allow customers to append custom distributions to the dcat:distribution array
#18
Conversation
sonofflynn89
force-pushed
the
f/1188-append-distributions-in-config
branch
from
70199f3 to
d1481de
Compare
| @@ -203,91 +195,154 @@ describe('formatDcatDataset', () => { | |||
| 'dct:title': 'OGC WMS', | |||
| }, | |||
| }; | |||
| const result = JSON.parse(formatDcatDataset(distDataset, defaultFormatTemplate)); | |||
There was a problem hiding this comment.
This looks like a big diff, but if you open in VSC mode (hit .), you'll see that I moved all of this into a nested distributions describe and made the following changes:
- Create a
distributionsconst that all tests can reference, reducing redundancy - Create a
getDistributionstesting helper that returns the underlying parseddcat:distributionarray for each test (greatly simplifies the expect statements for each test) - Verify order of all distributions explicitly (since we were implicitly checking before)
- Add tests around appending custom distributions
There was a problem hiding this comment.
Nice, well thought out
| @@ -46,9 +46,10 @@ export function formatDcatDataset(dcatDataset: any, template: DatasetFormatTempl | |||
|
|
|||
| const formattedDataset = adlib(template, dcatDataset, transforms); | |||
|
|
|||
| setLeftoverInterpolations(formattedDataset); | |||
| // reset uninterpolated, non-editable fields since customers cannot remove them | |||
| resetUninterpolatedPaths(formattedDataset, nonEditableFieldPaths); | |||
There was a problem hiding this comment.
When I revisited setLeftoverInterpolations for the first time, I could not for the life of me remember why we had this function. I eventually figured it out, but if I (who approved the original PR) had a hard time remembering the purpose of the function, I guarantee you that others will as well.
Changing the name to something like resetUninterpolatedPathsOfUneditableFields seemed way too verbose, so I opted to create a generic function name, explicitly pass in nonEditableFieldPaths and add in a clarifying comment. I'm hoping this semantic change will be a little more friendly for future devs (including ourselves).
| @@ -46,9 +46,10 @@ export function formatDcatDataset(dcatDataset: any, template: DatasetFormatTempl | |||
|
|
|||
| const formattedDataset = adlib(template, dcatDataset, transforms); | |||
|
|
|||
| setLeftoverInterpolations(formattedDataset); | |||
| // reset uninterpolated, non-editable fields since customers cannot remove them | |||
| resetUninterpolatedPaths(formattedDataset, nonEditableFieldPaths); | |||
| @@ -203,91 +195,154 @@ describe('formatDcatDataset', () => { | |||
| 'dct:title': 'OGC WMS', | |||
| }, | |||
| }; | |||
| const result = JSON.parse(formatDcatDataset(distDataset, defaultFormatTemplate)); | |||
There was a problem hiding this comment.
Nice, well thought out
Part of https://devtopia.esri.com/dc/hub/issues/1188.
To allow customers to add custom distributions to their feed via their dcat config, I made a few changes:
dcat:distributionfrom the list of non-editable fieldsdcat:distributionthrough adlibdcat:distributionarray IFF the adlib result is an arrayWe are not adding any validation as part of this PR. If the dcat config's
dcat:distributionis an array, we just throw it in there.