ai_dataset/import_data/peyma_dataset.py

import json
import requests
from decimal import Decimal
import os

TOKEN = 'eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpYXQiOjE3MTg0NTE3MTUsImp0aSI6InNWaDljNkdlRWdEK0IzM1N2UHNzbXkybkxUK0tWWnBjIiwiaXNzIjoiaHR0cHM6XC9cL2NwLnRhdmFzaS5pciIsImV4cCI6MTcxOTc1MTcxNCwiYXVkIjoiaHR0cHM6XC9cL2NwLnRhdmFzaS5pciIsImRhdGEiOnsiaWQiOjQwLCJmaXJzdF9uYW1lIjoiXHUwNjM5XHUwNjQ1XHUwNjI3XHUwNjMxIiwibGFzdF9uYW1lIjoiXHUwNjJjXHUwNjQ4XHUwNmE5XHUwNjI3XHUwNjMxIiwiZW1haWwiOiJham9rYXI5MUB5YWhvby5jb20iLCJ1c2VybmFtZSI6ImFqb2thciIsInVzZXJfbGV2ZWwiOjF9fQ.frkLaR1HYyfZ8sFhLtiuEZaBPwDPWKSxKDQpql6aOZc'
ACCEPT = "application/json"
HEADERS = {"Authorization": TOKEN, "Accept": ACCEPT}

url = "https://api.tavasi.ir/repo/dataset/multi/add/peyma/ner"
headers = HEADERS

address = os.getcwd()
if 'impoert_data' in address:
    address += '/data/peyma_train.txt'
else:
    address += '/impoert_data/data/peyma_train.txt'
# باز کردن فایل متنی
with open(address, 'r', encoding='utf-8') as file:
    input_text = file.read()

# تبدیل متن به لیستی از خطوط
lines = input_text.strip().split('\n')

key = ''
begin = -1
end = -1
tokenNumber = -1
content = ''
result_token = []

class JSONEncoder(json.JSONEncoder):
    def default(self, obj):
        if isinstance(obj, Decimal):
            return float(obj)
        return json.JSONEncoder.default(self, obj)



# def createIndex(content, result_token):
#     output = {
#         "content": content,
#         "domain": "پیکره پیما",
#         "ref_id": "",
#         "ref_url": "",
#         "result_token": result_token,
#     }
#     # print(output)
#     # print(json.dumps(output, indent=4, ensure_ascii=False))
#     return output

def createIndex(content, result_token):
    result_objects = [{
          "task":"ner",
          "key":"peyma_ner",
          "label":"خروجی تیم پیما",
          "values":result_token
          } ]
    output ={
        "content": content,
        "domain": "پیکره پیما",
        "ref_id": "",
        "ref_url": "",
        "result_objects": result_objects,
    }
    # print(output)
    # print(json.dumps(output, indent=4, ensure_ascii=False))
    return output

def appendResultToken(text, key, begin, tokenNumber, result_token):
    end = -1
    # if key == 'HALFREFERENCE' :
    #     key = 'H_REF'
    # elif key == 'REFERENCE' :
    #     key = 'REF'
    if key:
        if key == 'ORG' :
            key = 'ORG'
        elif key == 'MON' :
            key = 'MON'
        elif key == 'LOC' :
            key = 'LOC'
        elif key == 'DAT' :
            key = 'DAT'
        elif key == 'TIM' :
            key = 'TIM'
        elif key == 'PER' :
            key = 'PER'
        elif key == 'PCT' :
            key = 'PCT'

        end = tokenNumber -1
        result_token.append({
        "begin": begin,
        "end": end,
        "result_key": key,
        "text" : text
        }) 
        begin = -1
        end = -1
        key = ''
    return  key, begin, end, result_token  

bulk_data = []
bulk_count = 1
count = 0
text = ''

for i, line in enumerate(lines):
    print('line: ' + str(i))
    count += 1
    tokenNumber += 1

    if line.strip() == '' : 
       key, begin, end, result_token = appendResultToken(text, key, begin, tokenNumber, result_token)

       data = createIndex(content, result_token) 
       tokenNumber = -1
       content = ''
       result_token = []

       bulk_data.append(data)    
       bulk_count +=1 
       if bulk_data.__len__() > 100:
            print('=' * 30 )
            print('count ' + str(count))
            payload = json.dumps(bulk_data, cls=JSONEncoder)  #Works!
            response = requests.request("POST", url, headers=headers, data=payload)
            print(response)
            bulk_data = []
            bulk_count = 1

       continue

    parts = line.split('|')
    if len(parts) != 2:
        continue

    result_key = parts[1]
    token_part = (parts[0].strip()).split()
    if len(token_part) > 1:
        new_token_part = token_part[0].strip() + '\u200c' + token_part[1].strip()
        content += ' ' + parts[0]
    else:
        content += ' ' + parts[0]
        
    if result_key.startswith('I_'):    
        text += ' ' + parts[0]
        continue

    if result_key == 'O' :     
        key, begin, end, result_token = appendResultToken(text, key, begin, tokenNumber, result_token)

    if result_key.startswith('B_'):        
        key, begin, end, result_token = appendResultToken(text, key, begin, tokenNumber, result_token)

        text = parts[0]
        begin = tokenNumber
        end = -1
        key = result_key.replace('B_', '')


if content != '' :
    key, begin, end, result_token = appendResultToken(text, key, begin, tokenNumber, result_token)
    data = createIndex(content, result_token) 
    bulk_data.append(data)    
    bulk_count +=1 


if  bulk_data.__len__() > 0:
    print(bulk_count)
    payload = json.dumps(bulk_data, cls=JSONEncoder)  #Works!
    response = requests.request("POST", url, headers=headers, data=payload)
    print(response.text)
    
# نمایش دیکشنری خروجی به صورت JSON
print("*****************   end ")