Commit graph

18 commits

Author SHA1 Message Date
trvon
982e9d2688
perf(search): favor lexical mixed defaults
Signed-off-by: trvon <git@trevon.dev>
2026-08-23 19:36:29 -06:00
Trevon
3f37e8f564
Sync experimental into main: Search retrieval improvements (#60)
* fix(graph): handle qualified C++ symbols

Signed-off-by: trvon <me@trevon.dev>

* refactor(vector)!: remove dead transition APIs

Split Simeon vector and retrieval linkage so consumers inherit only the runtime surface they use.

BREAKING CHANGE: remove unused async/model/config helpers and the undefined vector-backend factory from public headers.

Signed-off-by: trvon <me@trevon.dev>

* fix(vector): harden PQ snapshot lifecycle

Validate and atomically install persisted generations, revalidate after races, and fall back to exact search when a snapshot is stale. Add the exact-scan reference lane and distinguish observed work counters from unavailable backend metrics.

Signed-off-by: trvon <me@trevon.dev>

* feat(topology): expose route coordinates

Persist chart distortion and local-dimension evidence, retain pre-scalarization route axes, and emit detailed trace rows without treating unavailable vector work as zero.

Signed-off-by: trvon <me@trevon.dev>

* feat(xplan): add vector decision gates

Track binary and corpus identity, preserve missing-work semantics, and add repeated PQ routing, crossover, and rerank plans with quality and cost summaries.

Signed-off-by: trvon <me@trevon.dev>

* feat(topology): formalize coordinate routing

Separate chart geometry from route-risk admission and prove candidate-stage recall and strict-focus boundaries. Record the measured atlas limitations and literature basis for the next construction experiment.

Signed-off-by: trvon <me@trevon.dev>

* fix(search): serialize SymSpell readers

SQLite-backed lookups reuse mutable prepared statements, so shared-reader locking is unsafe. Advance the owned SymSpell revision for batched delete insertion and cover concurrent root lookups under TSAN.

Signed-off-by: trvon <me@trevon.dev>

* fix(vector): make mutations atomic

Preserve stored vectors on replacement and commit failures, reject cross-thread writes during explicit transactions, and restore exact reranking for quantized-primary PQ rows.

* fix(search)!: surface SymSpell write errors

Propagate persistent dictionary failures through metadata ingestion, roll back failed batches, and make clear remove stored terms.

BREAKING CHANGE: SymSpellSearch mutations now return Result values.

* feat(topology): formalize candidate expansion

* feat(vector): add coordinate-aware embeddings

* feat(topology): preserve protected relation cover

* feat(search): add graph candidate rescue

* feat(daemon): expose retrieval hydration telemetry

* feat(xplan): add rescue and hydration gates

* refactor(content): dedupe media handlers

Reuse shared graph presenter helpers and centralize repeated audio/video/archive buffer processing, validation, lifecycle, and stats helper logic.

* feat(lifecycle): prune stale corpus data

Keep cleanup preview-first while retention stays opt-in and long prune work no longer holds IPC responses open.

* feat(search): add topology candidate rescue

Preserve high-value routed candidates across vector backends while keeping exact rerank diagnostics explicit.

* feat(search): expose evidence paths

* docs: define mobile host contract

* formal: model bounded candidate rescue

* fix(prune): inspect extensionless signatures

* fix(prune): tolerate stale extensions

---------

Signed-off-by: trvon <me@trevon.dev>
2026-07-24 19:00:22 -04:00
Trevon
7df64b1161
perf(search): bound topology routing and harden runtime paths (#59)
* fix: address review build issues

* perf(search): bound topology routing

Cache route membership, shortlist centroids with static ANN, and route SPQ through explicit allowed-document budgets.

* test(xplan): measure routed vector work

Pin benchmark binary identity and export route, cache, and phase costs for repeated equal-budget comparisons.

* test(search): use portable jthread wrapper

* fix(daemon): serialize detached cancellation

* fix(cli): honor pinned daemon sockets

* fix(xplan): support Windows contract imports

* fix(list): apply metadata filters before paging

* docs(prompts): require memory hydration

Treat search results as candidates and hydrate selected artifacts before use. Quote argument hints for strict YAML loaders.

* test(cli): keep add on pinned daemon

* test(metadata): isolate performance timing

* fix(graph): tolerate stale snippet ranges

Source files may change before persisted symbol metadata is refreshed. Omit stale snippets instead of aborting graph exploration.

