* docs: add spider2-specs handoff directory for benchmark-driven feature specs
* feat(cli): connection-scoped wiki pages
Add an optional `connections` frontmatter field so database-specific wiki
knowledge can be scoped to a connection without polluting searches about other
databases, while page keys stay a flat, globally-unique namespace.
- connections: single string or list; absent/empty ⇒ unscoped (applies to all)
- wiki_search (MCP) and `ktx wiki --connection` return unscoped ∪ matching
pages, filtered at the disk-load seam so all three search lanes draw their
candidate pool from the already-scoped set (not a post-filter)
- wiki_write accepts connections with REPLACE semantics and rejects a
connection-scoped write whose key collides with a disjoint-connection page
(data-loss guard; hard error, no silent clobber)
- explicit connection-id args (wiki_search, memory_ingest, ktx wiki) are
validated against ktx.yaml via a shared assertConfiguredConnectionId, which
also closes the prior gap where memory_ingest's connectionId was unvalidated;
persisted ids absent from config warn (not fail) in `ktx status`
- prompt guidance in the wiki_capture skill and external-ingest prompt; the
session connectionId is surfaced to the memory agent and ingest work units
Implements spider2-specs/specs/01-connection-scoped-wiki.md; intake draft moved
to spider2-specs/done/.
* docs(spider2-specs): add specs/ refinement stage and composite-key join spec
Describe the todo/ → specs/ → done/ pipeline in the README (refined specs are
the durable artifact; intake drafts move to done/ on ship) and add a
MEDIUM-priority spec for multi-column composite-key join detection found during
the first sqlite smoke test.
* feat(cli): add --verbatim ingest mode for authoritative documents
Store each --text/--file document body unchanged as a GLOBAL wiki page
instead of routing it through the memory agent, which may rewrite,
condense, or re-title it. The LLM derives only metadata (summary, tags,
sl_refs) and only for frontmatter fields the document does not already
set; the stored body is written by code and never edited.
- Deterministic page key: files derive it from the filename, inline
text from its leading Markdown heading (headless inline text is
rejected — pass it as --file instead).
- Idempotent: re-running the same body is a no-op; a different body at
the same key fails loudly rather than overwriting.
- Works with llm.provider.backend: none, deriving a degraded summary
from the heading or first sentence.
- Existing frontmatter (including unmodeled fields like effective_date)
passes through untouched; --connection-id scopes the page.
* feat(cli): SQL-authoring craft and per-dialect notes tool for the analytics skill
Spec 07: add a dialect-agnostic <sql_craft> block to the ktx-analytics skill (schema discovery, composition, window-function correctness, numeric precision, answer completeness) with one worked window-then-filter example. Workflow steps gain pointers into it; existing guidance is unchanged.
Spec 08: add a read-only sql_dialect_notes MCP tool returning a connection's engine SQL conventions (FQTN form, identifier quoting/case, date/time, top-N idiom, JSON access), resolved through the existing sqlAnalysisDialectForDriver path. Notes are per-dialect markdown files under context/sql-analysis/dialects, served by the tool and copied to dist (package-internal, never installed). Non-SQL connections return a clear KtxExpectedError. The flat skill gains a one-line pointer to the tool.
Both spider2-specs intake drafts move to done/ with implementation notes.
* feat(cli): tolerate objects that fail introspection during scan
Isolate per-object introspection failures so one broken or inaccessible object no longer zeroes out a connection's whole semantic layer: the sqlite and bigquery connectors introspect each object defensively (tryIntrospectObject), the live-database adapter records a scan outcome and fetch report, and enabled_tables accepts catalog.db.name, db.name, or bare names with a clear no-match error. Includes matching ktx-daemon introspection changes, docs, and tests.
* docs(spider2-specs): add 06-scan-tolerate-broken-objects spec
* feat(cli): generalize analytics fan-out rule to multi-hop join chains
The ktx-analytics skill's fan-out rule only reliably caught single-hop
inflation; agents still silently fanned out on multi-hop chains where the
offending one-to-many join sits several hops below the SUM/COUNT and is easy
to miss.
Rewrite the Composition rule so the danger reads as cumulative across the whole
chain (pre-aggregate per measure-owning table), add an affirmative
grain-verification habit (default: pre-aggregate to grain; escape hatch:
COUNT(DISTINCT key) for pure counts only; SUM/AVG of a fanned-out measure must
pre-aggregate), and add one generic wrong-vs-right worked example. Content-only
and dialect-agnostic; no new tool, flag, or config.
Implements spider2-specs/specs/09 and annotates spec 07's one-example
constraint as superseded.
* feat(cli): add panel-completeness, time-series window, and text-encoded numeric SQL craft
Extend the analytics skill's <sql_craft> with three correctness habits and
route the dialect-specific halves through sql_dialect_notes:
- Panel completeness (spec 10): full-domain spine -> LEFT JOIN -> COALESCE for
"each/every/all/per" questions, defaulted by measure additivity.
- Time-series windows (spec 11): explicit cumulative frames, calendar-range
rolling windows with minimum-periods guards, and period-over-period via LAG.
- Text-encoded numerics (spec 12): sample distinct values, strip/scale/cast in
one early CTE, and confirm coverage with a failure-detecting cast.
Add per-dialect Series, Rolling window, and Safe cast notes to all seven
dialect files so the skill stays dialect-agnostic while the engine-specific
syntax lives in sql_dialect_notes. Tests updated and passing (19).
* docs(spider2-specs): add specs 10-12 for analytics SQL-craft additions
Refined specs and completion records for the panel-completeness spine (10),
time-series window recipes (11), and text-encoded numeric parsing (12)
implemented in the preceding commit.
* docs(spider2-specs): add backlog intake drafts 13-14
- 13: canonical authoritative-source measures
- 14: output-completeness final check
* skill(analytics): spec 14 output-completeness + iter1 (active column planning)
Bundles two changes (entangled in SKILL.md; future spider2 iterations land as
separate commits):
- spec 14 (output-completeness): multi-part "answer every requested output" rule
+ a "Final completeness check" in workflow Step 6 and <sql_craft>; analytics
skill-content test updated; intake draft -> done/, refined spec added.
- iter1 experiment: spec 14's passive end-check did not change behavior on the
benchmark's output-completeness failures, so (a) the Plan step now writes the
exact output-column list UP FRONT as a contract the final SELECT must match,
and (b) "expose identity" -> "project BOTH the entity id and its name" (covers
both omission directions). All generic craft.
Driven by the Spider 2.0-Lite failure analysis (incomplete output was the
largest failure bucket); benchmark only as motivation.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* skill(analytics): iter2 — deterministic order in string/array aggregation
GROUP_CONCAT/string_agg/array_agg element order is undefined without an explicit
ORDER BY; also note SQLite's default text sort is binary/case-sensitive (uppercase
before lowercase) vs case-insensitive (COLLATE NOCASE). Generic SQLite craft.
Spider 2.0-Lite motivation: an ordered-ingredient-list question failed only on the
within-string element order (right elements, wrong order); benchmark as motivation only.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(mcp): structured, leveled logging for the MCP server
Add one synchronous pino logger per MCP server process, written through the
io.stderr sink: plain JSON when stderr is not a TTY, colorized pino-pretty
(sync, in-process) when it is. Every tool call logs tool.start with its raw
params BEFORE the handler runs and tool.end after (info / warn past
KTX_MCP_SLOW_TOOL_MS / error), correlated by callId plus sessionId, so a
runaway sql_execution leaves a recoverable start line with its exact SQL and
no matching end. HTTP logs session.open/close and wires the previously-dead
transport.onerror to transport.error; stdio routes its transport error
through the logger. Level via KTX_MCP_LOG_LEVEL (default info). Existing
mcp_request_completed telemetry and registerParsedTool are unchanged; no
worker/async transport and no redaction in v1 (logs are local-only).
