[Tumblr] Blogs who only accept a "https://www.tumblr.com/<user name>/<ID>/" fail to download

[arisboch]

There are some Tumblr blogs, that won't accept a "https://<user name>.tumblr.com/post/" url, but only a "https://www.tumblr.com/<user name>/<ID>/" url. An attempt to download from these blogs fails. One of thee blogs would be https://www.tumblr.com/suf-fering. Could you take a look at it, please?

[mikf]

I'll look into it, but you can also manually rewrite those URLs to put them into the format expected by gallery-dl: https://suf-fering.tumblr.com/

[mikf]

So I took a look, but I'm not really sure what the problem is supposed to be.

Both https://suf-fering.tumblr.com/
as well as https://www.tumblr.com/suf-fering work as a blog URL, and

https://www.tumblr.com/suf-fering/700951855904227328,
https://www.tumblr.com/blog/suf-fering/700951855904227328,
https://www.tumblr.com/blog/view/suf-fering/700951855904227328,
https://suf-fering.tumblr.com/post/700951855904227328, and
https://suf-fering.tumblr.com/image/700951855904227328
get accepted as post URLs.

[arisboch]

I have no idea what kinda sorcery is happening on my computers, but it doesn't work with either of these URLs, I'm gonna post the verbose output and the Tumblr section of the .conf file:

gallery-dl --verbose  https://www.tumblr.com/suf-fering/700951855904227328/oomelet-i-want-to-see-her-in-another-outfit
[gallery-dl][debug] Version 1.25.0-dev
[gallery-dl][debug] Python 3.9.2 - Linux-5.10.102.1-microsoft-standard-WSL2-x86_64-with-glibc2.31
[gallery-dl][debug] requests 2.28.2 - urllib3 1.26.14
[gallery-dl][debug] Configuration Files ['${HOME}/.gallery-dl.conf']
[gallery-dl][debug] Starting DownloadJob for 'https://www.tumblr.com/suf-fering/700951855904227328/oomelet-i-want-to-see-her-in-another-outfit'
[tumblr][debug] Using custom api_key authentication
[tumblr][debug] Using TumblrPostExtractor for 'https://www.tumblr.com/suf-fering/700951855904227328/oomelet-i-want-to-see-her-in-another-outfit'
[urllib3.connectionpool][debug] Starting new HTTPS connection (1): api.tumblr.com:443
[urllib3.connectionpool][debug] https://api.tumblr.com:443 "GET /v2/blog/suf-fering.tumblr.com/posts?id=700951855904227328&offset=0&limit=50&reblog_info=true&api_key=[redacted] HTTP/1.1" 404 150
[tumblr][error] NotFoundError: Requested user or post could not be found

        "tumblr":
        {
            "avatar": false,
            "external": false,
            "inline": true,
            "posts": "all",
            "reblogs": true,
	    	"api-key": "[redacted]",
	    	"api-secret": "[redacted]",
	    	"filename": "tumblr {blog[name]} {id} {num:>02}.{extension}",
	    	"directory": ["tumblr {blog[name]} {id}"],
	    	"postprocessors": 
	    	[{
		 		"name": "metadata",
				"event": "post",
				"filename": "tumblr {blog[name]} {id} {num}.html",
				"mode": "custom",
				"format": "<meta charset='UTF-8'/>{body}<br><br>{caption}<br><br>{question}<br>{answer}<br>{tags!S}"
			}]
        },

Maybe I should create a new API key for gallery-dl? Is it possible, that another .conf setting is somehow affecting the retrieval? Is it possible, that the python environment is somehow borked?

[arisboch]

https://www.tumblr.com/hismajestythebeggar

That one also doesn't work (warning, NSFW) ;(

[mikf]

Found the problem: Those blogs are "dashboard-only".

You need all four OAuth tokens to access them:

• api-key and api-secret, which you already have
• access-token and access-token-secret, which you can get by running gallery-dl oauth:tumblr

You should see OAuth1.0 authentication instead of api-key authentication in the debug output when using all four tokens.

[tumblr][debug] Using custom OAuth1.0 authentication

[arisboch]

I tried that, got the following URL back:

http://localhost:6414/?oauth_token=[redacted]&oauth_verifier=[redacted]#_=_

The "oauth_token" from the URL is the "access-token", and the "oauth_verifier" is the "access-token-secret", right? I tried that, but it still doesn't work ;-(

[gallery-dl][debug] Version 1.25.0-dev
[gallery-dl][debug] Python 3.9.2 - Linux-5.10.102.1-microsoft-standard-WSL2-x86_64-with-glibc2.31
[gallery-dl][debug] requests 2.28.2 - urllib3 1.26.14
[gallery-dl][debug] Configuration Files ['${HOME}/.gallery-dl.conf']
[gallery-dl][debug] Starting DownloadJob for 'https://www.tumblr.com/suf-fering/700951855904227328/oomelet-i-want-to-see-her-in-another-outfit'
[tumblr][debug] Using custom OAuth1.0 authentication
[tumblr][debug] Using TumblrPostExtractor for 'https://www.tumblr.com/suf-fering/700951855904227328/oomelet-i-want-to-see-her-in-another-outfit'
[urllib3.connectionpool][debug] Starting new HTTPS connection (1): api.tumblr.com:443
[urllib3.connectionpool][debug] https://api.tumblr.com:443 "GET /v2/blog/suf-fering.tumblr.com/posts?id=700951855904227328&offset=0&limit=50&reblog_info=true HTTP/1.1" 404 None
[tumblr][debug] {'meta': {'status': 404, 'msg': 'Not Found'}, 'response': [], 'errors': [{'title': 'Not Found', 'code': 4012, 'detail': 'This Tumblr is only viewable within the Tumblr dashboard'}]}
[tumblr][info] Run 'gallery-dl oauth:tumblr' to access dashboard-only blogs
[tumblr][error] AuthorizationError: This Tumblr is only viewable within the Tumblr dashboard

What am I doing wrong?

[mikf]

You didn't do anything wrong. I made a mistake in commit 70ce45d which stopped any oauth procedure before it could finish and display the two access-token values.

oauth_token and oauth_verifier are intermediate values that are supposed to be send to Tumblr, but that never happened due to this bug. Fixed in 75570ad, by the way.

[arisboch]

@mikf Thanks a lot, now all works perfectly!