fix(schema): emit INDEXMISSING after the vector attribute block

[RaphaelFakhri]

What

Vector fields now pass index_missing to the redis-py VectorField constructor instead of adding INDEXMISSING to the algorithm attribute dict.

Why

With index_missing: True on a vector field, SearchIndex.create() fails. field_data put "INDEXMISSING": True inside the attribute block, so the client raises Invalid input of type: 'bool', and the server rejects the token there anyway (Bad arguments for algorithm HNSW: INDEXMISSING). FT.CREATE expects INDEXMISSING after the block, which is where redis-py places it for the index_missing keyword. The review on #742 described this failure.

Fixes #770.

Changes

• redisvl/schema/fields.py: drop INDEXMISSING from BaseVectorFieldAttributes.field_data, and pass index_missing=self.attrs.index_missing in FlatVectorField, HNSWVectorField and SVSVectorField.as_redis_field().
• tests/unit/test_fields.py: replace the test that asserted INDEXMISSING in field_data with a parametrized test (FLAT, HNSW, SVS-VAMANA) that checks INDEXMISSING follows the attribute block in redis_args().
• tests/integration/test_field_modifier_ordering_integration.py: create FLAT and HNSW indexes with a vector index_missing field and check that ismissing(@embedding) returns only the document without the vector.

from_existing still does not recover index_missing on vector fields (see the comment in tests/unit/test_convert_index_info.py). That can follow separately now that the field can be recreated.

Testing

• On main, the 5 new test cases fail (DataError: Invalid input of type: 'bool' for the integration cases). With this change, tests/unit/test_fields.py and tests/integration/test_field_modifier_ordering_integration.py give 72 passed against redis:8.4.
• tests/unit (without test_mcp) gives the same result as on main, apart from the 2 added cases. The only failures are the existing test_repr and google_genai optional-dependency ones.
• black, isort and mypy redisvl pass.

---

Note

Medium Risk
Touches vector index schema serialization for all algorithms; wrong ordering would break index creation, but behavior is narrowly scoped to the index_missing flag with added regression tests.

Overview
Vector index_missing now emits INDEXMISSING in the correct place for FT.CREATE, so indexes with missing-vector search can be created again.

Previously INDEXMISSING was stuffed into the vector algorithm attribute dict (field_data), which broke redis-py (Invalid input of type: 'bool') and Redis (Bad arguments for algorithm …: INDEXMISSING). The change removes it from BaseVectorFieldAttributes.field_data and passes index_missing into the redis-py VectorField constructor for FLAT, HNSW, and SVS-VAMANA, matching how lexical fields handle the modifier (token after the attribute block).

Tests were updated: a parametrized unit test asserts INDEXMISSING is not inside the counted attribute block but appears in redis_args() suffix; integration tests cover FLAT/HNSW indexes and ismissing(@embedding) queries.

^{Reviewed by Cursor Bugbot for commit 723e8b1. Bugbot is set up for automated code reviews on this repo. Configure here.}

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[vishal-bala]

Hi, thanks for your contribution! Can you rebase onto the latest origin/main and follow the new contribution process?