Implements spider2-specs/specs/15-mcp-server-structured-logging.md and moves
the intake draft to done/.
* feat(mcp): report uptimeMs in MCP server /health
The /health endpoint now includes uptimeMs (monotonic elapsed time since
the server started), mirroring the Python daemon's uptime_ms telemetry
field.
* feat(cli): bound read-query execution with a per-connection deadline
Enforce one shared query deadline (default 30s, overridable per connection via
query_timeout_ms) on every executeReadOnly path, so an accidentally-expensive
LLM-authored query returns a fast "query exceeded Ns" KtxQueryError instead of
hanging the MCP server.
- New shared contract context/connections/query-deadline.ts
(resolveQueryDeadlineMs, queryDeadlineExceededError); query_timeout_ms added to
the shared warehouse schema; BigQuery's job_timeout_ms removed.
- SQLite runs the read query in a short-lived forked child process and enforces
the deadline with SIGKILL. worker_threads + terminate() was tried first but
cannot interrupt a synchronous better-sqlite3 scan (the native loop never
yields); SIGKILL reclaims the process in ~2ms and keeps the event loop free.
- Remote connectors apply a real server-side statement timeout and re-wrap their
own timeout signal as KtxQueryError: Postgres statement_timeout/57014, MySQL
max_execution_time/3024, Snowflake STATEMENT_TIMEOUT_IN_SECONDS/604, ClickHouse
max_execution_time + aligned request_timeout/159, SQL Server requestTimeout/
ETIMEOUT, BigQuery jobTimeoutMs.
- Relationship validation skips a candidate to review on a deadline timeout
instead of aborting the pass; the deadline surfaces through the existing MCP
pino logger as a matched tool.start/tool.end(error) pair (no new logging code).
Also fixes a pre-existing, unrelated invalid cast in mcp-server-factory.test.ts
that was breaking tsc -p tsconfig.test.json.
* docs(spider2-specs): mark spec 16 (bounded query execution) done
Append Implementation notes to the refined spec (what shipped, where, and the
worker-thread -> child-process+SIGKILL deviation with its evidence) and move the
intake draft from todo/ to done/.
* skill(analytics): iter3 — measure-as-amount, inter-event gap, top-per-metric career
Three generic interpretation rules: a named business measure (sales/revenue/spend)
means its amount not a row count; "inter-event duration/gap" is LAG/LEAD time-between
events not a magnitude column; "highest across several achievements" aggregates per
metric over the whole history. All three demonstrably FIRE (verified on local008/003/152
SQL). local008 flips to correct (mechanism-aligned). 003/152 still fail on a different
axis (source-column / grouping). Generic craft; benchmark only as motivation.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* skill(analytics): spine-for-extreme-selection + aggregate-over-selected-set
Two generic answer-completeness refinements:
- Selecting the extreme group (lowest/highest count over a period/category
domain) must rank over the COMPLETE spine, not only groups with fact rows —
an empty period is a genuine 0 and often the true minimum.
- An aggregate scoped to a per-entity selected set ('avg revenue per actor in
those top-3 films') is computed ACROSS that set, distinct from the per-item
value; project both.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* skill(analytics): iter2 — sharpen extreme-selection spine + top-N ranking-measure
- spine-for-extreme: concrete cue that a zero-row period never appears in a
GROUP BY of the facts; generate the full calendar, LEFT JOIN, COALESCE, then rank.
- aggregate-over-selected-set: top-N selection ranks by the named ranking measure
(the item's own revenue), independent of the per-item share that feeds the aggregate.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* skill(analytics): iter3 — comparison-between-two-extremes is one wide row
Distinguishes a cross-item comparison ('the difference between the highest and
lowest month' -> single wide row, both extremes side by side + the comparison
column) from 'report a metric for each group' (-> stays long). Generic, question-
derived; targets the wide-vs-long shape gap without affecting per-group long output.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* skill(analytics): iter4 — anchor a period bucket to the named lifecycle event
When a record carries multiple lifecycle timestamps (created/placed, approved,
shipped, delivered, completed, settled) and the question counts/measures records
in a named *completed state* by period ("delivered orders by month", "shipped
items per week"), bucket the period by that named event's own timestamp, not the
record-creation timestamp; the state value is the qualifying filter, the matching
timestamp is the time anchor. Wording priority is explicit — purchased/placed/
created/submitted/ordered keep the start-event timestamp — and a non-temporal
state filter (counts by customer/city/seller with no period) introduces no anchor.
Generic analytics craft: counting completed-state records by their creation date
silently answers "records that later reached that state, grouped by when they
started" instead of the question asked. Surfaced via the spider2-autofix loop;
FAIR_PRODUCT (adversary-screened, restatable from question wording + schema/
semantic-layer lifecycle descriptions, no gold dependency).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* skill(analytics): iter5 — canonicalize observed URL-path variants before page-level analysis
When a question groups/filters/sequences web pages by a path/url column, sample
its distinct values; if the data itself shows /route and /route/ variants for the
same page context, canonicalize in an early CTE (preserve / as root, strip trailing
slashes from non-root paths, map an observed empty path to / only when the column is
a URL path with blank root-page events) and use the canonical path everywhere above.
Explicitly forbids inventing aliases the data doesn't show: no merging different
route names, no stripping query/fragment/host/scheme, no lowercasing, and no
canonicalization when the question asks for raw URL/path or slash-vs-no-slash diffs.
Generic web-analytics craft: raw request logs routinely store the same user-visible
page with and without a trailing slash, so grouping raw labels silently splits one
page into several. Surfaced via the spider2-autofix loop (Codex runner, round r2);
FAIR_PRODUCT (adversary-screened, restatable from URL-path semantics + page-grain
question wording + solver-observed distinct values, no gold dependency). The rule
fired mechanism-aligned on both targets; flipped local330 (landing/exit page counts),
local331 residual is a separate sequence-semantics axis beyond canonicalization.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* skill(analytics): iter6 — coverage over a selected group is a set-membership aggregate
When a question first selects a group of entities ("the top 5 actors", "these
products") and then asks what count/share/percentage of a DIFFERENT subject domain
relates to *these* selected entities ("what % of customers rented films featuring
these actors"), the subject set is the UNION across the whole group: count DISTINCT
subject ids once across the selected entities and return one collective value at the
subject-domain grain — not one row per selected entity (which double-counts subjects
related to more than one entity and answers a different question). Narrowly guarded:
emit one row per entity only when the wording says "for each / per / by / list" or
asks for each entity's own metric ("top 5 players and their batting averages").
The collective-coverage cousin of the existing per-entity selected-set rule. Generic
analytics craft (per-entity metric vs set-level coverage). Surfaced via the
spider2-autofix loop (Codex runner, round r3); FAIR_PRODUCT (adversary-screened,
restatable from wording alone, no gold dependency). Flipped local195 mechanism-aligned
(union COUNT(DISTINCT customer)/total, one scalar); 0 regression across 5 passing
per-entity top-N guards (local023/024/029/212/221 stayed long).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* skill(analytics): label-only joins must LEFT JOIN — incomplete dims silently drop fact rows
Mirror of the existing fan-out rule for the DROP direction: an inner JOIN to a
dimension table used only to attach a display attribute silently discards every
fact row whose key has no parent when the dimension is incomplete (trimmed
catalogs, late-arriving / SCD-gap rows), shrinking counts/sums and the universe
over which shares/averages/medians are computed. Guidance: LEFT JOIN pure
enrichment; inner-join a dimension only when intended as a filter; key the
aggregate/GROUP BY on the fact column, not the dimension column.
