How to fix <Response 500> error in python requests?

Question

I am using an API, which receives a pdf file and does some analysis, but I am receiving Response 500 always

Have initially tested using Postman and the request goes through, receiving response 200 with the corresponding JSON information. The SSL security should be turned off.

However, when I try to do request via Python, I always get Response 500

Python code written by me:

import requests

url = "https://{{BASE_URL}}/api/v1/documents"
fin = open('/home/train/aab2wieuqcnvn3g6syadumik4bsg5.0062.pdf', 'rb')
files = {'file': fin}
r = requests.post(url, files=files, verify=False)
print (r)
#r.text is empty

Python code, produced by the Postman:

import requests

url = "https://{{BASE_URL}}/api/v1/documents"

payload = "------WebKitFormBoundary7MA4YWxkTrZu0gW\r\nContent-Disposition: form-data; name=\"file\"; filename=\"aab2wieuqcnvn3g6syadumik4bsg5.0062.pdf\"\r\nContent-Type: application/pdf\r\n\r\n\r\n------WebKitFormBoundary7MA4YWxkTrZu0gW--"
headers = {
    'content-type': "multipart/form-data; boundary=----WebKitFormBoundary7MA4YWxkTrZu0gW",
    'Content-Type': "application/x-www-form-urlencoded",
    'cache-control': "no-cache",
    'Postman-Token': "65f888e2-c1e6-4108-ad76-f698aaf2b542"
    }

response = requests.request("POST", url, data=payload, headers=headers)

print(response.text)

Have masked the API link as {{BASE_URL}} due to the confidentiality

Response by Postman:

{
    "id": "5e69058e2690d5b0e519cf4006dfdbfeeb5261b935094a2173b2e79a58e80ab5",
    "name": "aab2wieuqcnvn3g6syadumik4bsg5.0062.pdf",
    "fileIds": {
        "original": "5e69058e2690d5b0e519cf4006dfdbfeeb5261b935094a2173b2e79a58e80ab5.pdf"
    },
    "creationDate": "2019-06-20T09:41:59.5930472+00:00"
}

Response by Python:

Response<500>

UPDATE:

Tried the GET request - works fine, as I receive the JSON response from it. I guess the problem is in posting pdf file. Is there any other options on how to post a file to an API?

Postman Response RAW:

POST /api/v1/documents
Content-Type: multipart/form-data; boundary=--------------------------375732980407830821611925
cache-control: no-cache
Postman-Token: 3e63d5a1-12cf-4f6b-8f16-3d41534549b9
User-Agent: PostmanRuntime/7.6.0
Accept: */*
Host: {{BASE_URL}}
cookie: c2b8faabe4d7f930c0f28c73aa7cafa9=736a1712f7a3dab03dd48a80403dd4ea
accept-encoding: gzip, deflate
content-length: 3123756

file=[object Object]

HTTP/1.1 200
status: 200
Date: Thu, 20 Jun 2019 10:59:55 GMT
Content-Type: application/json; charset=utf-8
Transfer-Encoding: chunked
Location: /api/v1/files/95463e88527ecdc94393fde685ab1d05fa0ee0b924942f445b14b75e983c927e
api-supported-versions: 1.0
X-Frame-Options: SAMEORIGIN
X-XSS-Protection: 1; mode=block
X-Content-Type-Options: nosniff
Strict-Transport-Security: max-age=31536000; includeSubDomains; preload
Referrer-Policy: strict-origin

{"id":"95463e88527ecdc94393fde685ab1d05fa0ee0b924942f445b14b75e983c927e","name":"aab2wieuqcnvn3g6syadumik4bsg5.0062.pdf","fileIds":{"original":"95463e88527ecdc94393fde685ab1d05fa0ee0b924942f445b14b75e983c927e.pdf"},"creationDate":"2019-06-20T10:59:55.7038573+00:00"}

CORRECT REQUEST

So, eventually - the correct code is the following:

import requests

files = {
    'file': open('/home/train/aab2wieuqcnvn3g6syadumik4bsg5.0062.pdf', 'rb'),
}
response = requests.post('{{BASE_URL}}/api/v1/documents', files=files, verify=False)
print (response.text)

Answer 1

A 500 error indicates an internal server error, not an error with your script.

If you're receiving a 500 error (as opposed to a 400 error, which indicates a bad request), then theoretically your script is fine and it's the server-side code that needs to be adjusted.

In practice, it could still be due a bad request though.

If you're the one running the API, then you can check the error logs and debug the code line-by-line to figure out why the server is throwing an error.

In this case though, it sounds like it's a third-party API, correct? If so, I recommend looking through their documentation to find a working example or contacting them if you think it's an issue on their end (which is unlikely but possible).