Vector field with index_missing=True cannot be created #770

Description

A vector field declared with index_missing: True makes SearchIndex.create() fail. The same review thread on #742 noted this, and the from_existing parser skips INDEXMISSING on vector fields for this reason.

BaseVectorFieldAttributes.field_data adds "INDEXMISSING": True to the algorithm attribute dict. redis-py's VectorField counts and sends every dict entry inside the attribute block, so:

1. The client rejects the Python bool before sending: Invalid input of type: 'bool'.
2. Even as a string, the server rejects the token inside the block: Bad arguments for algorithm HNSW: INDEXMISSING.

FT.CREATE expects INDEXMISSING after the attribute block, which is where redis-py puts it when you pass index_missing=True to the field constructor.

Reproduction

Redis 8.4, redis-py 8.1.0, RedisVL main (e5d7598):

from redisvl.index import SearchIndex

schema = {
    "index": {"name": "im", "prefix": "im", "storage_type": "hash"},
    "fields": [
        {"name": "id", "type": "tag"},
        {"name": "emb", "type": "vector",
         "attrs": {"dims": 3, "algorithm": "hnsw", "index_missing": True}},
    ],
}
SearchIndex.from_dict(schema, redis_url="redis://localhost:6379").create()

redisvl.exceptions.RedisSearchError: Failed to create index 'im' on Redis:
Invalid input of type: 'bool'. Convert to a bytes, string, int or float first.

The same happens with flat and svs-vamana.

Expected behavior

The index is created with INDEXMISSING on the vector field, and ismissing(@emb) matches documents without the field.