chatGptNew/text_analysis/tools/tool.py

#coding:utf8
import re
from jsonpath_ng import parse
import json
import traceback
import regex

def parse_data(data,para):
    param_split = str(para).split(":")
    datasourcestr = data[param_split[0]]
    datasource = json.loads(datasourcestr)
    # 创建 JsonPath 表达式对象
    expr = parse(param_split[1])
    # 使用表达式来选择 JSON 元素
    match = [match.value for match in expr.find(datasource)]
    val = match[0]
    return val

def parse_gptResult(output,gptContent):
    json_result = {}
    try:
        json_gpt=json.loads(str(gptContent).replace("```json","").replace("```",""))
        for key in output.keys():
            if key in json_gpt.keys():
                json_result[key]=json_gpt[key]

        return json_result
    except  Exception as e:
        try:
            # 直接解析失败，使用正则表达式匹配外层的 {}
            pattern = r'\{(?:[^{}]*|(?R))*\}'
            match = regex.search(pattern, gptContent, flags=regex.DOTALL)
            # match = re.search(pattern, gptContent, re.DOTALL)
            if match:
                json_gpt = json.loads(match.group())
                for key in output.keys():
                    if key in json_gpt.keys():
                        json_result[key]=json_gpt[key]
                return json_result
            else:
                return None
        except:
            traceback.print_exc()
            return None


def get_content(inputdata,logging):
    """
    重新组装参数
    :param inputdata:原json数据
    :return: 组装的prompt及其他参数
    """
    res={}
    input=inputdata["input"]
    if "data" not in inputdata.keys():
        data=""
    else:
        data=inputdata["data"]
    prompts=input["prompt"]
    prompt_res=""
    if prompts:
        for prompt in prompts:
            if str(prompt["type"])=="1":
                prompt_res+=prompt["value"]
            elif str(prompt["type"])=="2":
                try:
                    tmp=parse_data(data,prompt["value"])
                    prompt_res +=tmp
                except:
                    logging.info("动态字段获取数据失败。{}-{}".format(prompt, traceback.format_exc()))

    # logging.info("拼接后的问题：{}".format(prompt_res))
    res["prompt"]=prompt_res
    res["authorization"]=input["authorization"]
    res["model"]=input["model"]
    res["temperature"]=input["temperature"]
    res["authorization"]=input["authorization"]
    res["top_p"]=input["top_p"]
    res["n"]=input["n"]
    return res

# def get_content(inputdata,logging):
#     """
#     重新组装参数
#     :param inputdata:原json数据
#     :return: 组装的prompt及其他参数
#     """
#     res={}
#     admin=inputdata["metadata"]["admin"]
#     data=inputdata["data"]
#     prompt=admin["prompt"]
#     if_user=re.findall("{{(.*)}}",prompt)
#     if_data=re.findall("@@(.*)@@",prompt)
#     if if_user != []:
#         user_data=inputdata["metadata"]["user"]
#         if if_user[0] in user_data.keys():
#             tmp=user_data[if_user[0]]
#             prompt=re.sub("{{(.*)}}",tmp,prompt)
#     if if_data!=[] and if_data[0] in data.keys():
#         tmp1=data[if_data[0]]
#         prompt=re.sub("@@(.*)@@",tmp1,prompt)
#     res["prompt"]=prompt
#     res["authorization"]=admin["authorization"]
#     res["model"]=admin["model"]
#     res["temperature"]=admin["temperature"]
#     res["authorization"]=admin["authorization"]
#     res["top_p"]=admin["top_p"]
#     res["n"]=admin["n"]
#     return res

