从Databricks向FHIR提交POST请求时遇编码解析错误

现象 #

• 从Databricks向FHIR插入动态生成的HL7 JSON数据时，返回400错误；但用Postman直接发送相同的POST请求可正常执行。
• 静态HL7 JSON在Databricks中测试无问题，动态生成的JSON在Postman中也能正常运行，仅在Databricks执行时触发错误。

错误详情 #

{"resourceType":"OperationOutcome","id":"21e75c73f1d000f2de7de02","meta":{"lastUpdated":"2022-08-23T18:30:17.9799097+00:00"},"issue":[{"severity":"error","code":"invalid","diagnostics":"Error occurred when parsing model: 'Unable to translate bytes [D3] at index 856 from specified code page to Unicode.'"}]}

Error Code: 400

动态生成JSON的代码 #

data = """
{
"resourceType" : "Patient",
"identifier" : [{ 
        "use" : "official", 
        "type" : { 
            "coding": [
            {
                "system": "http://terminology.hl7.org/CodeSystem/v2-0203",
                "code": "NNcol"
            }, {
                "system": "https://registraduria.gov.co/identificaciones",
                "code": "CC",
                "display": "Cedula de Ciudadania"
            }
        ] }, 
        "value" : """ +    str(v_id_paciente)   + """
},
{ 
    "use" : "usual", 
    "type" : { 
        "coding": [
        {
            "system": "http://terminology.hl7.org/CodeSystem/v2-0203",
            "code": "MR",
            "display": "Medical record number"    
        }
    ] }, 
    "value" :  """  + str(v_registro_medico) + """
}
],
"active": true,
"telecom": [
    {
        "system": "phone",
        "value": """ + str(v_telefono) + """,
        "use": "mobile"
    }
],
"name": [
    {
        "use": "official",
        "family": """ + str(v_nombre) + """
    }
],
"gender": """ + str(v_sexo) + """,
"birthDate": """ + str(v_fecha_nac) + """,
"address" : [
    {
        "text" : """ + str(v_dire_part) + """ ,
        "city": """ + str(v_region) + """
    }
]
}"""

Postman可正常运行的生成结果 #

{
"resourceType" : "Patient",
"identifier" : [{ 
        "use" : "official", 
        "type" : { 
            "coding": [
            {
                "system": "http://terminology.hl7.org/CodeSystem/v2-0203",
                "code": "NNcol"
            }, {
                "system": "https://registraduria.gov.co/identificaciones",
                "code": "CC",
                "display": "Cedula de Ciudadania"
            }
        ] }, 
        "value" : "17.003.627-K"
},
{ 
    "use" : "usual", 
    "type" : { 
        "coding": [
        {
            "system": "http://terminology.hl7.org/CodeSystem/v2-0203",
            "code": "MR",
            "display": "Medical record number"    
        }
    ] }, 
    "value" :  0.1449191882121713
}
],
"active": true,
"telecom": [
    {
        "system": "phone",
        "value": 9217432,
        "use": "mobile"
    }
],
"name": [
    {
        "use": "official",
        "family": "Patrick Mess"
    }
],
"gender": "male",
"birthDate": "1988-08-01",
"address" : [
    {
        "text" : "AV. CONCÓN 0132, VALPARAÍSO" ,
        "city": "VALPARAÍSO"
    }
]
}

解决方案 #

问题源于Databricks发送请求时的编码处理逻辑与Postman不同，需对含特殊字符的字段做Unicode规范化处理。处理代码如下：

vv_region = re.sub(
        r"([^n\u0300-\u036f]|n(?!\u0303(?![\u0300-\u036f])))[\u0300-\u036f]+", r"\1", 
        normalize( "NFD", vv_region), 0, re.I
    )

# 转换为NFC编码
N_v_region = normalize( 'NFC', vv_region)

print(N_v_region)

内容的提问来源于stack exchange，提问作者Gonza