coding a API POST request to upload image

[modymp]

greetings!

I am trying to code a script that will help me post data to the Szurubooru instance. And the plan is convert this into an iOS/macOS Shortcut so I can upload images.

a beginner in Python or interpreting API doc and hence finding it hard to get this done. Any help appreciated!

• where-iam-struggling: I could not figure out how to "post" image data as per the API guidelines
• api: https://github.com/rr-/szurubooru/blob/master/doc/API.md#file-uploads
• api: https://github.com/rr-/szurubooru/blob/master/doc/API.md#creating-post

what works: a GET works fine

import requests
ENDPOINT = "https://booru/api"

get_headers = {
    'Content-Type': 'application/json',
    'Accept': 'application/json',
    "Authorization": "TOKEN LTRmOWUtOTAyNC0x",
}

get_response = requests.get(url=f"{ENDPOINT}/post/1", headers=get_headers, verify=False)

failing: a POST to upload a file fails

import requests

ENDPOINT = "https://booru/api"

authheader = {
'Authorization': 'TOKEN LTRmOWUtOTAyNC0x',
}

data17 = {
    "tags": "2process",
    "safety": "safe",
    "metadata": open("Downloads/bor/unknown.png", 'rb'),
}

post_response = requests.post(url=f"{ENDPOINT}/posts/", verify=False, headers=authheader, files=data17)
print(post_response.text)

As I was not certain of adhering to the API guidelines I did try to seek help on forums that are specific for Python and on Reddit.. I could not make progress, so, I am trying here to check..

I tried to modify the code by reading a few articles & other GitHub projects, but I am sure I missing some thing.

• https://blog.jetbridge.com/multipart-encoded-python-requests/
• https://stackoverflow.com/questions/9733638/how-to-post-json-data-with-python-requests
• https://github.com/reluce/szurubooru-toolkit/tree/main/src/szurubooru_toolkit

thanks!

[neobooru]

Just putting this here for completeness as this has been solved in #572

I used noirscape's code and adapted it a bit.

import base64
import requests


def btoa(txt: str):
    return base64.b64encode(txt.encode()).decode()


# conventional names for a session include session, s and r. i use session here. pick your poison.
# the auth token is created by taking base64 encoding the `username:login_token`
# you can find the login token in the webui under Account -> Login tokens
session = requests.Session()
session.headers = {
    "Accept": "application/json",
    "Authorization": "Token " + btoa("username:01ec0fd6-2dd4-40a6-907a-c36dfec07592"),
}

API_URL = "https://booru/api"

# context manager means that the file handle will always be properly closed, even if something fails.
# some operating systems don't clean up loose file handles which can lead to OS weirdness.
with open("image.png", "rb") as uploadfile:
    # first we upload the file to temporary storage and get the token from the response
    temporary_data = session.post(
        f"{API_URL}/uploads", files={"content": uploadfile}
    ).json()
    file_token = temporary_data["token"]

    # here you can use that token for anything else you want to do pre-upload, such as reverse image search.

    # next we upload the file for real
    # you can add your own parameters here - token and safety are the only required ones.
    # setting content-type here explicitly means its added to the session headers
    post_data = session.post(
        f"{API_URL}/posts",
        json={"contentToken": file_token, "safety": "safe", "tags": ["tag1", "tag2"]},
        headers={"Content-Type": "application/json"},
    )

    print(post_data.json())  # finally we print the uploaded post data.