if __name__=="__main__":
    inputdata={
    "output":{
        "id":"id",
        "content":"content"
    },
    "address":"http://172.18.1.181:9011/chatGpt/",
    "input":{
        "authorization":"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",
        "top_p":"1",
        "temperature":"1",
        "model":"gpt-3.5-turbo-16k",
        "prompt":[{"type":1,"value":"下面我给出一段文本，这段文本的内容是"},{"type":2,"value":"2_内容翻译:$['content']"},{"type":1,"value":"请总结上述内容"}],
        "n":"1"
    },
    "data":{
        "1_Youtube采集":"{\"isDownload\":\"true\",\"imageCount\":0,\"groupRules\":[],\"commentUrl\":\"\",\"channel\":\"网络视频\",\"readCount\":55,\"resolution\":\"best\",\"srcimagePath\":[],\"fileCount\":0,\"subjectId\":304864,\"forwardQuoteCount\":-1,\"province\":\"\",\"crawlDataFlagType\":\"0\",\"price\":0,\"quoteCount\":-1,\"translateContentLength\":\"\",\"forwardUserType\":0,\"brand\":\"\",\"createTimeStr\":\"2023-11-02 14:30:24\",\"ocrLength\":0,\"hasOCR\":0,\"authornickname\":\"\",\"contentSimHash\":\"a870a5b58416c163e2618400adf2ea0f\",\"dns\":\"https://www.youtube.com/\",\"crawlDay\":1698854400000,\"asrText\":\"\",\"nomorprice\":0,\"mentionAccountUrl\":[],\"titleSimHash\":\"fc47b579bbee866173216c05bf3a8dcd\",\"videoPath\":[\"/group16/default/20231102/14/25/6/sbbJSkJdgF4.mp4\"],\"brandId\":\"\",\"commentId\":\"\",\"impression\":\"\",\"originalPhrase\":\"[\\\"8473704592933378\\\",\\\"video\\\"]\",\"userTypeContent\":\"\",\"forwardContent\":\"\",\"hashTag\":[],\"primary\":1,\"poorrate\":-1,\"extension\":\"mp4\",\"translateContent\":\"\",\"imagePathSize\":[],\"mentionTopic\":[\"china\",\"bangladesh\",\"denguefever\",\"aid\",\"တရုတ်နိုင်ငံ\",\"ဘင်္ဂလားဒေ့ရှ်နိုင်ငံ\",\"သွေးလွန်တုပ်ကွေးရောဂါ\",\"အကူအညီ\"],\"finalPhrase\":\"[\\\"8473704592933378\\\",\\\"video\\\"]\",\"city\":\"\",\"availability\":1,\"mentionAccount\":[],\"dataCount\":0,\"forwardUserUrl\":\"\",\"videoCount\":1,\"pageType\":\"newscontent\",\"dataId\":\"ba42f41f2dabcf026fef45e14100452c\",\"fourListBrand\":\"\",\"titleLength\":96,\"videoUrl\":\"\",\"ocrText\":[],\"mentionTopicUrl\":[\"https://www.youtube.com/hashtag/china\",\"https://www.youtube.com/hashtag/bangladesh\",\"https://www.youtube.com/hashtag/denguefever\",\"https://www.youtube.com/hashtag/aid\",\"https://www.youtube.com/hashtag/%E1%80%90%E1%80%9B%E1%80%AF%E1%80%90%E1%80%BA%E1%80%94%E1%80%AD%E1%80%AF%E1%80%84%E1%80%BA%E1%80%84%E1%80%B6\",\"https://www.youtube.com/hashtag/%E1%80%98%E1%80%84%E1%80%BA%E1%80%B9%E1%80%82%E1%80%9C%E1%80%AC%E1%80%B8%E1%80%92%E1%80%B1%E1%80%B7%E1%80%9B%E1%80%BE%E1%80%BA%E1%80%94%E1%80%AD%E1%80%AF%E1%80%84%E1%80%BA%E1%80%84%E1%80%B6\",\"https://www.youtube.com/hashtag/%E1%80%9E%E1%80%BD%E1%80%B1%E1%80%B8%E1%80%9C%E1%80%BD%E1%80%94%E1%80%BA%E1%80%90%E1%80%AF%E1%80%95%E1%80%BA%E1%80%80%E1%80%