None
    gh_id = None
    trace_id = None

    def __init__(self, info, gh_id):
        SearchABTest.set_class_properties(info, gh_id)

    @classmethod
    def set_class_properties(cls, info, gh_id):
        """
        初始化搜索策略实验类
        :param info: kimi 挖掘的基本信息
        :param gh_id: 公众号账号 id
        :return:
        """
        cls.ori_title = info["ori_title"]
        cls.article_summary = info["content_title"]
        cls.article_keys = info["content_keys"]
        cls.trace_id = info["trace_id"]
        cls.gh_id = gh_id

    @classmethod
    async def base_line(cls):
        """
        兜底策略
        """
        result = await SearchMethod().search_v1(
            text=cls.article_keys[0],
            trace_id=cls.trace_id
        )
        if result:
            return result
        else:
            sub_result = await SearchMethod().search_v1(
                text=cls.article_keys[1],
                trace_id=cls.trace_id)
            if sub_result:
                return sub_result
            else:
                return await SearchMethod().search_v1(
                    text=cls.article_keys[2],
                    trace_id=cls.trace_id
                )

    @classmethod
    async def ab_0(cls):
        """
        默认原标题搜索
        :return:
        """
        search_result = await SearchMethod().search_v1(
            text=cls.ori_title,
            trace_id=cls.trace_id
        )
        if search_result:
            return search_result
        else:
            return await cls.base_line()

    @classmethod
    async def ab_1(cls):
        """
        使用 content_summary搜索
        :return:
        """
        search_result = await SearchMethod().search_v1(
            text=cls.article_summary,
            trace_id=cls.trace_id
        )
        if search_result:
            return search_result
        else:
            return await cls.ab_0()

    @classmethod
    async def ab_2(cls):
        """
        使用文本关键词搜索
        :return:
        """
        search_result = await SearchMethod().search_v1(
            text=cls.article_keys[0],
            trace_id=cls.trace_id
        )
        if search_result:
            return search_result
        else:
            return await cls.base_line()

    @classmethod
    async def ab_3(cls):
        """
        使用文本关键词搜索
        :return:
        """
        search_result = await SearchMethod().search_v1(
            text=cls.article_keys[1],
            trace_id=cls.trace_id
        )
        if search_result:
            return search_result
        else:
            return await cls.base_line()

    @classmethod
    async def ab_4(cls):
        """
        使用文本关键词搜索
        :return:
        """
        search_result = await SearchMethod().search_v1(
            text=cls.article_keys[2],
            trace_id=cls.trace_id
        )
        if search_result:
            return search_result
        else:
            return await cls.base_line()

    @classmethod
    async def ab_5(cls):
        """
        增量搜索, 返回result_list
        :return:
        """
        result_list = await SearchMethod().search_v2(
            text=cls.article_summary[:15],
            trace_id=cls.trace_id
        )
        if len(result_list) > 3:
            return result_list
        else:
            result_list += await SearchMethod().search_v2(
                text=cls.ori_title[:15],
                trace_id=cls.trace_id
            )
            if len(result_list) > 3:
                return result_list
            else:
                result_list += await SearchMethod().search_v2(
                    text=cls.article_keys[0],
                    trace_id=cls.trace_id
                )
                if len(result_list) > 3:
                    return result_list
                else:
                    result_list += await SearchMethod().search_v2(
                        text=cls.article_keys[1],
                        trace_id=cls.trace_id
                    )
                    if result_list:
                        return result_list
                    else:
                        result_list += await SearchMethod().search_v2(
                            text=cls.article_keys[2],
                            trace_id=cls.trace_id
                        )
                        return result_list


class SearchMethod(object):
    """
    搜索召回模式
    """
    s_words = []

    @classmethod
    async def search_v1(cls, text, trace_id):
        """
        dy ---> baidu ---> xigua
        :param text:
        :param trace_id:
        :return:
        """
        douyin_result = douyin_search(keyword=text, sensitive_words=cls.s_words, trace_id=trace_id)
        if douyin_result:
            return {"platform": "dy_search", "result": douyin_result[0]}
        else:
            time.sleep(1)
            baidu_result = hksp_search(key=text, sensitive_words=cls.s_words, trace_id=trace_id)
            if baidu_result:
                return {"platform": "baidu_search", "result": baidu_result[0]}
            else:
                xigua_result = xigua_search_v2(keyword=text, sensitive_words=cls.s_words)
                if xigua_result:
                    return {"platform": "xg_search", "result": xigua_result[0]}
                else:
                    return None

