today's report

[Jason Priem]

did today:

• provenance url construction works; happens on creation of items in the ItemFactory, so can be used by everyone else down the chain. i don't think all the providers actually have the right urls in the configs, but that's just a matter of copying them down from wherever.
• refactored the items/item endpoint methods. focusing on the item (singular) method, as that's what the UI will use.
• item method now serves json or html depending on an extension (/item/:tiid.json or /item/:tiid.html). Left a slot to also allow switching based on request headers...is quite easy. but leaving it for now, in interests of getting something working. tried the https://github.com/nickstenning/negotiate library and gave up after fixing two bugs and finding a third.
• set up template switching based on genre (articles' metadata will look quite different from datasets', for instance.). I think in the future we'll let jinja switch in sub-templates, but seemed good to have the genre logic out of the presentation layer, at least at first.
• the article html display code mostly works...still have to finish the metrics bits. using BeautifulSoup to compare the expected/actual results in testing, as string matching over strings that hairy was becoming nightmarish.

tomorrow:

• finish the item html article. then, probably
• start on the collection report page.

holdups:

• I'm not totally grokking the backend changes because i didn't look at them much today. will want to spend some time on that.
• biblio still doesn't work as far as i know. display won't work without biblio.

woohoo, it's looking good!

j

--
Jason Priem
UNC Royster Scholar
School of Information and Library Science
University of North Carolina at Chapel Hill

[Heather Piwowar]

great work, Kevin and Jason!

• biblio still doesn't work as far as i know. display won't work without biblio.

I think biblio is working at least in part, as of this commit:

https://github.com/total-impact/total-impact-core/commit/bbcff8fdd00259be3db5bd673998c6e9936adfa5

for example, this seems to pass in test_alias_queue integration test:

assert_equals(resp_dict["biblio"]['data']["year"], "2010")

h