CODEWITHSUNDEEP
pythonpython-3.xtumblr
errata
errata

Reputation: 6031

Tumblr API posting to secondary blog

I am trying to use Tumblr API to post something to my secondary blog. I am using Tumblpy Python package because official Tumblr client is for Python 2 only and it seems abandoned anyway.

I authorized the app and I took a look at Tumblr console where I picked up all the keys from. I started to play around with it and I have noticed that I can successfully post to my primary blog, but not on secondary (getting {TumblpyError} 404 'There was an error making your request.' error all the time).

This is the code I've been trying out:

from tumblpy import Tumblpy


def post_tumblr(
        url,
        comment='',
        tags='',
        **kwargs
):
    t = Tumblpy(
        APP_KEY, APP_SECRET,
        OAUTH_TOKEN, OAUTH_TOKEN_SECRET
    )

    blog_url = t.post('user/info')
    blog_url = blog_url['user']['blogs'][0]['url']  # POSTING TO PRIMARY BLOG WORKS
    # blog_url = blog_url['user']['blogs'][1]['url']  # CANNOT POST TO SECONDARY BLOG?

    post_url = t.post(
        'post',
        blog_url=blog_url,
        params={
            'type': 'video',
            'embed': url,
            'caption': comment,
            'tags': tags,
        }
    )

    return True

Is there some catch in posting to secondary blog in Tumblr API?

Upvotes: 1

Views: 240

Answers (1)

ceyko
ceyko

Reputation: 4852

This is an interesting one. The first thing to note is that in the Tumblr API, any blog.url is a full url, including the scheme: http:// or https://. However, in any /blog/{blog-identifier}/* API request, it does not accept a full url, only the hostname: example.com or demo.tumblr.com.

Given this information, it seems like neither of your requests should work, so I checked out the code for Tumblpy a bit. It turns out that Tumblpy does accept a full url for a {blog-identifier}, by stripping away everything except the hostname. However, it only works with http:// urls, and not https:// urls, as seen here.

If you have a blog with SSL enabled, the url field will be an https:// link, and then Tumblpy will not handle it correctly. I'm assuming this is the case for your secondary blog; you can check on your settings page for that blog.

Assuming this is the case, the best fix will be to just construct the hostname yourself before calling Tumblpy.post(). Just use user.blogs[i].name + ".tumblr.com". Or, you can parse the hostname from the url too. Either approach will work.

Upvotes: 1

Related Questions