Add paradigm and analysis inference

[andrewdotn]

A first pass at adding analysis-matching/paradigm-assignment code when creating the ndjson.

Sample output fragment:

"pos": "NA-1",
"senses": [{ "source": "CW", "definition": "star, little star" }],
"analysis": [[], "acâhkos", ["+N", "+A", "+Sg"]],
"paradigm": "NA",
"linguistInfo": {
  "inflectional_category": "NA-1",
  "pos": "N",
  "wordclass": "NA"
}

Notes:

• This adds a single TypeScript file; run npx --no-install tsc --watch to get TypeScript to automatically transpile it to JavaScript when editing. I’m not sure where best to document that here.

• The code has largely been ported from https://github.com/andrewdotn/cree-intelligent-dictionary/commit/86a59a0026c6a36aff7907ea32c4c2a94f582c66. See also https://github.com/giellalt/lang-crk/blob/main/tools/shellscripts/add-explicit-fields-to-crkeng.sh.

• Many warnings are printed about cases where the dictionary entry is not analyzable, or has ambiguous analyses. These entries do not currently get analyses assigned, as that will require linguistic work.

• The pos field from CW means something different from the pos field in the current cree-intelligent-dictionary display code; I’ll leave it to the linguists to sort out what should be renamed to what, and how to prioritize a rename alongside other morphodict work.