Null check field map before using it

[labkey-adam]

Rationale

After merging the related PR, OConnorExperimentTest started failing when attempting to render a multi-valued field while rendering a bulk edit page. This seems to fix it.

https://teamcity.labkey.org/buildConfiguration/bt401/3514983

I can't explain with certainty what changed here, but here's my guess, based on what I've seen:

• Previously (25.3) the RenderContext passed to MutiValuedRenderContext failed to resolve ParentExperiments/Container and ParentExperiments/ExperimentNumber (in MultiValueRenderContext constructor)
• MultiValuedDisplayColumn.values() then saw only null values for these field keys, so next() returned false, bypassing calls to the passed in Function
• Early in 25.4, a display column refactor caused us to bypass the MV getInputValue() code path entirely
• Later, Matt made changes that populated more values into the context (??)
• Nick & I just fixed the bypass with Fix multi-valued input form fields #6674, so MV getInputValue() code path is live again. It now resolves those field keys and invokes MultiValuedRenderContext.get(), which unconditionally uses the field map, causing an NPE in this case.

Related Pull Requests

• Fix multi-valued input form fields #6674

[labkey-adam]

OConnorExperimentTest passed here: https://teamcity.labkey.org/buildConfiguration/bt401/3515707?buildTab=overview