Spider2 autofix round 'joindim': flips complex_oracle local050 (FAIL->PASS,
official scorer) — solver dropped the gratuitous products inner-join and
recovered the exact gold. local060/063 also adopt LEFT JOIN (rule fires) but
remain gold-convention-blocked. Guards local061/067 held.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(spider2-specs): add todo/17 — lifecycle-event metrics (semantic-layer)
Draft intake spec surfaced by the spider2-autofix loop (round r1): the model-layer
form of the shipped iter4 lifecycle-date-anchoring skill rule — infer per-state
lifecycle-event metrics (e.g. delivered_orders with defaultTimeDimension = the
delivery timestamp) during enrichment so the correct time anchor is the default for
any consumer, not only an agent that loaded the skill. Generic; FAIR_PRODUCT.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(connectors): accept leading underscore in connection/identifier ids
The safe-identifier validator regex /^[a-zA-Z0-9][a-zA-Z0-9_-]*$/ allowed an
underscore everywhere except the first character, so a connection id / database
name that legitimately starts with '_' (valid in Snowflake, e.g. _1000_GENOMES)
could never be ingested or queried. Allow a leading underscore across all 16
duplicated validators (connection ids, source ids, page/wiki keys, warehouse-
verification tool schemas). Path-safety is unaffected — '.' and '/' remain
excluded, and assertSafePathToken still blocks traversal.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(analytics): generic geospatial query guidance
Add a Snowflake ST_* dialect note (ST_MAKEPOINT lon-first, ST_DWITHIN/ST_CONTAINS/
ST_WITHIN/ST_INTERSECTS, bbox->polygon via ST_MAKEPOLYGON/ST_MAKELINE) and a
dialect-agnostic 'Spatial predicates' recipe in the analytics skill (resolve the
entity geometry, build an area-of-interest polygon, test with the engine's
containment/proximity/overlap predicate; mind lon/lat argument order). Steers the
solver off hand-rolled lat/lon BETWEEN boxes toward correct, index-assisted
geospatial predicates.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(analytics): parse code/dependency text by language grammar
Add two generic <sql_craft> rules: (1) parse imported/required/loaded packages by
the language or manifest format (Java import keep-package-path allowing underscores/
mixed-case; Python import/from + alias stripping; R library/require; .ipynb parse
JSON cell source before language rules; JSON manifests flatten the dependency object
keys), stripping comments/prose and splitting multi-import lines; (2) on a
de-duplicated table with a documented copy/occurrence count, choose COUNT(*) vs the
weight column from the population the question names, not silently. Steers off one
broad regex that drops valid identifiers and matches prose.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(analytics): source filters/dates/measures from the owning fact grain
Add a <sql_craft> rule for joined fact tables at different grains (parent order
vs child line item): read each predicate, calendar bucket, and measure from the
table whose grain the question names, not whichever is in scope post-join. An
order-grain filter ("orders that are Complete", "the order's creation date")
must come from the parent even though the child carries its own status/created_at;
line price/cost come from the child. Mirror at metric grain: don't combine a
parent-grain count with child rows (num_of_item * SUM(line_price) per line) —
aggregate each measure at its own grain before combining.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(analytics): collapse multi-valued classes to one representative per entity before counting/concentration
When an entity carries a multi-valued classification array (IPC/CPC codes, tags)
and the methodology counts entities-per-class or a concentration/diversity metric
(HHI, originality, share), pick ONE representative per entity first (the array's
main/primary/first flag, else a defined fallback like most-frequent), then
aggregate; and use COUNT(DISTINCT entity) when the denominator is defined as a
count of entities. Unnesting the array otherwise multiplies an entity's weight by
its code count, inflating per-class frequencies and skewing the ranking/score.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(connectors): introspect BigQuery datasets hosted in foreign projects
A dataset_ids/dataset_id entry may now be written `project.dataset` to
introspect a dataset hosted in another project while query jobs still bill to
credentials.project_id. Entries are parsed once at the config boundary into
canonical {project, dataset} pairs; introspection, primary-key discovery,
testConnection, getTableRowCount, and listTables (grouped per project) all
resolve in the dataset's own project, and scanned tables are labeled with that
project so sampling, distinct-value, and read queries resolve. Bare entries are
unchanged.
Implements spider2-specs/specs/18-bigquery-cross-project-datasets.md.
* feat(scan): durable, resumable, bounded relationship detection during enrichment
Move the enrichment persistence boundary to the cost boundary and bound the
open-ended relationship stage (spec 19).
- Checkpoint descriptions + embeddings into the queryable `_schema` manifest
(and the raw enrichment artifacts) before relationship detection runs, via a
new `onCheckpoint` hook + `writeLocalScanEnrichmentCheckpoint`. An interrupted,
budget-truncated, or failed relationship stage now degrades to "no joins",
never "no descriptions".
- Resume the enrichment cache by content identity: re-key the SQLite stage store
on `(connection_id, stage, input_hash)` so a re-run with a fresh runId resumes
finished descriptions/embeddings instead of re-paying for LLM work. The
disposable cache recreates its table if the on-disk key shape differs.
- Make the relationship stage observable and bounded: a sticky wall-clock budget
(`scan.relationships.detectionBudgetMs`, default 600000 ms) + per-unit progress
+ honored `ctx.signal`, threaded through profiling, validation, and composite
detection. On exhaustion/abort it stops scheduling, finalizes, and returns a
partial result instead of throwing or hanging.
- Mark a budget/abort-truncated result partial (diagnostics `partial`/`partialReason`
+ recoverable `relationship_detection_partial` warning). A graceful partial saves
as a completed stage and resumes cheaply; raising the budget changes inputHash
and forces a fresh, fuller run. A process killed mid-stage saves nothing.
Document `detectionBudgetMs` in the ktx.yaml reference. Append implementation
notes to specs/19 and move the intake draft to done/.
Also carries the in-tree per-table enrichment LLM timeout work it builds on
(`description-generation.ts` + the `enrichment_timeout` warning code), which is
intertwined in `local-enrichment.ts`/`types.ts` and cannot be split into a
separately-building commit.
* feat(scan): bound + retry the per-table enrichment LLM call
The batched table-description call had no retry (sampleTable retried 3x, this did
not), so a single transient backend error (e.g. an overloaded/burst rejection when
many tables enrich concurrently) silently nulled a whole table's descriptions —
observed dropping ~70% of a db's tables during a bad window despite ample quota.
- Wrap generateObject in retryAsync (3 attempts + backoff; KTX_ENRICH_LLM_ATTEMPTS).
- Fresh per-attempt timeout (KTX_ENRICH_LLM_TIMEOUT_MS, default 120s) still bounds a
wedged wide table; a timeout is surfaced as KtxAbortedError so it is NOT retried
(one wedge stays one timeout, not 3x).
- Granular per-table progress + start/done/retry/timeout logging.
Composes with spec 19 (its non-goal #1): spec 19 makes completed descriptions durable;
this makes more of them complete.
* feat(scan): survive a hung LLM enrichment backend and resume descriptions
Two compounding failure modes on the per-table description-enrichment path (spec 20):
Enforced per-table timeout for subprocess backends. The runtime declares whether it owns an SDK subprocess (subprocessForkSpec on KtxLlmRuntimePort); codex/claude-code calls run behind a ktx-owned detached child that is tree-killed (SIGKILL of the process group on POSIX, taskkill /T on Windows) on the deadline or ctx.signal, reaping the wedged model grandchild. HTTP backends keep native fetch abort. Default stays 120s, one-wedge-one-timeout.
Incremental, resumable descriptions persistence. generateDescriptions flushes enriched tables per batch to an inputHash-tagged durable record (at a stable, non-syncId path) plus only the changed manifest shards, skips already-enriched tables on resume, and never lets one table's failure discard the stage (a skipped table costs one missing description, not the whole stage's output).
Spec 20 refined + intake draft moved to done/.
* feat(scan): selective enrichment stages (--stages) + per-stage cache keys
Split the single coarse enrichment cache key into per-stage hashes
(descriptions <- snapshot + LLM identity; embeddings <- snapshot + embedding
identity + description digest; relationships <- snapshot + relationship settings
+ LLM identity), so changing one stage's inputs invalidates only that stage and
never throws away the expensive per-table descriptions on an unrelated edit.
