Updated elife-article-xml to: 2020-02-07 17:41:45 +0000

[elife-alfred-user]

I have run https://alfred.elifesciences.org/job/dependencies/job/dependencies-bot-lax-adaptor-update-elife-article-xml/157/display/redirect which resulted in this pull request.

[lsh-0]

re: fixes version of werkzeug

jarus/flask-testing#143

[lsh-0]

had a bunch of timeouts as we hammered iiif without a cache. I bet it's also responsible for loggly complaining about hitting out 80% quota for the day.

good to know that iiif still can't survive a complete backfill

kicking >:)

[lsh-0]

much better result:

[2020-02-11T03:37:34.905Z] Input: 13658 JSON articles
[2020-02-11T03:37:34.905Z] Output valid: 13657 JSON articles
[2020-02-11T03:37:34.905Z] Output invalid: 1 JSON articles
[2020-02-11T03:37:34.905Z] Validated green,red,total: 13657,1,13658
[2020-02-11T03:37:34.905Z] No more than 0 articles should be failing validation (with placeholders)
[2020-02-11T03:37:34.905Z] INFO - 2020-02-11 03:36:25,626 - elife-23292-v2.xml.json => failed

we still have an invalid article though:
bot-lax-validation-failure.log.txt

cc @gnott @giorgiosironi

[lsh-0]

dedicated werkzeug fix here: #604

[gnott]

The most likely reason I can see for the failure to validate output from elife-23292-v2.xml is a figure in the JSON output with a width: None, height: None. The particular image in IIIF is https://iiif.elifesciences.org/lax/23292%2Felife-23292-fig5-figsupp1-v2.tif, which is showing a width and height. I attribute the error to being a random glitch getting image dimentions from IIIF. I will re-run the pipeline tests to see if it will pass or repeat the same error.

[lsh-0]

to being a random glitch getting image dimentions from IIIF.

with the recreation of elife-libraries, the bot-lax caches were recreated as well, putting IIIF under a bit of stress. I'm pretty sure it dropped a number of requests initially. The next run was improved but it's possible that single request failed