Python多边界API调用失败，请求代码修正指导

我在Postman里有个能正常运行的API调用，想在Python里复现。修改Postman提供的模板代码后，API调用无法正常工作。用Wireshark对比Postman的正常请求和我的Python请求，发现第一个边界缺少application/json类型，也不确定JSON字符串传递是否正确。

Wireshark对比发现的问题 #

Postman的请求中，documentAttributes字段带有Content-Type: application/json标识，但我的Python请求里该字段没有这个类型声明。

当前Python请求体内容 #

>>> response.request.body
b'--affc9f680a99e3d8509e8c9f24b422b3\r\nContent-Disposition: form-data; name="documentAttributes"\r\n\r\n{"batchInfo":{"batchName":"YMTESTALIAS","createDateTime":"2023-05-22T14:00:00.000Z"},"batchDocumentInfo":{"documentTypeName":"Photo Identification","subject":"Alias Test","metadata":[{"fieldName":"MRN","fieldValue":"9305940"},{"fieldName":"ENCTR_NBR","fieldValue":"65013813"}]},"processQueue":"BATCH","indexImmediately":"true"}\r\n--affc9f680a99e3d8509e8c9f24b422b3\r\nContent-Disposition: form-data; name="documentContent"; filename="hello.tif"\r\nContent-Type: image/tiff\r\n\r\nII*\x002\x19\x00\x00\x80?\xe0@\x0

我的Python代码 #

import requests
import json

domain = 'cert'
client_mnemonic = 'xxxx_fl'
auth_token = 'Basic .....=='
file_path = 'C:/Users/system/Desktop/hello.tif'

url = f'http://1.1.1.1/docimaging-access-1.0/{domain}.{client_mnemonic}.cernerasp.com/service/documents'

batch_name = 'TESTALIAS'
document_type = 'Photo Identification'
subject = 'Alias Test'
mrn = '9305940'
fin = '65013813'

# build payload
payload = {
    'documentAttributes': {
        'batchInfo': {
            'batchName': f'{batch_name}',
            'createDateTime': '2023-10-31T14:00:00.000Z'
        },
        'batchDocumentInfo': {
            'documentTypeName': f'{document_type}',
            'subject': f'{subject}',
            'metadata': [
                {
                    'fieldName': 'MRN',
                    'fieldValue': f'{mrn}'
                },
                {
                    'fieldName': 'ENCTR_NBR',
                    'fieldValue': f'{fin}'
                }
            ]
        },
        'processQueue':'BATCH',
        'indexImmediately':'true'
    }
}

headers = {
    'Authorization': auth_token
}

files = {
    'documentContent': ('hello.tif', open(file_path,'rb'), 'image/tiff')
}

response = requests.post(url, headers=headers, data=payload, files=files)

修复方案 #

问题根源是requests库同时使用data和files参数时，会把data中的字段默认以普通文本格式发送，无法为单个字段指定application/json类型，和Postman的请求结构不匹配。

修改后的代码：

import requests
import json

domain = 'cert'
client_mnemonic = 'xxxx_fl'
auth_token = 'Basic .....=='
file_path = 'C:/Users/system/Desktop/hello.tif'

url = f'http://1.1.1.1/docimaging-access-1.0/{domain}.{client_mnemonic}.cernerasp.com/service/documents'

batch_name = 'TESTALIAS'
document_type = 'Photo Identification'
subject = 'Alias Test'
mrn = '9305940'
fin = '65013813'

# 将documentAttributes序列化为JSON字符串
document_attributes = json.dumps({
    'batchInfo': {
        'batchName': f'{batch_name}',
        'createDateTime': '2023-10-31T14:00:00.000Z'
    },
    'batchDocumentInfo': {
        'documentTypeName': f'{document_type}',
        'subject': f'{subject}',
        'metadata': [
            {
                'fieldName': 'MRN',
                'fieldValue': f'{mrn}'
            },
            {
                'fieldName': 'ENCTR_NBR',
                'fieldValue': f'{fin}'
            }
        ]
    },
    'processQueue':'BATCH',
    'indexImmediately':'true'
})

headers = {
    'Authorization': auth_token
}

# 把JSON字段和文件一起放入files参数，为JSON字段指定Content-Type
files = {
    'documentAttributes': (None, document_attributes, 'application/json'),
    'documentContent': ('hello.tif', open(file_path,'rb'), 'image/tiff')
}

response = requests.post(url, headers=headers, files=files)

关键修改说明 #

1. 将documentAttributes的字典结构序列化为JSON字符串，不再放在data参数中
2. 在files参数中定义documentAttributes字段：第一个参数None表示该字段不是文件（无文件名），第二个参数是JSON字符串，第三个参数指定Content-Type: application/json
3. 移除data参数，所有表单字段都通过files传递，确保每个字段的Content-Type符合API要求

内容的提问来源于stack exchange，提问作者Yitzhak