Add `ktx ingest --stages <list>` to force-re-run a chosen subset on an
already-ingested connection: a named stage bypasses the completed-stage
short-circuit while the per-table descriptions resume record still skips
already-enriched tables, and unselected stages are left untouched on disk. Feed
embeddings + relationships their description context from the on-disk _schema
when descriptions do not run this invocation, and carry descriptions into the
llmProposals evidence packet (closing a latent gap on the full-run path too).
Surface an enrichment_stage_stale warning when an unselected stage's inputs have
drifted, rather than silently cascading the work.
Implements spider2-specs/specs/21-selective-enrichment-stages.md.
* test(analytics): realign SKILL.md acceptance test with the evolved skill
Three assertions in analytics-skill-content.test.ts drifted from the analytics
SKILL.md as later iterations edited the skill without updating the test:
- the sub-heading was renamed Window functions -> Ordering & aggregation
determinism (iter2), so follow the source name;
- the rule "Expose identity, not just the label" was renamed to "Project BOTH
identity and label" (spec 14), so match the new wording;
- the dialect-FQTN guard false-positived on the Java package example
com.planet_ink.coffee_mud, whose backticks made a 3-segment package path read
as a BigQuery/Snowflake `a.b.c` table reference. Drop the backticks so the
guard stays at full strength without weakening it.
* fix(scan): --stages subset must not delete unselected stages' on-disk artifacts
A --stages subset that omitted descriptions wiped all on-disk ai/db descriptions
from the written _schema. runLocalScan writes the structural manifest shard from
the bare snapshot BEFORE enrichment runs, and the shard merge treats ai/db as
scan-managed and overwrites them with whatever the run emits — none, on a subset
that skips descriptions. Enrichment then read the already-wiped shard via
loadPriorDescriptions and had nothing to restore.
runLocalScanEnrichment now returns the best-available descriptions (fresh-this-run
if descriptions ran, else loaded from the on-disk _schema) instead of [], and
runLocalScan captures the prior descriptions before the structural write and feeds
them to both the structural write and enrichment, so an unselected stage's
artifacts survive. Joins were already preserved for --stages descriptions via the
manual/inferred preservedJoins path.
Tests: a full runLocalScan --stages relationships path test (RED without the fix,
GREEN with it — the earlier unit test missed the structural-pre-write ordering),
plus enrichment-layer contract tests for both directions. Validated live on
northwind: --stages relationships keeps all 110 descriptions + 22 joins (was
wiping to 0); --stages descriptions restores descriptions from the spec-20 resume
record (no LLM calls) while keeping joins.
* feat(dialects): bigquery nested-data (ARRAY/STRUCT/UNNEST), geospatial (GEOGRAPHY), SAFE_DIVIDE
bigquery.md lacked the two sections that define BigQuery analytics (present in snowflake.md):
- Nested & repeated data: UNNEST to flatten arrays of STRUCTs (GA360 hits, GA4 event_params),
dot-notation field access, key-value param scalar-subquery extraction, fan-out/COUNT(DISTINCT) guard.
- Geospatial (GEOGRAPHY): ST_GEOGPOINT (lon-first), containment/proximity/distance/intersection
predicates, areal allocation via ST_AREA(ST_INTERSECTION()).
- SAFE_DIVIDE for zero-denominator-safe rates; sharded-table shard-presence note.
Generic BigQuery craft surfaced by sql_dialect_notes; product-completeness (any BQ analyst benefits).
* feat(dialects): sqlite ROUND half-up FP-underflow note (+1e-9 before ROUND)
SQLite ROUND(x,n) rounds half-away-from-zero, but binary FP stores an exact
half-way value just below it, so ROUND(6.475,2) returns 6.47 not 6.48. Add a
dialect note: nudge by a tiny epsilon (1e-9) below display precision before
rounding for deterministic half-up, leaving non-boundary values unchanged.
Generic SQLite craft surfaced by sql_dialect_notes (any analyst rounding a
displayed average/rate/price benefits).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(analytics): list-as-delimited-string, answer-literally, drop free-text columns
Add SKILL.md guidance to emit list-valued answer cells as delimited
STRING (not ARRAY/repeated column), answer the literal ask without
unrequested transformations (HAVING for aggregate bounds), and avoid
projecting unrequested free-text columns that corrupt row-delimited output.
* fix(scan,mcp): gitignore runtime logs, budget-guard LLM proposal, validate enrich timeout
- gitignore `.ktx/logs/` in both scaffold + setup-merge lists: the managed MCP
daemon writes raw tool params (SQL, memory_ingest content) to mcp.log under a
version-controlled `.ktx/`, and snowflake.log already sat there unprotected.
- gate the LLM relationship proposal on the detection budget/abort signal so an
exhausted or aborted stage cannot start a fresh LLM call; document the boundary.
- validate KTX_ENRICH_LLM_TIMEOUT_MS (NaN/0 → 120s default) like enrichAttempts,
so a bad value no longer times out every table immediately.
- daemon introspection now warns on malformed column/FK rows instead of dropping
them silently, matching the table-row path and the "surface broken objects" goal.
- docs: document `ktx wiki -c/--connection`; fix the SQLite query-deadline schema
doc (forked-subprocess SIGKILL, not worker-thread termination).
* fix(scan,wiki,mcp): address PR #312 review findings
- scan: key the description pipeline (resume map, enriched-schema and
embedding-text lookups, manifest write/read) by full table identity via
tableRefKey/buildTableRef, so two same-named tables in different schemas no
longer cross-assign descriptions or skip a sibling on resume
- scan: re-throw a genuine context cancel during the batched description LLM
call so Ctrl-C resumes the stage instead of nulling tables and recording it
completed; per-table timeouts still degrade (context.signal not aborted)
- scan: report statisticalValidation 'skipped' (not 'completed') when a
budget/abort stop leaves relationship profiling partial
- wiki: sync the full page corpus into the sqlite index and filter only the
candidate/result set, so a connection-scoped search no longer prunes other
connections' pages and cached embeddings from the shared index
- wiki: route verbatim ingest through the canonical writePageAndSync so
contentHash is set and later syncs can short-circuit
- mcp: drop the as-unknown-as cast in serializeMcpError
- dialects/analytics: document the integer-division trap on postgres/sqlite/tsql
Adds regression tests for each behavior change.
* fix(wiki): scope connection filter before SQLite lane limit
Connection-scoped wiki search applied the connectionId allowlist after
the lexical/semantic lanes had already truncated to laneCandidatePoolLimit
over the full (connection-agnostic) corpus. When the requested connection
was a minority of a large corpus, its pages were crowded out of the
candidate pool before filtering, so a semantic-only match could be missed
outright and lexical hits under-ranked.
Push the path allowlist into searchLexicalCandidates/searchSemanticCandidates
so LIMIT applies to in-scope rows, matching what the token lane already did,
and drop the now-redundant post-limit JS filters.
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* refactor(connectors): split KtxDialect into core and KtxSqlDialect
Separate the dialect contract into a driver-agnostic core (display/ref
formatting and type mapping) and a SQL-only extension (query generators).
The catalog and entity-details paths resolve the core dialect for any
snapshot driver, so it must stay free of SQL generation; this is the
prerequisite refactor for adding non-SQL primary sources.
- KtxDialect keeps type, formatDisplayRef, parseDisplayRef,
columnDisplayTablePartCount, mapDataType, mapToDimensionType
- KtxSqlDialect extends it with quoteIdentifier, formatTableName, and the
query/sample/statistics generators; the 7 SQL dialects implement it
- add getSqlDialectForDriver for SQL drivers; the 7 connectors and the
relationship-benchmark harness consume it
- thread the relationship pipeline (profiling/validation/composite/
discovery) as KtxSqlDialect | null so a non-SQL source skips coverage SQL
and its candidates stay in review; local-enrichment builds the SQL
dialect only when the connector advertises readOnlySql
Pure extraction: no behavior change for the existing 7 drivers.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(connectors): add MongoDB connector for issue #305
Add a read-only MongoDB connector that treats a database as a primary
context source: collections map to tables and inferred top-level fields to
columns. MongoDB is the first non-SQL source (readOnlySql: false), so
ktx sql and metric compilation do not apply, but its collections flow
through ingest, descriptions, and relationship discovery.
