I took 10 random DOIs from en.wiki.
You can see the results here.
10.1038/scientificamerican0200-90 still not working. - due to 401 at website pointing to log-in only .pdf. Not sure there's much we can do here.
10.2307/3677029 still says "JSTOR: An Error Occurred Setting Your User Cookie" . Forked to T93877
10.1542/peds.2007-2362 still gives " Check date values in: |date= (help)" . Forked to T95016
Hope this helps.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | • mobrovac | T101589 Automatic Cite generation results in malformed URL after fed valid LeMonde.fr link | |||
| Resolved | Mvolz | T100464 Insertion of <i> tag | |||
| Resolved | Mvolz | T94209 Results of a test with 10 random DOIs from en.wiki on the Beta site | |||
| Resolved | Mvolz | T93877 Accept and serve cookies | |||
| Resolved | • mobrovac | T95016 Citoid should be validating date fields | |||
| Resolved | Mvolz | T101878 Improve Zotero Validation | |||
| Resolved | Mvolz | T103238 Add crossref metadata request to lib/Scraper.js | |||
| Resolved | Mvolz | T95295 Create a cleaner import and export of natively scraped and zotero scraped citations | |||
| Resolved | Mvolz | T75993 Promisify Citoid and base it on service-template-node / service-runner | |||
| Resolved | Mvolz | T96363 Promisify html-metadata | |||
| Resolved | Mvolz | T104174 Add COinS scraping to html-metadata |
https://gerrit.wikimedia.org/r/#/c/199921/ should help when it gets merged, but I will check all of these with that change and give an update.
@Elitre, scratch that, they've been merged but aren't live yet :). I'll let you know when they are.
10.1038/scientificamerican0200-90 still not working.
10.2307/3677029 still says "JSTOR: An Error Occurred Setting Your User Cookie" .
10.1542/peds.2007-2362 still gives " Check date values in: |date= (help)" .
Note that you get the JSTOR error when using a standard JSTOR URL as well, like http://www.jstor.org/stable/25177324 - should that get a separate bug open?
The problem lies within CS1 templates like the Cite journal one, which do not accept dates with slashes, and is quickly fixed by replacing them with dashes (this is explained in https://en.wikipedia.org/wiki/Help:CS1_errors#bad_date , and it's good that the error message links to the explanation page, on en.wiki). So I don't know if this can be solved by changing the templates, or by having Citoid convert the date to an accepted format - I suspect the former. I can fork this task if Marielle says it's needed.
The problem with the scientific american one is it resolves to a pdf that requires you to log-in to access. 10.1038/scientificamerican0200-90
http://www.nature.com/scientificamerican/journal/v282/n2/pdf/scientificamerican0200-90.pdf
That website gives us a 401 response code, so we won't try to scrape it. If it had resolved to an actual pdf, we wouldn't be able to scrape that either, because we can't scrape actual pdfs.
In the future we may try to scrape data from certain non-200 response codes and try to generate a citation anyway, but probably not now. You could try filing this one as a site specific issue, as all these dois are probably bad since they try to point to pdfs that you we can't scrape, but I doubt much progress will be made here. Scientific american should make their DOIs point to abstract pages instead of log-in only pdfs :).