comments

[bdarcus]

So it might just be me, and the end of a long day, but I'm struggling a bit with the API.

Can you perhaps include a hello world that just lists entries in a library and prints out the titles?

More specifically, I guess I'm looking for really straightforward methods like:

• library.items - list of item dictionaries, maybe with optional filter parameter
• library.item - item by id
• library.collections - list of collections

So that I can just do:

for item in library.items:
    print item.title

Or:

library.item(125429).title

[urschrei]

At the moment, a basic hello world:

import pyzotero.zotero as z
zot = z.Zotero(user id, user key)
items = zot.items_data('top_level_items')
for i in items:
   print i['title']

This would return the first 50 top-level library items (the API will only retrieve 50 at a time), and print their titles. I'm not currently abstracting the returned data into new objects (such as Items, Collections, Groups), just returning lists containing dicts, mainly because I haven't gotten as far as thinking about what people might want to do with them once they've been retrieved. Admittedly, this is a bit minimal. I'll have a think about how best to deal with this. One problem that leaps out is how to create properties for a notional Item object when properties vary from item to item; they always have titles and IDs, but not always abstracts, publishers etc. I suppose I could just add them as properties on the fly, and handle missing properties gracefully with a custom error. I could also give methods like items_data more intuitive names. Let me know what you think.

[urschrei]

The latest commit on the dev branch: SHA: 937cbf42208638eaa4e7a52cd1781dc0ed955ab7 creates Item objects for any calls to items_data. So you can do: import pyzotero.zotero as z zot = z.Zotero(user id, user key) items = zot.items_data('top_level_items') for item in items: print 'Author: %s\nTitle: %s\nID: %s' % (item.author, item.title, item.id)

You'll just get None for any attribute that doesn't exist. I'm not 100% sure this is the way to go (since creating instance attributes on the fly is Bad), but still.

[bdarcus]

Not sure if this is helpful in terms of ideas, but I came across a twitter client library a bit ago which sounded interesting:

http://mike.verdone.ca/twitter/

In particular, the statement "Its methods are directly bound to the Twitter API's URLs."

[urschrei]

Hmm. I could do that, it's only about 25 methods.

[urschrei]

SHA: 9c34318376a94d9274fccad83720841c161e0fad (https://github.com/urschrei/pyzotero/tree/dev)

I like this a lot better, even though it's going to entail a lot more work to implement every possible API call.

[bdarcus]

I tried this again, using the example in the updated README, and got this error:

$ python zot.py 
Traceback (most recent call last):
  File "zot.py", line 4, in <module>
    items = zot.items()
  File "/usr/local/lib/python2.6/dist-packages/pyzotero/zotero.py", line 109, in wrapped_f
    return eval(output_func)(retr)
  File "/usr/local/lib/python2.6/dist-packages/pyzotero/zotero.py", line 353, in items_data
    return self.standard_items(retrieved)
 File "/usr/local/lib/python2.6/dist-packages/pyzotero/zotero.py", line 367, in standard_items
   [e.text.lower() for e in elem.iter('th')])
AttributeError: _ElementInterface instance has no attribute 'iter'

[urschrei]

I suspect you have an older version of elementtree installed. Can you try pip install elementtree --upgrade, then re-running your script and/or tests.py

[bdarcus]

I ended up upgrading the default ubuntu 2.6 to 2.7; that solved it.