- schema-inference: infer a flat column schema from the most recent
sample_size documents (by _id desc, or order_by for non-ObjectId keys).
Union BSON types per field, mark multi-type fields mixed (string), keep
sub-documents/arrays as a single opaque json column, derive nullability
from presence, treat _id as the primary key
- connector: KtxMongoDbScanConnector behind an injectable client seam;
strictly read-only (find/listCollections/estimatedDocumentCount only),
no executeReadOnly; resolves env:/file: via resolveKtxConfigReference
- core-only KtxMongoDbDialect and a live-database introspection adapter
- wire the mongodb driver: driver union, dialect registry, driver
registration (scopeConfigKey databases), mongodbConnectionSchema,
connection-drivers, normalizeDriver, the live-database route, and the
ktx setup picker. ktx sql is refused by the read-only SQL capability gate
- tests: schema inference, connector snapshot via a fake client, dialect,
driver-schema parsing, and the ktx sql rejection
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(integrations): document the MongoDB primary source
Add a MongoDB section to the primary-sources reference: connection config
(url, databases, enabled_tables, sample_size, order_by), mongodb+srv/TLS/
Atlas notes, the schema-inference explainer, a features matrix, and the
non-SQL caveat. Update the frontmatter and connection field reference.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(connectors): address review blockers on the MongoDB connector
- introspect: skip estimatedDocumentCount for views. The count command is
rejected on a MongoDB view (CommandNotSupportedOnView), so counting a view
aborted introspect for the whole connection; compute estimatedRows only for
real collections, as ClickHouse does.
- sl: refuse a semantic-layer query against a non-SQL connection instead of
defaulting it to the Postgres dialect. compileLocalSlQuery (the shared CLI +
MCP path) now rejects a driver with no SQL dialect via the new
isSqlQueryableDriver authority, keeping MongoDB context-only per issue #305.
- tests: cover input.tableScope and the empty-scope skip for the Mongo
connector (the scan layer does not post-filter), the view no-count path, and
the ktx sl query refusal for a mongodb connection.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* polish(mongodb): compute sampled nullCount and document sampling caveats
Address the non-blocking review notes:
- sampleColumn now counts null/absent values over the sampled window instead of
returning nullCount: null, since the documents are already in hand
- warn that a custom order_by must be indexed (an unindexed sort hits MongoDB's
in-memory sort limit on large collections) in the connection schema and docs
- note that sampled values for nested fields are stringified, not faithfully
serialized, so the json opacity is deliberate
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(examples): add a MongoDB connector example
A manual, container-backed example mirroring examples/postgres-historic:
- docker-compose.yml + init/seed.js seed a representative dataset (nested
documents, arrays, a Decimal128, a mixed-type field, a nullable field, an
ObjectId reference, and a view) on first container start
- scripts/smoke.sh + introspect-smoke.mjs assert the connector's inferred
schema with no LLM credentials — the same introspection entry point ktx
ingest's database-schema stage uses, including the view-no-count path
- README.md documents the smoke and a full keyless ktx ingest run
(claude-code LLM + managed sentence-transformers embeddings)
Works with Docker Compose or podman compose. Verified end to end.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore: ignore examples/** in knip to fix dead-code false positives
The MongoDB connector example files (examples/mongodb/init/seed.js and
examples/mongodb/scripts/introspect-smoke.mjs) are used at runtime but were
flagged as unused by knip. Add examples/** to the ignore array, matching the
existing .context/** entry.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0114qQV8fJ5a5ME3XbMVRzbL
* fix(mongodb): refuse non-SQL connections before SQL analysis
`ktx sql` and the MCP sql_execution tool resolved a SQL-analysis dialect
(falling back to Postgres for a non-SQL driver) and ran read-only
validation before the connector capability gate refused the connection.
For a MongoDB connection that spun up the parser/daemon and produced
Postgres parser diagnostics instead of a clean non-SQL refusal.
Route both entry points through a shared assertSqlQueryableConnection
guard before dialect selection, mirroring compileLocalSlQuery. The
federated duckdb path has no driver and is exempted at each call site.
Add CLI and MCP regression tests asserting validation/connector work
never starts for a MongoDB connection.
* fix(mongodb): pass CI gates (dialect boundary, secrets, setup test)
Three latent failures in the connector surfaced once CI ran on the branch:
- connector.ts imported the concrete KtxMongoDbDialect, which the connector
dialect-import boundary forbids. Route it through getDialectForDriver('mongodb')
and widen inferKtxMongoCollectionColumns to the base KtxDialect (it only uses
mapDataType/mapToDimensionType).
- detect-secrets flagged a test ObjectId hex and the mongodb+srv example URL;
annotate both with allowlist pragmas.
- the "shows every supported database" setup test omitted the new MongoDB option.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Luca Martial <48870843+luca-martial@users.noreply.github.com>
Co-authored-by: Luca Martial <lucamrtl@gmail.com>
Co-authored-by: Andrey Avtomonov <andreybavt@gmail.com>
* feat(docs): add system theme option to theme toggle
* fix(docs): address review on theme toggle
Render the switcher as a proper radio group (role="radiogroup"/role="radio"
with aria-checked, roving tabindex, and arrow-key navigation) instead of
independent aria-pressed toggles, since the three options are mutually
exclusive. Update the component docstring to describe the three-option
switcher and note that it reads `theme` (not `resolvedTheme`) so the System
option can show as selected. Drop unrelated formatting churn.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Luca Martial <lucamrtl@gmail.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Luca Martial <48870843+luca-martial@users.noreply.github.com>
* docs(readme): improve CLI flags table readability
* feat: add more details
* docs(cli-reference): keep table default markers consistent
Revert the "no default" markers for --target and --no-input from em
dash back to the plain hyphen used in every other CLI-reference table,
and drop the code-font styling on the descriptive "ktx project dir"
default so it reads as text rather than a literal value.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Luca Martial <lucamrtl@gmail.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* fix(gdrive): validate folder access, run config test, harden Drive API
Connection test and setup validation now verify folder_id resolves to an accessible Drive folder before counting Docs, via a shared verifyGdriveFolderAndCountDocs helper, so a wrong or unshared folder fails instead of passing with 0 docs.
Move gdrive-config.test.ts under test/ so Vitest's test/** glob actually runs it; escape folder_id in the Drive query; add retry/backoff on transient Google API responses; and record skipped non-Google-Doc files in the staged manifest.
* chore: sync uv.lock to ktx-daemon/ktx-sl 0.13.1
* fix(sl): parse user filter expressions as predicates, not projections
User-authored filters and segments were parsed in a projection context
(`SELECT {expr}`). On T-SQL a top-level `col = 'value'` projection is the
`alias = expression` aliasing syntax, so an equality filter parsed this way
became `'value' AS col` — dropping the comparison entirely and silently
skipping computed-column expansion (the column hid behind the alias).
Parse user fragments as predicates (`SELECT * WHERE {expr}`) at every parse
site — the parser cache, measure-filter CASE WHEN generation, computed-column
expansion, and measure-filter/segment column qualification. For plain
non-condition expressions the column set is identical, so this is a no-op
everywhere except the T-SQL alias case it fixes.
Add cross-dialect regression tests (tsql, postgres, snowflake, bigquery)
locking equality filters/segments to comparison shape and confirming `= 'x'`
now matches `IN ('x')` on T-SQL.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Shorten T-SQL predicate comments
* docs(sl): tighten T-SQL predicate docstrings and AGENTS docstring rule
Trim the parser and regression-test docstrings to the 1-3 line bar and
extend the AGENTS.md comment guidance to cover docstrings explicitly.
