medplum / medplum / 22793998241 / 1

92%

main: 92%

push

github

FooMedical Scheduling Demo (#8599)

* Mark `<Scheduler>` props as optional

There are already guards in place to handle usage of this component when
these props are not passed. Let's make the type information reflect
reality by marking these props as optional.

Signed-off-by: Noah Silas <noah@medplum.com>

* Move Schedule.actor fetching into separate effect

This doesn't depend on the current `month` state, so putting this into a
separate effect that runs less frequently is better illustrative of how
things should work.

In practice this is pretty minimal in effect, as the schedule should get
cached by the Medplum client and so re-running the query will be fast on
subsequent renders.

Signed-off-by: Noah Silas <noah@medplum.com>

* Remove a type assertion

Due to its type signature, `isReference()` can't actually be used to
refine a union type of `Resource | Reference<Resource>`. On the other
hand, `isResource()` does work the way we want!

So, by switching our condition here from `isReference(props.schedule)`
to `isResource(props.schedule)`, both branches end up with improved
typing, and we can drop an `as Schedule` typecast.

Signed-off-by: Noah Silas <noah@medplum.com>

* Improve Scheduler practitioner extraction

We had been assuming that a Schedule we get has a single actor of type
Practitioner, and using a typecast to bypass checks. Let's replace that
assumption with something we validate at runtime and in the type
checker.

Signed-off-by: Noah Silas <noah@medplum.com>

* Add new `fetchSlots` prop to Scheduler component

Previously this was being passed in via the `schedule` prop. That has
some complications:
- Requires a lot of  type checking
- Can't provide custom slot fetching (such as $find) while also
  contextualizing with a single practitioner's schedule

Splitting this out into a separate prop will facilitate future
scheduling work.

This is technically a breaking change (passing a function is no longer
accepted in the `schedule` prop), ... (continued)

17821 of 20208 branches covered (88.19%)

Branch coverage included in aggregate %.

32351 of 34148 relevant lines covered (94.74%)

15132.86 hits per line