Use filename instead of name for id in HarvestList

[dafeder]

We are using the name element from file_scan_directory to create source_ids that are stored in the migrate map tables for harvest. name attempts to remove the file extension (specifically, it uses the basename element from PHP's pathinfo(), so asdf1234.json in the cache would be simply stored as source ID "asdf1234." However, files are not cached with a file extension. In most cases this doesn't matter, but if there is a period (.) in ID, what comes after it can be mistaken for a file extension.

So, source ID knb-lter-jrn.210001001.61 gets cached the the filename "public://dkan-harvest-cache/[harvest-machine-name]/knb-lter-jrn.210001001.61", but when gathered by the getIdList() function will then be interpreted as simply the id knb-lter-jrn.210001001. Harvest will then think this is an orphaned dataset and unpublish it.

Since we don't cache either JSON or XML files with extensions, using filename should remove this problem with no ill effects.

QA Steps

1. Create a harvest from https://api.jsonbin.io/b/5eb59eb9a47fdd6af15fd015 - it should bring in 4 datasets.
2. Delete one of the datasets
3. Re-run the Harvest. The deleted dataset should come back

Tests

I did not add a new test for this, it felt like an edge case. Reviewer can lmk if they think a unit test is needed.