* test(daemon): await request readiness

Reuse the retrying list probe before asserting post-shutdown error quality under sanitizer load.

* ci(release): use hosted cmake

Resolve Conan CMake tool requirements from PATH so cold ARM64 caches do not depend on cmake.org. Default BUILD_DIR for failure cleanup.
2026-07-19 14:53:10 -04:00
Trevon
a3d065dd8d
feat(search): make topology-assisted retrieval the default (#57)
* fix(release): repair package repo publish

Arch repo metadata and R2 uploads need to be reproducible from CI, with retry hardening for transient Cloudflare failures.

* fix(daemon): stop shutdown log storm

Init waiter must treat ShuttingDown as terminal so daemon shutdown tests do not flood Meson logs while the daemon is stopping.

* refactor(cli): delegate serve to MCP binary

Keep yams serve UX while removing MCP server linkage from yams-cli. The standalone MCP binary owns stdio serving and exits on stdin EOF.

* refactor(cli): move daemon path helpers

Keep CLI path resolution out of YamsDaemon so later link cuts can separate CLI and daemon internals. Full yams_daemon removal still needs a socket-only daemon-client split.

* refactor(daemon): split embedded client

* perf(search): add lean-guided benchmark gates

Add Lean topology/search models, ingestion ablation docs, and focused regression tests for topology and compression threshold behavior.

Extract search observability, keep benchmark-only ablation controls, and make retrieval quality benchmarks default to Simeon while preserving explicit ONNX opt-in.

* refactor(config): move daemon policies

* refactor(daemon): extract plugin runtime

* refactor(daemon): untangle service coordinators

Untangle CLI/MCP from daemon service coordinators and the daemon static lib.

- MetadataRepository uses a daemon-neutral tree-diff callback. The
  GraphComponent wrapper moves to src/daemon/components/metadata_graph_bridge.cpp,
  owned by the daemon library.
- WriteCoordinator moves to a dedicated build-only yams_write_coordinator
  static lib. Drop ResourceGovernor coupling inside the writer loop; batch
  sizing still respects config.maxBatchSize.
- AppContext gains enqueuePostIngest callback. ServiceManager wires it to
  its own enqueuePostIngest; app services no longer call ServiceManager
  directly.
- Embedding repair uses BulkIngestLease/BeginBulkIngestCallback instead of
  concrete VectorIndexCoordinator::BulkScope. Daemon dispatchers provide a
  lease factory backed by the real coordinator.
- CLI daemon_command uses config helpers instead of YamsDaemon::getXDGRuntimeDir.
- Remove explicit dependency('yams_daemon') from src/cli and tools/yams-mcp.
- Declare missing yams_extraction/yams_ingest deps on yams_app_services so
  app services link cleanly without daemon.

Validated:
  meson compile -C build/release -j4 yams-daemon yams-cli yams-mcp-server
  ninja -C build/release -t commands tools/yams-cli/yams-cli tools/yams-mcp/yams-mcp-server | grep -c libyams_daemon.a -> 0
  yams-cli --version, yams-mcp-server init, yams serve --help smoke tests

Size drop on macOS release: yams-cli 41M -> 32M, yams-mcp-server 37M -> 22M.

* fix(cli): keep stdin add local fallback

No embedded daemon now. When yams add - buffers stdin and daemon IPC is missing, store via local app services instead of failing socket_missing.

Validated:
  meson compile -C build/debug -j4 yams-cli catch2_cli_commands
  MESON_BUILD_ROOT=/Users/trevon/work/tools/yams/build/debug build/debug/tests/catch2_cli_commands 'AddCommand - reads from piped stdin and stores content'
  MESON_BUILD_ROOT=/Users/trevon/work/tools/yams/build/debug build/debug/tests/catch2_cli_commands
  meson compile -C build/release -j4 yams-cli yams-mcp-server
  meson compile -C build/asan-api -j2 yams-cli catch2_cli_commands
  MESON_BUILD_ROOT=/Users/trevon/work/tools/yams/build/asan-api ASAN_OPTIONS=halt_on_error=1:abort_on_error=1:print_summary=1 build/asan-api/tests/catch2_cli_commands 'AddCommand - reads from piped stdin and stores content'
  ninja -C build/release -t commands tools/yams-cli/yams-cli tools/yams-mcp/yams-mcp-server | grep libyams_daemon.a -> no output

* feat(bench): add xplan multi-arm experiment harness

Unified plan runner, ablation axes, REPORT/ablation docs, and
KPI snapshot so optimization loops stop living in shell matrices.

