alphagov / digitalmarketplace-content-loader / 252 / 1
99%
master: 99%

LAST BUILD BRANCH: bk-add-boolean
DEFAULT BRANCH: master
Ran
Files 8
Run time 0s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 99.228% (-0.002%) from 99.23%
252.1

push

travis-ci

galund 
Taxonomy (category tree) implementation - simplification

  - https://trello.com/c/y79rP1KA/197-suppliers-choose-a-category-backend
  - less is more - don't need to build another tree structure
  - also tidy up inheritance for summary, multiple inheritance avoided
    as not needed

Explanation of why we are storing the parent values:

  - really, these additional categories are a denormalization. Arguably,
    the frontend should just assume that they are selected - the place
    they actually need to be selected is in the search engine;
  - however, the indexer doesn't have access to the tree at the moment,
    so moving the denormalization there is rather hard;
  - the next best place to put this might then be the API, because that
    would ensure a consistent view on the data. But that would require
    teaching the API about this as a special field, and no other fields
    get processing applied - either the data posted in is accepted, or
    not (i.e. we only have validation, nothing beyond that). And again,
    the API doesn't currently have access to the original framework data,
    so even if we wanted to have the additional categories added there,
    it's hard in the current structure.
  - that leaves us with having to do the denormalization in the frontends
    (i.e. in the content loader)
  - having the API validate that the denormalization was applied means
    duplicating that content loader logic into the API, and again we don't
    actually have the data available that would allow us to do that anyway...
  - and therefore data validly posted to the API may not have the
    denormalization applied (i.e. if someone posted to the API via some other
    method than the content loader).
  - and that's weird, but we can't do anything about it, and weird stuff
    shouldn't be silent, so here we are.

771 of 777 relevant lines covered (99.23%)

0.99 hits per line

Source Files on job 252.1