elastic / elastic/semantic-code-search-indexer
enhancement: make code_vector mapping conditional when ENABLE_DENSE_VECTORS is false
- Dominant language
- TypeScript
- Stars
- 19
- Forks
- 10
- PR merge metrics
- No merged PRs in 30d
Description
## Problem
The index mapping always defines `code_vector` as an indexed `dense_vector` field:
- `index: true`
Even when `ENABLE_DENSE_VECTORS` is disabled, this mapping can still increase index overhead (storage, indexing CPU) and potentially require extra cluster resources.
## Where in code
- `src/utils/elasticsearch.ts` in `createIndex()` mapping for `code_vector`.
## Suggested fix
Options:
1) Only include the `code_vector` field mapping when `ENABLE_DENSE_VECTORS=true`.
2) Keep the field but set `index: false` unless enabled.
## Notes
This is a mapping change; deployments will need a clean reindex to pick it up.
## Test plan
- Unit test for `createIndex` mapping generation (mock client) verifying `code_vector` is omitted or non-indexed when disabled.
Contributor guide
No contributing guide indexed for this repository
Research direction
Read src/utils/elasticsearch.ts, focusing on createIndex() and the code_vector mapping. Run or add the planned unit test with a mocked client, then verify the generated mapping does not index code_vector when ENABLE_DENSE_VECTORS is disabled while preserving the enabled case.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- elasticsearch, typescript
- Domain
- search
- Issue type
- Feature
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100