Open Issues Need Help
View All on GitHub Hybrid retrieval: RRF fusion of lexical and vector results about 2 months ago
enhancement help wanted
CONTRIBUTING.md is stale — still says "do not add vector/ANN code" about 2 months ago
documentation good first issue
Python client `indices.get_mapping` returns a hardcoded, DecoverAI-specific mapping about 2 months ago
bug good first issue
CLI: ship shell completions and a man page about 2 months ago
enhancement good first issue
enhancement good first issue
Add a `GET /metrics` Prometheus endpoint about 2 months ago
enhancement help wanted
Implement the CreateNamespace / DeleteNamespace HTTP routes about 2 months ago
enhancement good first issue
Tokenizer drops terms on interior punctuation and is not Unicode-aware about 2 months ago
enhancement help wanted
Honor `FieldSchema.analyzer`: configurable analyzers, stopwords, stemming about 2 months ago
enhancement help wanted
Add `histogram` and `date_histogram` aggregations about 2 months ago
enhancement help wanted
Add metric aggregations: min / max / avg / sum / value_count about 2 months ago
enhancement good first issue
Add a `prefix` filter clause about 2 months ago
enhancement good first issue
Add an `exists` filter clause about 2 months ago
enhancement good first issue
Aggregations ignore the query and filter — they aggregate the whole segment about 2 months ago
bug help wanted
vector: late-interaction (ColBERT-style) multi-vector retrieval about 2 months ago
enhancement help wanted
ops: minimal built-in dashboard for namespaces, stats, and a query console about 2 months ago
enhancement good first issue help wanted
cache: usage-driven and manifest-change-driven warming (warmup is static and one-shot today) about 2 months ago
enhancement help wanted
query: score modifiers — rank by attribute (recency/boost decay) and rank by distance (geo) about 2 months ago
enhancement help wanted
query: result diversification — cap hits per attribute value (limit.per / collapse) about 2 months ago
enhancement help wanted
vector: i8 scalar-quantized vectors (~4x storage and I/O reduction) about 2 months ago
enhancement help wanted
server: native embedding at ingest and query time about 2 months ago
enhancement help wanted
control-plane: namespace branching via copy-on-write manifests about 2 months ago
enhancement help wanted
control-plane: namespace sharding (fixed shard count first) about 2 months ago
enhancement help wanted
core: nested attributes (objects and arrays) in documents, filters, and aggregations about 2 months ago
enhancement help wanted
query: RRF fusion and weighted reranking across subqueries about 2 months ago
enhancement help wanted
aggregations: metric aggs (sum, min, max, avg, value_count) and metrics nested under buckets about 2 months ago
enhancement help wanted
query: fuzzy / typo-tolerant term matching about 2 months ago
enhancement help wanted
indexing: the schema `analyzer` field is accepted and silently ignored — tokenization is hardcoded about 2 months ago
enhancement help wanted
query: add a `prefix` filter clause about 2 months ago
enhancement good first issue
clients/python: `get`, `mget`, and `msearch` are missing from the OpenSearch-compatible client about 2 months ago
enhancement good first issue
observability: no /metrics endpoint — the engine computes the numbers but never exports them about 2 months ago
enhancement help wanted
aggregations: no metric aggregations (min / max / sum / avg) about 2 months ago
enhancement help wanted
query: highlighting only supports a single field per request about 2 months ago
enhancement help wanted
query: add an `exists` filter clause about 2 months ago
enhancement help wanted
cli: `kosha search` has no --from, --sort, or --search-after flags about 2 months ago
enhancement good first issue
cli: add `compact`, `import`, and `create-api-key` commands for the remaining admin endpoints about 2 months ago
enhancement good first issue
cli: add a `kosha exists` command for the /v1 exists endpoint about 2 months ago
enhancement good first issue
chore: add GitHub issue and pull request templates about 2 months ago
documentation good first issue
docs: CONTRIBUTING.md is stale and tells new contributors not to write vector code about 2 months ago
documentation good first issue
docs: add a configuration reference for the KOSHA_* environment variables about 2 months ago
documentation good first issue