library/module/convertTxtServer.py

# coding:utf-8

from servicerd.client import RdClient
import grpc
from proto import wordToPdf_pb2_grpc
from proto import wordToPdf_pb2
from loguru import logger

MAX_MESSAGE_LENGTH = 256 * 1024 * 1024
RD = RdClient(rd_server="192.168.3.12:10021",
              service_name="Pdf2Txt",
              balance_type=3)


def read_into_buffer(filename):
    try:
        with open(filename, 'rb') as f:
            buf = f.read()
        f.close()
        return True, buf
    except Exception as e:
        logger.warning(e)
        return False, ""


@RD.fn_wrap
def call_say(requestFile: list, **kwargs):
    '''

    :param requestFile:
    :param kwargs:
    :return:
    '''
    address = '{}:{}'.format(kwargs['ip'], kwargs['port'])
    WordFileName = requestFile[0]
    Word = requestFile[1]
    print(address, "--->")
    try:
        with grpc.insecure_channel(address, options=[
            ('grpc.max_send_message_length', MAX_MESSAGE_LENGTH),
            ('grpc.max_receive_message_length', MAX_MESSAGE_LENGTH),
        ]) as channel:
            stub = wordToPdf_pb2_grpc.AnalysisDocumentsStub(channel)
            response = stub.Extract(wordToPdf_pb2.ParseRequest(WordFileName=WordFileName,
                                                               Word=Word))
        return response
    except Exception as e:
        print(e)
        return None


def convert_txt_start(pdf_file_path: str, suffix: str = "pdf")->(bool,str):
    '''
    转换pdf开始
    :param file_path:
    :param suffix:
    :return:
    '''
    state, word = read_into_buffer(pdf_file_path)
    if not state:
        return False, ""
    request_ret = call_say(["0.%s" % suffix, word])
    if not (request_ret and request_ret.State):
        return False, ""
    bytedata = request_ret.Pdf
    try:
        strdata = bytedata.decode("gbk","ignore").encode("utf-8").decode("utf-8")
    except Exception as e:
        print(e)
        strdata = bytedata.decode("utf-8")
    return True, strdata


if __name__ == '__main__':
    from util.file_operations import save_file

    st, request_ret = convert_txt_start("../data/0.docx", ".docx")
    print(request_ret)
    # state = save_file(request_ret, text_path)
    # print(state)