.. _DEVNOTES:

Developer Notes
===============

Export Dependencies
-------------------

The `digitaledition-jekylltheme <https://github.com/emory-libraries-ecds/digitaledition-jekylltheme>`_
dependency is setup so as a minimal python module so it can be installed
as a python dependency.  It is currently included in `requirements/minimum.txt`,
which can be modified to require a branch or tagged version, and the
appropriate version is automatically installed as part of the fab deploy process.

However, note that the appropriate version of the
`teifacsimile-to-jekyll <https://github.com/emory-libraries-ecds/teifacsimile-to-jekyll>`_
Ruby gem must currently be installed manually.


Useful Queries
--------------

Find Volume objects in Fedora that do *not* have a cover image:

.. code-block:: sparql

    SELECT ?pid
    WHERE {
        ?pid <fedora-model:hasModel> <info:fedora/emory-control:ScannedVolume-1.0> .
        OPTIONAL {
            ?pid <http://pid.emory.edu/ns/2011/repo-management/#hasPrimaryImage> ?img
        }
        FILTER (!BOUND(?img))
    }

.. Note::
   Because Fedora's RISearch currently only supports SPARQL 1.0, this
   query requires the optional/not bound filter rather than the more
   straightforward NOT EXISTS that is supported in SPARQL 1.1+.

Other tasks
-----------

Sync volume content and annotations between instances
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

How to sync a volume with all its related content (relevant book and page
objects in Fedora) along with a set of annotations between two
environments, e.g. between production and QA.

In the source installation of Readux (e.g. production), use the
``find_test_pids`` script to get a list of volume, book, and page pids.
This reequires an input file, so create a file with a list of pids, one
per line (only one line if you are only synchronizing one volume)::

  echo "pid" > /tmp/vol_pid.txt
  python manage.py find_test_pids /tmp/vol_pid.txt > /tmp/vol_all_pids.txt

Use the eulfedora repo-cp script to sync the data between Fedora repositories
using the list of pids you generated::

  repo-cp prod qa --file /tmp/vol_all_pids.txt


Use the annotations api from your source installation to find the annotations
associated with your volume, e.g.:

  http://readux.library.emory.edu/annotations/api/search?user=<username>

or search by volume uri (available as of readux 1.4)::

  http://readux.library.emory.edu/annotations/api/search?volume_uri=http://readux.library.emory.edu/books/pid:###/

Save annotations as JSON and edit to replace the base source urls with
destination site urls  (e.g., readux.library to testreadux.library; but
note that this *must* match the url configured in your Django sites,
so if you are using a proxy use that url).  Then import the annotations
into your destination readux instance::

  python manage.py import_annotations my_annotations.json

Note that this requires equivalent user accounts to exist in both instances
(and if a different user happened to have the same username in the second
location, you have just given them access to another person's annotations).


Remove post-1922 yearbooks from the Solr index
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

For historical reasons, the Emory repository includes digitized yearbooks
that are not public domain but are inaccurately described as public
domain in the metadata.  For now, these must be suppressed from discovery
within Readux.

Start up a django python shell (``python manage.py shell``) and do
the following::


  from readux.utils import solr_interface
  from readux.books.models import VolumeV1_0, SolrVolume
  solr = solr_interface()
  vols = solr.query(content_model=VolumeV1_0.VOLUME_CONTENT_MODEL,
     collection_id='emory-control:LSDI-EmoryYearbooks').results_as(SolrVolume)
  solr.delete([{'pid': vol['pid']} for vol in vols if int(vol.volume[:4]) >= 1923])
  solr.query(content_model=VolumeV1_0.VOLUME_CONTENT_MODEL,
     collection_id='emory-control:LSDI-EmoryYearbooks').count()