    @classmethod
    async def search_v2(cls, text, trace_id):
        """
        dy ---> baidu ---> xigua
        :param trace_id:
        :param text:
        :return:
        """
        L = []
        douyin_result = douyin_search(keyword=text, sensitive_words=cls.s_words, trace_id=trace_id)
        for vid_obj in douyin_result:
            L.append({"platform": "dy_search", "result": vid_obj})
        if len(L) >= 3:
            return L
        else:
            baidu_result = hksp_search(key=text, sensitive_words=cls.s_words, trace_id=trace_id)
            if baidu_result:
                L.append({"platform": "baidu_search", "result": baidu_result[0]})
            xigua_result = xigua_search_v2(keyword=text, sensitive_words=cls.s_words)
            if xigua_result:
                L.append({"platform": "xg_search", "result": xigua_result[0]})
            return L


async def video_sender(video_obj, user, trace_id, platform, content_id):
    """
    异步处理微信 video_obj
    公众号和站内账号一一对应
    :param content_id：
    :param platform:
    :param user:
    :param trace_id:
    :param video_obj:
    :return:
    """
    Video = VideoProducer()
    if platform == "xg_search":
        mq_obj = Video.xg_video_producer(
            video_obj=video_obj,
            user=user,
            trace_id=trace_id,
        )
    elif platform == "baidu_search":
        mq_obj = Video.baidu_video_producer(
            video_obj=video_obj,
            user=user,
            trace_id=trace_id,
        )
    elif platform == "wx_search":
        mq_obj = Video.wx_video_producer(
            video_obj=video_obj,
            user=user,
            trace_id=trace_id,
        )
    elif platform == "dy_search":
        mq_obj = Video.dy_video_producer(
            video_obj=video_obj,
            user=user,
            trace_id=trace_id,
        )
    else:
        mq_obj = {}
    mq_obj['trace_id'] = trace_id
    mq_obj['content_id'] = content_id
    header = {
        "Content-Type": "application/json",
    }
    response = await request_etl(
        url="http://192.168.203.137:4612/etl",
        headers=header,
        json_data=mq_obj
    )
    return response
    # response = await request_etl(
    #     url="http://localhost:4612/etl",
    #     headers=header,
    #     json_data=mq_obj
    # )
    # return response


async def search_videos(params, trace_id, gh_id, mysql_client):
    """
    search and send msg to ETL
    :param mysql_client:
    :param params:
    :param gh_id: 通过账号 id 来控制实验策略
    :param trace_id:
    :return:
    """
    K = KimiServer()
    kimi_info = await K.search_kimi_schedule(params=params)
    kimi_title = kimi_info['k_title']
    content_title = kimi_info['content_title'].replace("'", "").replace('"', "")
    content_keys = json.dumps(kimi_info['content_keys'], ensure_ascii=False)
    update_kimi_sql = f"""
                    UPDATE {db_article} SET
                    kimi_title = %s,
                    kimi_summary = %s,
                    kimi_keys = %s
                    WHERE trace_id = %s;
                """
    await mysql_client.async_insert(
        sql=update_kimi_sql,
        params=(kimi_title, content_title, content_keys, trace_id)
    )
    kimi_info["trace_id"] = trace_id
    SearchAB = SearchABTest(info=kimi_info, gh_id=gh_id)
    # 启三个搜索，每个搜索都保证要搜索到， 分别用key1， key2， key3去搜索
    recall_list = await SearchAB.ab_5()
    logging(
        code="1006",
        info="搜索到{}条视频".format(len(recall_list)),
        data=recall_list,
        trace_id=trace_id
    )
    # 按照标题相似度排序
    ranked_list = title_similarity_rank(content_title=params['title'].split("@@")[-1], recall_list=recall_list)
    index = 0
    for recall_obj in ranked_list:
        if recall_obj:
            platform = recall_obj['platform']
            recall_video = recall_obj['result']
            if recall_video:
                response = await video_sender(
                    video_obj=recall_video,
                    user=gh_id_dict.get(gh_id, DEFAULT_ACCOUNT_ID_DICT),
                    trace_id=trace_id,
                    platform=platform,
                    content_id=params['content_id']
                )
                if response['status'] == "success":
                    index += 1
                    logging(
                        code="1007",
                        info="成功请求etl",
                        data=recall_video,
                        trace_id=trace_id
                    )
                if index >= 3:
                    print("already downloaded 3 videos")
                    logging(
                        code="1008",
                        info="成功下载三条视频",
                        trace_id=trace_id
                    )
                    return index
    return index