Robot Framework中含PDF的Multipart POST请求实现问题

问题根源 #

你的代码在Evaluate步骤中直接通过字符串插值传递二进制PDF数据，导致bytes类型被错误转换为字符串，引发格式问题；另外使用Get File For Streaming Upload时，没有正确处理返回的文件对象。

解决方案一：正确在Evaluate中引用二进制数据 #

避免直接在Evaluate的字符串中插入${filedata}，改用Python的globals()函数引用Robot Framework变量，保留bytes类型的原始状态：

Create a session for indexing api
        Create Session    indexing_session    ${base_url}    verify=True

Execute indexing api for mortagenumber "${mortgagenumber}"
        [Arguments]    ${type}    ${description}    ${sendtime}    ${source}    ${filename}
    
        ${filedata}=    Get Binary File   ${CURDIR}/../testdata/${filename}
        
        ${token}=        Get new bearer token
        ${headers}=      Create Dictionary    Authorization=Bearer ${token}
        # 通过globals()引用RF变量，避免二进制数据被错误转换
        ${formdata}=     Evaluate    {'type': (None, '${type}', 'text/plain'), 'description': (None, '${description}', 'text/plain'),'sendTime': (None, '${sendtime}', 'text/plain'), 'source': (None, '${source}', 'text/plain'), 'files': ('${filename}', globals()['filedata'], 'application/pdf')}
        ${response}=     POST On Session   indexing_session    /indexing/${mortgagenumber}    headers=${headers}    files=${formdata}
    
        ${json}=    Set Variable  ${response.json()}
        Set Global Variable    ${json}

解决方案二：使用Get File For Streaming Upload优化文件传递 #

Get File For Streaming Upload返回可直接用于流式上传的文件对象，无需手动读取二进制数据，结合Create Dictionary和Set To Dictionary更直观地构建请求参数：

Create a session for indexing api
        Create Session    indexing_session    ${base_url}    verify=True

Execute indexing api for mortagenumber "${mortgagenumber}"
        [Arguments]    ${type}    ${description}    ${sendtime}    ${source}    ${filename}
    
        ${file_object}=    Get File For Streaming Upload   ${CURDIR}/../testdata/${filename}
        
        ${token}=        Get new bearer token
        ${headers}=      Create Dictionary    Authorization=Bearer ${token}
        ${formdata}=     Create Dictionary
        # 添加纯文本参数
        Set To Dictionary    ${formdata}    type=(None, '${type}', 'text/plain')
        Set To Dictionary    ${formdata}    description=(None, '${description}', 'text/plain')
        Set To Dictionary    ${formdata}    sendTime=(None, '${sendtime}', 'text/plain')
        Set To Dictionary    ${formdata}    source=(None, '${source}', 'text/plain')
        # 添加文件对象
        Set To Dictionary    ${formdata}    files=('${filename}', ${file_object}, 'application/pdf')
        
        ${response}=     POST On Session   indexing_session    /indexing/${mortgagenumber}    headers=${headers}    files=${formdata}
    
        ${json}=    Set Variable  ${response.json()}
        Set Global Variable    ${json}

关键注意事项 #

1. 不要手动设置Content-Type头：POST On Session底层依赖requests库，会自动为multipart请求生成正确的Content-Type: multipart/form-data; boundary=xxx头，手动设置会覆盖该头导致请求格式错误。
2. 参数格式规范：files字典中，纯文本项格式为(None, 参数值, 'text/plain')（None表示不需要关联文件名）；文件项格式为(文件名, 文件对象/二进制数据, MIME类型)。

内容的提问来源于stack exchange，提问作者Peter Franssen