everything() returns a 403

[stakats]

zot.top() functions fine, but zot.everything(zot.top()) generates a 403 using a private individual library. Is it possible that the API key is being dropped in the everything call? In my script output I see URL: https://api.zotero.org/users//items/top?limit=5&content=json&start=5 with no key appended.

[stakats]

I copied the wrong output URL, but the situation is the same: https://api.zotero.org/users/3/items/top?content=json&start=50

[urschrei]

I wasn't getting a 403, but it was still falling over, due to an unrelated bug. Can you try again using the dev branch?

[stakats]

Tried again with the dev branch and still encountered a 403.

[urschrei]

I can't reproduce this, running v0.9.7 (the latest version on PyPI). I just tried again, and retrieved all 606 items in my library.

Are you running Python 2.7?

[stakats]

Yes, running 0.9.7 with Python 2.7.1. The problem seems to be that when everything() gets to items.extend(self.follow()), it's just chasing down link rel="next" which in my case is https://api.zotero.org/users/3/items/top?content=json&start=50. Since no API key is appended in the next links, it fails.

[urschrei]

I've done some more testing, omitting my API key, and the top() call still works, even though I should be getting an immediate 403. I'll post to zotero-dev before doing anything else, but unless I'm missing something obvious, there are two issues here:

• a simple bug in Pyzotero, causing your subsequent top() calls to fail (correctly, as it turns out), which should be an easy fix
• some as-yet-unknown problem with my library ID, which is causing API calls to be accepted, when they should be failing with a 403.

[stakats]

Easily solved: 1) there's a bug in pyzotero; 2) your library is public. If you test with a private library, you'll find the bug.

[urschrei]

Hah. That explains everything.

[urschrei]

OK, commit c8fab2c360451e53935c42c327a1654dd3f167f5 on the dev branch should fix this.