* feat(formal): expand topology/search Lean contracts

Boundary models for search provenance, construction gates,
and cluster/route objectives so Lean guides correctness while
xplan ranks performance levers.

* feat(search,topology): construction purity + graph expansion path

Raise topology build defaults (min edge 0.25, max component 64) and
split oversized CC pieces so hybrid assist is quality-positive on
scifact. Add GraphNeighbors expansion with medoid-anchor fallback,
structured helpers instead of nested session logic, and BEIR-default
xplan plans. Search engine extraction (fanout/classic pipeline) rides
along as the integration surface for routing.

* feat(topology): add KMeans engine and SGC identity contracts

Register kmeans in the factory with deterministic spherical k-means
over embeddings, and enforce SGC document-identity invariants in
debug builds with matching unit coverage.

* feat(daemon): topology readiness gates and selectable reranker

Drive topologyReady from published epoch/freshness/rebuild/dirty
ingest state (not just KG drain), surface those fields in search
stats, and wire onnx|simeon|auto|none cross-reranker selection.

* chore(search): drop meta_path tuner knobs

Meta-path routing is gone; remove dead TunedParams fields and the
ConfigResolver test for [search.meta_path].

* feat(bench): add topology_purity_validate plan

Repeats=3 validation matrix for construction-purity winners vs
topo_off so MRR deltas can clear run-to-run noise.

* chore: format add_command and config include order

Mechanical clang-format / include reorder only; no behavior change.

* feat(bench): decision-grade xplan loop for search engine

Document search-engine measurement in AGENTS; harden xplan with
repeats=3 plans, quality×latency primaries, mechanism metrics from
debug.jsonl, multi-rep log promotion, and archive superseded topology
plans.

* feat(graph): denser product semantic_neighbor topK

Raise default YAMS_GRAPH_SEMANTIC_TOPK fallback from 4 to 8 so product
graphs leave fewer degree-0 docs for GraphNeighbors expansion.

* feat(search): extract topology assist and cross-rerank seams

Pull topology assist into search_topology_stage (seed merge, skip
reasons, optional vector seed probe default 0). Extract cross_rerank
compute from search_engine. Wire probe via typed config/ConfigResolver.

* feat(bench): product-path search engine ablation plans

Add topology-off component, vector-weight, and tiered-execution xplan
plans (repeats=3) and wire them into the AGENTS measurement loop table.

* fix(ci): repair pre-push fallbacks

* feat(bench): lock-free mid-run builds + multi-corpus gates

xplan skips meson rebuild when the quality binary exists (flock +
YAMS_BENCH_FORCE_BUILD override) to stop multi-arm link races. Add
graph_vector weight and nfcorpus transfer-gate plans; document in AGENTS.

* fix(daemon): keep SearchEngine alive via shared_ptr cache

getCachedEngine previously published a raw pointer that metrics polling
used through rebuild, racing destructor of SearchEngine::Impl (TSAN).
Store and return a shared_ptr snapshot so diagnostic readers pin the
engine for the duration of getConfig/health checks.

Also adds the cautious vectorWeight 0.30→0.20 multi-corpus xplan plan.

* feat(search): route vectors through cached topology

Build and reuse validated topology snapshots, rank routed members by query similarity, and narrow the vector leg to the accepted document set with explicit work diagnostics.

Remove the retired HDBSCAN engine and vendor dependency, preserve typed topology policy through search tuning, and cover the runtime/configuration contracts.

* feat(bench): consolidate search decision plans

Enforce deterministic experiment identity and declared baselines, expose routed/global vector work metrics, and register the focused benchmark contracts.

Promote the compact repeats=3 routing-budget gate, archive superseded topology mechanism screens, and add multi-corpus product calibration plans.

* feat(search): refine topology vector routing

* fix(search): propagate daemon readiness

* feat(bench): measure generalized memory health

* fix(cli): honor in-process fallbacks

* ci: harden smolvm debug builds

* perf(search): route topology through filtered ANN

* feat(bench): measure read-write pressure

* perf(metadata): index symbol lookups on read pool

* fix(daemon): isolate startup reads and maintenance

* feat(bench): replay isolated generalized corpora

* feat(search): add evidence routing pipeline

Replace the blocking smolvm pre-push path with host ASan/TSan validation, harden sanitizer-sensitive smoke tests, and add recovery cleanup for malformed SQLite WAL sidecars.