* refactor(sl): route all filter parsing through parse_predicate
Consolidate the predicate-context parse into a single parse_predicate
helper and route every filter-parsing call site through it: measure
CASE-WHEN filters, segments, computed-column-in-filter, the
aggregate-locality HAVING rewrite, and the planner OR-mixing /
top-level-AND split. The locality and split paths still parsed user
filters in projection context, so a named-measure equality filter
compiled to `0 AS measure` on T-SQL. Add a locality regression test
covering the HAVING rewrite path.
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Andrey Avtomonov <andreybavt@gmail.com>
#301 strengthened resolveLocalConnectionId to reject connections absent
from ktx.yaml, but three sl query tests configured the warehouse
connection only in memory while runKtxSl reloads the project from disk,
so they returned exit code 1 once the check landed on main. Persist the
connection to ktx.yaml so it survives the reload and the tests exercise
the real on-disk load path.
* feat(duckdb): add @duckdb/node-api dependency for federation
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* refactor(connectors): extract resolveStringReference to shared module
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* refactor(connectors): route all identical connectors through shared resolveStringReference
Collapse the 5 remaining private copies in bigquery, clickhouse, mysql,
snowflake, and sqlserver into the shared module. Fix a latent bug in the
shared module where `~/path` was incorrectly sliced (dropping only `~`,
leaving the leading `/` and making resolve() ignore homedir). Add a
tilde-expansion test that caught the bug and now covers that branch.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(sl): reserve _ktx_ connection-id prefix for virtual connections
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(connections): derive virtual federated connection from compatible members
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(duckdb): federated executor builds READ_ONLY attaches and runs SQL
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(duckdb): close federated DuckDB instance and escape quotes in attach url
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(sl): union member source directories for _ktx_federated
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(query): route _ktx_federated through DuckDB executor
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(sl): use duckdb dialect for federated query compilation
Bypass assertSafeConnectionId for _ktx_federated in resolveLocalConnectionId
and loadComputableSources, and resolve the compute dialect to 'duckdb' when
connectionId is FEDERATED_CONNECTION_ID instead of falling through to the
default postgres lookup.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* test(duckdb): end-to-end cross-catalog federated join
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* test(duckdb): harden federated join test with multi-book join-key coverage
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(ingest): keep declared cross-DB joins to federated siblings
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(setup): surface federated connection availability after adding a member
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(setup): mark federationNoticeFor @internal for dead-code gate
Also marks attachTypeForDriver, buildAttachStatements, and
isReservedConnectionId @internal — all three are exported solely for
unit-test access with no production cross-file consumer.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs(concepts): document cross-database federation
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs(concepts): correct sqlite two-part naming in federation doc
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(duckdb): quote federated catalog alias so hyphenated connection ids attach
* refactor(duckdb): single-source federation driver list, dedup attach loads
Collapse the parallel ATTACH_COMPATIBLE_DRIVERS set and ATTACH_TYPE_BY_DRIVER
map into one map in federation.ts whose keys are the membership rule. Replace
FederatedMember.config (read only via a type-erasing cast) with a typed url
field extracted at derive time. Emit INSTALL/LOAD once per distinct driver
type instead of once per member.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(duckdb): close federated DuckDB instance on connect failure; dedup id validation
Wrap the federated DuckDB instance in its own try/finally so a failing
connect() or a throwing connection.closeSync() no longer leaks the native
instance. Route setup-sources connection-id validation through the canonical
assertSafeConnectionId so the reserved _ktx_ prefix guard applies there too.
Derive the federated dialect through sqlAnalysisDialectForDriver instead of a
hardcoded literal.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(federation): carry member connection config and projectDir on FederatedMember
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(federation): resolve per-member attach targets via canonical connector resolvers
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): quote mysql attach-string values like postgres
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): resolve member attach targets via canonical resolvers, supporting sqlite path:
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(federation): thread projectDir through deriveFederatedConnection callers
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(federation): add shared project read-only SQL executor that routes _ktx_federated
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* test(federation): exercise shared executor default federated path with real DuckDB
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(federation): route ingest query executor through shared executor
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): route MCP sql_execution _ktx_federated through shared executor
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): preserve cross-DB joins to federated siblings in manifest re-emit
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): preserve declared cross-DB joins through scan re-ingest
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(federation): document sibling-ref invariant, drop unsafe casts in test
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): namespace federated source names by member to avoid collisions
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(federation): document member-namespaced federated source names
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): preserve member SSL/search_path in attach, classify federated MCP errors
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(federation): simplify federated dispatch and parallelize sibling reads
Dedup the federated driver ternary in local-query, derive the prefixed
source.name from the already-built name, drop the duplicated error in
federatedAttachTarget's exhaustive switch, inline the one-line
cleanupConnector wrapper, and parallelize federatedSiblingTargets' shard
reads (was sequential await-in-for on the scan hot path).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(federation): carry headerTypes through shared SQL executor
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(federation): add shared federated connection listing builder
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): route ktx sql through shared executor for _ktx_federated parity
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(federation): show _ktx_federated in ktx connection list
Surfaces the virtual federated connection in the output of
`ktx connection list` so agents and users can discover cross-database
querying when 2+ attach-compatible connections are configured.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* feat(federation): surface _ktx_federated in MCP connection_list
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* test(federation): ktx sql federated cross-file join end-to-end
Drive runKtxSql with the real federated DuckDB executor against two on-disk
sqlite files, stubbing only SQL validation. The test surfaced that the JSON
output path could not serialize bigint values DuckDB returns for integer
columns; printJson now coerces bigint to JSON numbers, matching the
plain/pretty paths.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* docs(federation): document direct _ktx_federated query surface
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): coerce DuckDB bigint to number in shared federated executor
DuckDB returns integer columns as JS bigint, which JSON.stringify cannot
serialize. The CLI --json path worked around this with a replacer, but the
MCP sql_execution tool serializes via plain JSON.stringify and crashed on
any federated query selecting an integer column. Coerce bigint to Number
once in executeFederatedQuery so every consumer (CLI, MCP, ingest, SL)
gets a JSON-safe result, and remove the now-redundant CLI replacer.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* refactor(federation): simplify driver map and collapse forked MCP SQL path
- Replace the identity-valued ATTACH_TYPE_BY_DRIVER record with a
ATTACH_COMPATIBLE_DRIVERS Set; the driver name doubles as the attach
type, so the map encoded nothing beyond membership.
- Switch federatedAttachTarget directly on the driver with a default
throw, dropping the unreachable post-switch throw and its comment.
- Route the MCP sql_execution standard-connection case through the
shared executeProjectReadOnlySql instead of reimplementing the
connector create/capability-check/execute/cleanup ceremony, so
federated and standard connections share one execution path.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* chore(federation): allowlist placeholder credentials for detect-secrets
The federation doc example URL and the federated-attach test fixtures use
literal placeholder credentials that trip detect-secrets. Mark them with
line-scoped pragma allowlist comments so a real secret added later is still
caught.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix(federation): correct SL addressing, join pruning, and id-quoting guidance
- Federated SL list/search records carry the virtual `_ktx_federated`
connection id (member origin stays in the prefixed source name), so rows
round-trip to `ktx sl -c _ktx_federated read` and the fts index no longer
clobbers per-connection partitions.
- Prune semantic-layer joins by membership in the connection's own source set
instead of matching the target's first dotted segment against other
connection ids; a same-connection join whose target name collides with a
sibling connection id is preserved, and orphan targets that would poison the
planner are dropped.
- Document double-quoting for connection ids that are not bare SQL identifiers
(e.g. "books-db".public.books) in the federated naming hint, the sl-query
rejection error, and the federation docs.
