ASJLanguageSys/controller/ComprehendController.py

import boto3
from boto3 import Session
from django.views import View


from model.models import EmailSortModel, EmailConfigModel
from Utils.NeteaseMail import NeteaseMail
from django.views.generic import TemplateView

from object.ResponseObject import ResponseObject


class ComprehendView(TemplateView):

    def get(self, request, *args, **kwargs):
        request.encoding = 'utf-8'
        operation = kwargs.get('operation')
        return self.validation(request.GET, request, operation)

    def post(self, request, *args, **kwargs):
        request.encoding = 'utf-8'
        operation = kwargs.get('operation')
        return self.validation(request.POST, request, operation)

    def validation(self, request_dict, request, operation):
        response = ResponseObject()
        if operation is None:
            return response.json(444, 'error path')
        else:
            if operation == 'createSort':
                return self.createSort(request_dict, response)
            elif operation == 'doStartSortWord':
                return self.doStartSortWord(request_dict, response)
            elif operation == 'getDocumentClassifier':
                return self.get_document_classifier(request_dict, response)

    def createSort(self,request_dict, response):
        client = boto3.client('comprehend', region_name='region')

        # Create a document classifier
        create_response = client.create_document_classifier(
            InputDataConfig={
                'S3Uri': 's3://S3Bucket/docclass/file name'
            },
            DataAccessRoleArn='arn:aws:iam::account number:role/resource name',
            DocumentClassifierName='SampleCodeClassifier1',
            LanguageCode='en'
        )
        print("Create response: %s\n", create_response)

        # Check the status of the classifier
        describe_response = client.describe_document_classifier(
            DocumentClassifierArn=create_response['DocumentClassifierArn'])
        print("Describe response: %s\n", describe_response)

        # List all classifiers in account
        list_response = client.list_document_classifiers()
        print("List response: %s\n", list_response)


    def doStartSortWord(self,request_dict, response):
        client = boto3.client('comprehend', region_name='region')

        start_response = client.start_document_classification_job(
            InputDataConfig={
                'S3Uri': 's3://srikad-us-west-2-input/docclass/file name',
                'InputFormat': 'ONE_DOC_PER_LINE'
            },
            OutputDataConfig={
                'S3Uri': 's3://S3Bucket/output'
            },
            DataAccessRoleArn='arn:aws:iam::account number:role/resource name',
            DocumentClassifierArn=
            'arn:aws:comprehend:region:account number:document-classifier/SampleCodeClassifier1'
        )

        print("Start response: %s\n", start_response)

        # Check the status of the job
        describe_response = client.describe_document_classification_job(JobId=start_response['JobId'])
        print("Describe response: %s\n", describe_response)

        # List all classification jobs in account
        list_response = client.list_document_classification_jobs()
        print("List response: %s\n", list_response)


    def  get_document_classifier(self,request_dict, response):
        # aws_key = "AKIA2E67UIMD45Y3HL53"  # 【你的 aws_access_key】
        # aws_secret = "ckYLg4Lo9ZXJIcJEAKkzf2rWvs8Xth1FCjqiAqUw"  # 【你的 aws_secret_key】
        # session = Session(aws_access_key_id=aws_key,
        #                   aws_secret_access_key=aws_secret,
        #                   region_name="us-east-1")  # 此处根据自己的 s3 地区位置改变
        # client = session.client("asj-amazon-comprehend")
        # client = boto3.client(
        #     'asj-amazon-comprehend',
        #     aws_access_key_id='AKIA2E67UIMD45Y3HL53',
        #     aws_secret_access_key='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',
        #     region_name='us-east-1'
        # )
        #client = boto3.client(service_name='comprehend',region_name='us-east-1',aws_access_key_id='AKIA2E67UIMD45Y3HL53',aws_secret_access_key='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')

        # response = client.describe_document_classifier(
        #     DocumentClassifierArn='arn:aws:comprehend:us-east-1:697864307463:document-classifier/txttwo'
        # )
        # print(response)
        ec_qs =EmailConfigModel.objects.filter(emailtag=1) #获取总邮件

        for ec in ec_qs:

            email_server = NeteaseMail().loginEmail(ec['emailserver'], ec['fromaddr'], ec['password'], ec['emailserverport'])
            subject_list, from_list, bodydata_list = NeteaseMail().getEmailContext(email_server)
            NeteaseMail().closeEmail(email_server)

            for i ,v in enumerate(subject_list):

                client = boto3.client('comprehend')
                response = client.classify_document(
                    Text=bodydata_list[i],
                    EndpointArn='arn:aws:comprehend:us-east-1:697864307463:document-classifier-endpoint/test'
                )
                #print(response.index(max(response['Classes']['Score'])))
                nums = []
                for cls in response['Classes']:
                    print(cls['Name'])
                    nums.append(cls['Score'])

                print(response['Classes'][nums.index(max(nums))])

                es_qs = EmailSortModel.objects.filter(sort=response['Classes'][nums.index(max(nums))]['Name'])
                if es_qs.exists():
                    uemail_qs = EmailConfigModel.objects.filter(userid=es_qs[0]['userid'], langconfig__langcode='en')
                    if uemail_qs.exists():
                        for uem in uemail_qs:
                            NeteaseMail().sentemail(bodydata_list[i], subject_list[i], from_list[i], uem.fromaddr)

        return response.json(0)
        #print(response)

# if __name__ == '__main__':
#     ComprehendAction().get_document_classifier()