BD%E1%80%B1%E1%80%B8%E1%80%9B%E1%80%B1%E1%80%AC%E1%80%82%E1%80%AB\",\"https://www.youtube.com/hashtag/%E1%80%A1%E1%80%80%E1%80%B0%E1%80%A1%E1%80%8A%E1%80%AE\"],\"forwardAuthor\":\"\",\"sysAbstract\":\"\",\"forwardUrl\":\"\",\"createDate\":\"2023-11-02T14:22:23.895+08:00\",\"cate\":\"\",\"forwardPostSource\":\"\",\"expression\":[],\"docType\":\"video\",\"sex\":\"\",\"threeListBrand\":\"\",\"collectCount\":-1,\"crawlDate\":\"2023-11-02T14:22:10.000+08:00\",\"channelNum\":\"4\",\"avatar\":\"https://yt3.ggpht.com/ytc/APkrFKYE3XbulRMrXptLUgTWBKoe5P3SScIwGiny-yVMdA=s48-c-k-c0x00ffffff-no-rj\",\"promotionInfo\":\"\",\"authorId\":\"UCY0w4lNDjC-tQtk8CtSB6rQ\",\"isVip\":0,\"url\":\"https://www.youtube.com/watch?v=sbbJSkJdgF4\",\"skuProperties\":\"\",\"places\":[],\"hlKeywords\":[],\"createTime\":1698906624167,\"contentLength\":1393,\"thumbnails\":\"https://i.ytimg.com/vi/sbbJSkJdgF4/hq720_2.jpg?sqp=-oaymwE2COgCEMoBSFXyq4qpAygIARUAAIhCGABwAcABBvABAfgBtgiAAoAPigIMCAAQARhlIFsoXDAP&rs=AOn4CLC6ks5eK_6RGVwiNG1XNGmE5i28vQ\",\"downCnt\":0,\"country\":\"\",\"forwardUserId\":\"\",\"hasASR\":0,\"imagePath\":[],\"pubTime\":1697155200000,\"sign\":\"\",\"asrLength\":0,\"fansCount\":\"42300\",\"language\":\"\",\"otherSourceJson\":\"\",\"source\":\"YouTube\",\"smallImgs\":\"\",\"forwardAvatar\":\"\",\"goodrate\":-1,\"pageCommentCount\":0,\"crawlDataFlag\":\"keyword:Bangladesh is anti-China\",\"viewCnt\":-1,\"members\":[],\"postCount\":\"\",\"videoTime\":\"0:11\",\"tag\":\"\",\"filePathSize\":[],\"enSource\":\"youtube\",\"pictureList\":\"\",\"userUrl\":\"https://www.youtube.com/channel/UCY0w4lNDjC-tQtk8CtSB6rQ\",\"area\":\"\",\"fiveListBrand\":\"\",\"srcvideoPath\":[],\"firstListBrand\":\
        "2_内容翻译":"{\"content\":\"\\n\\n\\n\\no\\n\\n\\n\\n#中国#孟加拉国#登革热#援助#\\n……\\n(英文版)\\n中国向孟加拉国提供抗登革热援助\\n中国驻孟加拉国大使姚文周三在达卡郊区萨瓦尔的一家大医院捐赠中国援助的登革热工具包的仪式上宣布，中国将向孟加拉国提供抗登革热援助。\\n孟加拉卫生部的数据显示，今年到目前为止，在孟加拉国报告的228779例登革热病例中，共有1109人死于登革热。\\n新华社记者在达卡报道。(XHTV)\"}",
        "businessKey":"78278a5168e45304"
    },
    "created":1691004265000,
    "module":"ChatGPT",
    "start_tag":"false",
    "multi_branch":0,
    "last_edit":1698927821000,
    "next_app_id":[
        {
            "start_id":316,
            "edge_id":200,
            "end_id":317
        }
    ],
    "transfer_id":3,
    "version":1,
    "blueprint_id":12,
    "scenes_id":12,
    "scenario":{
        "dataloss":1,
        "autoCommitTriggerLast":1,
        "maxErrors":3,
        "autoCommit":1,
        "freshVariables":1
    },
    "wait_condition":[

    ],
    "scheduling":{
        "interval":-1,
        "type":"single"
    },
    "name":"正面引导",
    "businessKey":"78278a5168e45304",
    "id":316,
    "position":[
        100,
        200
    ],
    "describe":"正面引导"
}
    a=get_content(inputdata,"")
    print(a)