- Preserve exact federated BIGINT values beyond 2^53 as strings instead of
rounding, and steer the setup federation notice to raw SQL against
`_ktx_federated`.
* fix(federation): carry ssl:true into postgres URL attach target
A postgres member configured with `url` plus `ssl: true` resolved to both a
connectionString and an ssl flag, but the federated attach builder early-returned
the bare URL and dropped the ssl intent. DuckDB then handed libpq a URL with no
sslmode, so the URL path silently diverged from the discrete-field path (which
emits sslmode=require) and from the direct scan path (which enforces TLS).
Append sslmode=require to the URL when the member sets ssl, unless the URL
already pins a stronger sslmode.
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Andrey Avtomonov <andreybavt@gmail.com>
The Claude Code auth probe wrapped every error as "authentication is not
usable", so a subscription session limit told users to re-authenticate —
which cannot help, since auth already succeeded and only a reset clears it.
Discriminate probe failures with describeClaudeProbeFailure: session-limit
and rate-limit hits get their own messages (preserving the upstream reset
text), and genuine auth errors keep the original guidance. Also add the
session/usage-limit markers to the shared rate-limit classifier so the
governor stops treating a cap as a generic error.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The MCP sql_execution/sl_query tools and the `ktx sql` CLI threw a plain Error naming no valid connection ids when an agent passed an unconfigured connectionId (or omitted it with multiple connections). The message reached the agent verbatim but gave it nothing to correct with, so it re-guessed for days, and each correct caller-driven rejection filed in PostHog Error Tracking as a ktx fault (issue 019eb10c, 8 occurrences on one install).
Add a shared resolver (resolveConfiguredConnection / resolveRequiredConnectionId) that throws KtxExpectedError listing the configured connections, and route the three SQL-execution call sites through it. Expected-error classification keeps these out of Error Tracking while the actionable message lets agents self-correct.
ktx commits under a synthetic identity (ktx <ktx@example.com>) that can
never own a GPG secret key. On a machine with commit.gpgsign=true, git
tried to sign every ktx commit and failed with "No secret key", breaking
ingest, scan, wiki, memory, and bootstrap commits.
Inject commit.gpgsign=false as a per-invocation -c override in the single
core git client factory every ktx commit flows through. This honors the
existing principle of not mutating the user's repo config, and is
locale-independent (no error-message matching).
Also harden the repo-isolation fixture helper to disable signing on its
raw commits so the suite is deterministic regardless of the contributor's
global git config.
Fixes KLO-735.
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Split the fused directory concept into projectDir (what the agent config
references) and installRoot (where project-scoped files are written), so
users can install .claude/, .mcp.json, skills, and rules where they open
their agent instead of only in the ktx project directory.
- Add --install-dir <path> (resolved against cwd, created if missing,
mutually exclusive with --global/--local, rejected for claude-desktop).
- Add an interactive directory menu: ktx project dir / Current directory
(hidden when it equals the project dir) / Custom directory… / Global
scope (shown only when every target supports it).
- Expand a leading ~ in typed/quoted paths so the ~/… menu hints round-trip.
- Record installRoot in the install manifest and merge key; thread it
through file planning, MCP config paths, summaries, and next actions.
- Refresh uv.lock to 0.12.0 for the editable ktx-sl and ktx-daemon packages.
Fixes a production crash-loop (PostHog issue 019eb68e): ktx mcp start
--foreground on a uv-less container eagerly installed the managed Python
runtime at boot, failed, and was restarted by its supervisor every ~62s
(122 exceptions from one install).
- MCP server factory now wires a lazy semantic-layer compute port that
defers the runtime install to the first call, mirroring the already-lazy
SQL-analysis port; the server boots and serves non-Python tools without
the runtime.
- ktx no longer requires uv on PATH: it downloads its own pinned,
sha256-verified uv build under the runtime root (KTX_RUNTIME_ROOT aware),
always musl-static on Linux. PATH uv is never consulted.
- uv is acquired before the version dir is wiped, so a failed download
cannot destroy an existing runtime.
- Acquisition failures (offline, intercepted download, unsupported
platform) throw KtxExpectedError and stay out of Error Tracking; a
missing binary inside a checksum-verified archive remains a plain Error.
- scripts/refresh-uv-manifest.mjs regenerates the pinned manifest
(packages/cli/src/managed-uv-release.ts) on uv bumps.
- Setup consent prompt now discloses the uv download; docs updated.
* fix(cli): double the height of the setup banner t crossbar
* fix(cli): unify setup multi-select hints and make Tab the select key
The six interactive multi-select surfaces in `ktx setup` documented three
different hint voices, one had no hint at all, and they named two different
select keys (Space vs Tab). Tab is the only key that can toggle selection
without colliding with type-to-search input, so make it the single documented
select key everywhere and compose every hint from one shared fragment
vocabulary in prompt-navigation.ts.
- Register `updateSettings({ aliases: { tab: 'space' } })` so Tab toggles flat
multiselects; the alias applies only to non-text prompts, leaving typed
search input (schema/Notion) untouched.
- Add the missing hint to the agent-targets prompt and drop the stray
"Space to select … Esc …" info line plus the now-dead writeSetupInfo helper.
- Replace the schema-scope ad-hoc hint with the searchable-multiselect voice
and standardize "filter" -> "search" vocabulary.
- Delete DEFAULT_TREE_PICKER_HELP_TEXT and the unused TreePickerChrome.helpText
seam; render the shared tree hint instead.
* refactor(cli): show LLM check progress for every setup backend
Rename runLlmHealthCheckWithProgress to validateModelWithProgress and
wrap the Claude subscription and Codex auth probes in the same spinner
progress as the Anthropic API and Vertex backends, so each backend shows
consistent "Checking <provider> LLM" output during setup.
* feat(cli): add ktx-orange progress spinners to setup steps
Add a shared runWithCliSpinner helper and a TTY-aware createCliSpinner:
an animated clack spinner in a terminal, and a static stderr-only spinner
before raw-mode pickers (the table tree picker and demo tour), where the
animated spinner's stdin grab would otherwise corrupt the next prompt.
Wrap the slow setup waits in progress spinners: managed runtime install,
embedding daemon start + first-run model download, embeddings health
check, the connection-test gate, and source validation / dbt clone /
Metabase discovery. Recolor every spinner frame from clack's magenta to
the ktx mascot orange (#FF8A4C) via the static helper and clack's
styleFrame option.
Move the live GitHub stars widget into the sidebar footer pill as a
type:"icon" link, sitting opposite the Slack mark (space-between) and
beside the theme switcher. Render it as inner content (GitHub mark +
star + compact count) so fumadocs supplies the anchor.
Replace the default fumadocs theme switcher with a custom two-icon
control where each icon selects its own theme. The built-in
"light-dark" mode is a single blind toggle that flips on any click, so
clicking the sun while already in light mode jumps to dark. useTheme is
sourced from fumadocs-ui/provider/base and the icons are inlined to
avoid bare next-themes / lucide-react imports.
* fix(cli): survive ktx.yaml version skew and derive repo ownership from disk
Loading ktx.yaml is now tolerant of keys this ktx version does not
recognize: they are stripped from the in-memory config (the file on disk
is never rewritten) and reported by ktx status as non-blocking warnings,
while invalid values on recognized fields still fail hard. Repo
ownership is derived from observed state (a .git directory plus a root
ktx.yaml) instead of a ktx.managed git-config marker, so projects
created by any past or future ktx classify identically. initKtxProject
now runs an explicit foreign-repo pre-check and writes ktx.yaml before
initializing git, so an interrupted init leaves only recoverable
residue instead of a bare .git misread as foreign.
* style(cli): trim comment blocks to constraint-only notes
* docs(agents): require constraint-only code comments
One ingest integration hiccup no longer discards a whole source:
- Replace the duplicated gate-repair and textual-resolver loops with one
shared constrained-repair loop whose success criterion is re-running
the failed check (verify), not whether the agent edited files. Verify
failures feed the retry prompt; maxAttempts is 2.
