A better case mapping

Instead of having different `CASE_ID_TO_NAME` and `CASE_NAME_TO_ID` mappings, just have a `CASE_MAPPING` that looks like this:

```js
CASE_MAPPING = [
{
id: IO....,
title: ...,
subject_ids: [],
keyword_ids: []
},
{
id: IO....,
title: ...,
subject_ids: [],
keyword_ids: []
}
]

KEYWORDS = {keyword_id: keyword_title}
SUBJECTS = {subject_id: subject_title}
```

Locate the existing CASE_ID_TO_NAME and CASE_NAME_TO_ID mappings, then inspect the code that reads or updates them. Replace the duplicated mapping structure with the proposed CASE_MAPPING, KEYWORDS, and SUBJECTS structures, and verify that all existing case, keyword, and subject lookups still work.

Written by the indexing model from the issue text.