* refactor(search): unify topology evidence path

Remove classic and sidecar fusion lanes so topology augments the recall-preserving ANN and lexical candidate pool through one product path.

* refactor(bench): prune superseded search lanes

Measure the topology-plus-ANN product path with stable experiment identity and remove obsolete sidecar and tiered controls.

* docs(search): retire stale experiment notes

Keep the benchmark handoff focused on the topology-plus-ANN product lane and its active decision plans.

* refactor(search): prune dead search surfaces

Remove unused fusion, reranker, policy, and router abstractions.
Bound topology materialization and keep document trace payloads opt-in.

* fix(cli): clear build warnings

* fix(xplan): reject stale run identity

* refactor(search)!: remove legacy query stack

Replace the unused AST/ranker/post-processor surface with focused facet aggregation used by product search.

BREAKING CHANGE: removes query parser, result ranker, search filter, and SearchResults container APIs from yams_search.

* refactor(search): type value stage traces

Remove synthetic ComponentResult markers for embedding and concept stages. Preserve failure and timeout states instead of overwriting them with fake successful results.

* refactor(search): isolate benchmark env overlays

Move test-only search environment parsing behind an injectable internal adapter. Keep all existing overlay keys and tuner pin semantics while restoring SearchEngineBuilder to composition.

* refactor(search): unify topology routing options

Snapshot topology policy once at the assist boundary and pass it through routing sessions and diagnostics. Remove repeated scalar copying without changing routing behavior or metric keys.

* fix(search): unify tuner config ownership

Use one complete config-to-params overlay for runtime seeding and no-tuner policy resolution. Reseed attached tuners on setConfig so reported and executed configuration cannot diverge.

* fix(search): snapshot tuner state per query

Read tuned config, params, and state under one lock before routing. Add a TSAN correlation test so adaptive observation cannot produce a torn logical query snapshot.

* fix(search): enforce topology routing contract

Generated configs must preserve hybrid-assist shadow routing, while daemon-backed search must abstain when published topology artifacts are stale. Keep standalone validation and global ANN recall intact.

* fix(daemon): let SQLite reclaim its file

Directly resizing a live database from sampled page counts can truncate pages allocated by concurrent writers. VACUUM must remain the only file-reclamation mechanism.

* fix(metadata): tolerate missing symbol FTS

Readiness lookups can race an unavailable optional FTS migration. Retain the indexed fast path, but retry against the base table when prepare or cached-statement execution proves the capability absent.

* build(fuzz): drop removed query parser target

The optional AFL build still referenced the deleted legacy query parser API and could no longer compile.

* fix(bench): include writes in drain oracle

A zero post-ingest queue is not durable drain while coordinated writes remain queued or in flight.

* fix(daemon): vacuum padded WAL databases safely
2026-07-13 20:26:43 -04:00
Trevon
bfe7cd94ca
test: expand storage-readiness coverage — 48.1%→59.8% broad, 74.3% narrow (#38)
* fix(daemon): avoid socket cancel on stop

Signal active sockets with shutdown only so pending I/O wakes without entering Boost.Asio's reactor cancel path during teardown.

* fix(metadata): open read pools read-only

* ci: fix coverage report generation

* fix(metadata): tag unscoped pool holders

* test(daemon): avoid ipc teardown takeover

* ci: skip plugin coverage objects

* fix(daemon): repair completed FTS5 ghosts

* build(vector): guard unusable FAISS headers

* ci: use clang gcov for coverage

* test(daemon): stabilize arm CI flakes

* Harden search and compression numeric config

* fix(daemon): reconcile orphan vector embeddings

* fix: address vector cleanup review comments

* chore(main): release 0.16.1

* Harden search ranking divisors

* Harden snippet distance and command registration

* test(storage): add readiness coverage gate

Document the distributed sync readiness surface and add direct adapter, refcount, and GC failure-path coverage.

* test(storage): cover strict resolver fallback

* fix(search): tighten retrieval ranking

Honor component weights in default RRF, over-fetch vector chunks before document aggregation, and persist topology centroids for sparse-guided routing.

* fix(search): route weak queries through topology

* fix(search): improve rerank and weak queries

Wire provider-backed cross reranking, add direct Simeon candidates for weak lexical pools, and report stage-level retrieval metrics.