- Let the resolver declare a conflicting patch redundant: a verified
no-change resolution is accepted as subsumed instead of failing the
source (duplicate wiki-page creation from parallel work units).
- Carry per-source validation errors through validateWuTouchedSources
into gate messages and work-unit failure reasons instead of
discarding them.
- Move join-neighbor expansion into the shared validation path so
work-unit validation and integration gates check the same set.
- Reject joins whose target resolves to no source, at sl_write time and
in the gates, attributed to the declaring source. Resolution mirrors
the Python engine exactly (case-sensitive name within the
connection), with a case-mismatch hint for the writing agent.
Align the tree with AGENTS.md/CLAUDE.md conventions:
- Rewrite user-facing strings, docs, and tests to lowercase `ktx`
(no bare uppercase `KTX` tokens remain outside literal identifiers).
- Drop the legacy `historicSql` migration path and its now-unused
helpers, per the no-backward-compat rule.
- Remove `as unknown as` / `any` casts: narrow `BaseTool` generics to
`z.ZodObject`, add a typed `createLookerClient`, and delete the dead
`getParametersSchema`/`toAnthropicFormat` pre-AI-SDK helpers.
- Use `InvalidArgumentError` for Commander parse failures.
- Finish the adapter→connector prose conversion in the `ktx.yaml` docs
while keeping the literal `adapters` config key.
Verified the dbt, MetricFlow, LookML, Metabase, Looker, and Notion
sections of context-sources.mdx against the adapter code and fixed
claims that did not match the implementation:
- dbt: replace "test coverage" framing with the actual constraint/enum/
join derivation; name both overlay and wiki outputs; fix work-unit
granularity (per models/ schema file above 25 YAML files).
- MetricFlow: relationships come from entities (not dimensions); surface
the join edges they produce.
- LookML: chunking is one work unit per model (not connected component);
add the wiki output; note that a connection: mismatch disables SL writes.
- Metabase: dashboards are never fetched (no dashboard endpoint); work
units are per collection; "usage patterns" is really card output schema.
- Looker: drop invented "purpose/audience" framing; describe triage as a
prioritization gate; include Looks alongside explores and dashboards.
- Notion: not knowledge-only (it writes SL sources for mapped non-Notion
targets); remove the nonexistent database-schema extraction; reframe
"What it provides" as inputs; document root_data_source_ids.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Render a lowercase ktx half-block wordmark with the brand-orange gradient
above the `ktx setup` intro on interactive TTYs. The banner degrades
through truecolor, xterm-256, and monochrome, and is skipped on non-TTY,
non-Unicode, or too-narrow terminals.
Extract shared color/Unicode capability detection into io helpers
(shouldUseColorOutput, colorDepthForOutput, unicodeSupported) so the
banner and doctor report route through one implementation.
The "prints provider setup guidance when a skip-llm setup project runs
ingest" test drives runKtxSetup in --no-input (inputMode: disabled) mode
with a postgres warehouse but databaseSchemas: []. PR #286 changed the
no-input contract to require an explicit database scope instead of
auto-scanning, so setup now exits 1 and the test's first assertion
(resolves.toBe(0)) fails. #286 updated setup-databases.test.ts but missed
this runKtxSetup call in ingest.test.ts. Pass databaseSchemas: ['public']
to satisfy the new contract, matching #286's own test updates.
* fix(cli): isolate ktx project git repos
* fix(cli): remove inert auto commit config
* test(cli): drop stale auto commit fixtures
* docs: document isolated ktx project repos
* test(cli): keep stale config grep clean
* fix(cli): guide setup away from foreign repos at the project dir
ktx owns the git repo rooted at the project dir and refuses to adopt one it
did not create (the Finding 3 isolation invariant). But setup steered users
straight into that failure: the interactive menu offers "Current directory"
first, and `--no-input --yes --project-dir <repo-root>` created directly in
place — both then threw a generic "Failed to initialize git repository:"
wrapper from deep in GitService.initialize().
Extract the ownership rule into a shared `classifyKtxRepoOwnership(dir)` used by
both GitService.initialize() (the invariant) and the setup wizard (pre-flight
guidance), so the decision derives from one rule. Setup now detects a foreign
repo before constructing GitService and: interactively re-prompts (the user
picks the existing `ktx-project` subfolder), or non-interactively returns a
clean missing-input with the actionable message. The typed foreign-repo error
is also surfaced verbatim instead of being buried under the generic wrapper.
Empty/non-repo current directories still work — only foreign repos are blocked.
* fix(cli): keep classifyKtxRepoOwnership total for non-directory paths
The setup ownership guard runs before the existing not-a-directory check, so
pointing a custom/--project-dir path at a file made classifyKtxRepoOwnership
lstat `<file>/.git`, hit ENOTDIR, and throw — crashing the setup step instead
of returning the friendly "path exists and is not a directory" result.
A path that is a file (or missing) holds no git repo for ktx to avoid, so treat
ENOTDIR like ENOENT and return 'unowned'. The downstream existingFolderState
check still rejects a non-directory with its friendly message, and the
classifier no longer throws raw errno for any caller.
* fix: read semantic sources safely
* test: retarget reindex per-scope error case to a broken manifest
Reading a broken standalone source was made non-fatal in de1f1a8d (it is
surfaced for repair instead of throwing), so the reindex per-scope error
test no longer captured an error. Point it at a corrupt manifest shard,
which is the remaining fatal read failure the per-scope catch must
isolate, and assert the captured error names the offending file.
* fix(sl): decouple semantic-layer file names from warehouse naming rules
The in-file `name:` field is now the sole source identity; the filename is
a derived label that never participates in identity. This removes the
"Unsafe semantic-layer source name" failure class entirely: any warehouse
identifier (Snowflake's uppercase SIGNED_UP, EVENT$LOG, dotted names) can
be read, overlaid, edited, and deleted.
- New `source-files.ts`: one total filename derivation (safe lowercase
names verbatim; otherwise slug + sha256-hash suffix, immune to
case-insensitive-filesystem collisions) and one by-name file resolver.
- Reads resolve by name everywhere; the path-from-name fast path and
`assertSafeSourceName` are gone.
- Writes resolve-then-write: rewrites land on the file that declares the
name (human renames survive); new sources get a derived filename; a
derived path occupied by a different source fails instead of clobbering.
- `readSourceFile` returns null for missing files instead of forcing every
caller to launder IO errors; `deleteSource` distinguishes manifest-backed
sources from not-found instead of silently succeeding.
- `sl_write_source` accepts verbatim warehouse identifiers (snake_case is
now a recommendation for new sources) and rejects sourceName/source.name
mismatches; `sl_edit_source` rejects name-changing edits.
- Ingest projection commits, gate-repair allowlists, and touched-source
derivation use resolved paths / in-file names instead of interpolating
`<connId>/<name>.yaml`.
- Collapsed the five parallel path derivations and duplicated path-token
helpers onto the shared module; dropped dead service methods.
* fix(sl): resolve sources by declared name end-to-end and gate warehouse SQL with the parser-backed validator
- Key broken/renamed semantic-layer files by their recoverable in-file
name (slSourceNameForFile) so mid-edit sources stay reachable under
their real identity in reads, listings, and search
- Derive finalization touched sources from composed-source diffs and
recover deleted files' declared names from the pre-change commit
instead of parsing hash-derived filenames
- Resolve revert/rollback paths against history (listFilesAtCommit) so
human-renamed files are restored where they lived at preHead
- Validate ingest sql_execution through the daemon's sqlglot
validateReadOnly in the connection's dialect, sharing one
driver-to-dialect map (sql-analysis/dialect.ts) across MCP and ingest
- Harden the local read-only SQL backstop: accept leading comments,
reject smuggled second statements, and strip trailing
semicolons/comments before row-limit wrapping
