The digests service serves content displayed at https://elifesciences.org/digests

While an article may contain a digest, a digest is also a separate thing with a friendlier title, a cover image and its own production workflow managed by the Features team and coordinated using the elife-bot.

The workflow to get content in to the digests service is roughly:

1. Features team creates a zip file containing a digest .docx file and a digest image and they place the zip file into an S3 bucket.
2. elife-bot runs an IngestDigest workflow that:
   • produces a simplifed .docx file and emails it to the Features team for production
   • copies the digest image to an output bucket
   • copies the .docx to an output bucket
   • transforms the .docx to JATS XML that is posted to the Typesetter's API endpoint
3. a VoR article zip file is eventually ingested that contains the final digest and the elife-bot will include the IngestDigestToEndpoint activity that:
   • will POST digest JSON content to the digest service endpoint as 'unpublished'
   • fetch the digest image via the IIIF server in order to include the image dimensions into the digest JSON
4. when that article is published the elife-bot workflow executes a PublishDigest activity that issues a POST to the digests service, setting the digest to 'published'

In case the digest was edited in the typesetter's interface between steps 2 and 3, it will use the content of the digest from the JATS XML of the article itself and not the original digest .docx file.

For consistency of environment, use the venv container:

docker-compose run venv pipenv install --dev uwsgi-tools
docker-compose run venv pipenv lock

docker-compose build
docker-compose run wsgi venv/bin/pytest app/digests/tests/test_digest_api.py

Images currently have to be re-built for modifications to any part of the code to be noticed.

The digests project is MIT licenced.