* fix(storage): honor explicit plugin names

Add readiness coverage for factory, URL backend, and compressed storage failure paths.

* fix(storage): avoid remote size listing

* fix(api): rollback failed content adds

* test(storage): cover engine readiness

* test(storage): cover compressed readiness

* fix(core): harden delete and recovery paths

Route prune and local delete fallback through DocumentService so content, metadata, vectors, and KG cleanup stay consistent.

Treat transient SQLite quick_check lock failures as ResourceBusy, refresh IPC activity on streaming writes, and preserve reranker config in tuner seeding.

* test(storage): cover URL backend readiness

* fix(api): harden content rollback

* feat(storage): add list seam, URL listing, and write-edge coverage

* feat(api): reconcile ambiguous remote manifest writes

* feat(storage): add deep compressed scan validation

* test(storage): add plugin loader error path coverage

* fix(gc): prune reference rows after collection

* test(api,storage): add health, compact, gc, and verify edge coverage

* docs: update storage-readiness coverage to 71.8%

* test(metadata): add connection pool shutdown-while-leased coverage

* test(metadata): add tree builder snapshot root stability coverage

* test(vector): add before-init and after-close operation rejection coverage

* test(metadata): add database post-close, open-failure, and cache-clear coverage

* test(integrity): add P2P wrong-hash rejection + batch matrix resolution

* docs: audit and resolve all matrix rows — 18 fully covered

* test(metadata): add failed-migration and rollback coverage

* test(storage): add getBlockSize-invalid-key, empty-verify, PUT/DELETE retry coverage

* feat(storage): add file-open and rename failure hooks + ContentStore retrieve edge coverage

* docs: add broad-surface coverage with per-subsystem breakdown

* test(integrity): add chunk validator manifest validation coverage

* test(vector): add batch insert, delete, and update operation coverage

* docs: update integrity subsystem coverage to 46.4%

* docs: consolidate coverage report — clean matrix, remove boilerplate

* docs: update coverage with complete-suite numbers — 59.8% broad, 74.3% narrow

* test(storage): add URLBackend HEAD exists with 2xx coverage

* bump submodules

* fix: address PR review — curl_slist leak, ec ordering, CAS rollback, production hooks

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2026-05-25 22:58:15 -04:00
trvon
7ae52e8b8b
fix: test fixes and search fusion update to correctly expose simeontext fusion so that it does not drop ft5s candidates. Still explore fusion optimizations 2026-05-04 15:54:57 -04:00
trvon
ae1f0c10d2
perf: cleaning up search experiments, moving to new routing system for search for larger corpus improvements, repair system now triggers semantic / graph repair 2026-04-29 20:36:02 -04:00
trvon
7de4d95a43
feat: repair logic improvements, expanding config options to disable rebuilding/repair for memory constrained systems 2026-04-24 10:53:15 -04:00
trvon
6174b13caf
feat: continued extraction from serach engine and vector improvements. This allows for better testing and experimentation 2026-04-22 12:24:28 -04:00
trvon
be0feac605
feat: refactor vector engine to reduce monoliths in codebase 2026-04-22 08:36:55 -04:00
trvon
a3f1d529c9
fix: addressing unoptimal copies of vectors in repair path 2026-04-22 02:09:58 -04:00
trvon
4617bf4d8f
feat: further removing hnsw references and centralizing on simeon pq implementation 2026-04-22 01:10:57 -04:00
trvon
bdcff5b636
fix: HNSW and topology rebuilds are now batched in repair. Cleaning up from previous migration work 2026-04-21 13:58:10 -04:00
trvon
2b60bed67b
feat(breaking): breaking change. removing onnxruntime for embeddings in favor of simeon. embeddings will now use less system resources, run faster but require avx/neon support per library 2026-04-21 12:23:38 -04:00
trvon
036e5e7d9e
feat: daemon lifecycle improvements for supervisor mode, search now uses topology logic, homebrew formula launches yams in foreground mode and test updates 2026-04-14 00:54:57 -04:00
trvon
12e4b1943b
feat: fixes for repair batching logic, hnsw rebuild and search refinements before graph building improvements land 2026-04-04 20:06:12 -04:00
trvon
81cc38a3cf
fix: expanding utf-8 sanitization and expand search tracing and test coverage 2026-03-21 10:12:42 -04:00
trvon
2360828eb7
feat: bug fixes in daemon lifecycle, cleaning up tests and resolving test failures 2026-02-17 11:04:22 -05:00