Release: dev into main - #717
Merged
Merged
Conversation
main → dev
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…nel shows (#480) * Classes, properties, attributes and knowledge edit in dialogs; the panel shows Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Groups fold, entries are rows, the hover card is a tooltip, every edge kind fans out Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Only the Danger zone icon carries the warning colour Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Relations reads From and To, the past folds in place, evidence is a source count Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The entity panel's rows carry the direction arrow, as on the ontology page Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A fact is one row: the relation on the left, the entity on the right Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Three records were written as 0026-0028 on three branches on the same day. By the order the pull requests were opened, the adjudication stack (#473-#475) keeps 0026-0028, the rules stack (#484, #485) becomes 0029 and 0030, and the event record (#487) becomes 0031. Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* A decision records why Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * An automatic merge is gated by what it can undo Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* The import history shows what each import actually did Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The import history does not say empty while it is loading Signed-off-by: WaylandYang <wayland0916@gmail.com> * One import is one line Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* The rule editor opens; only saving waits for a condition Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Rules read as a table, and a rule is written in a dialog Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The rules page has a search, on the same row as its buttons Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* A rule may say or, once Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The rule sentence says or where the rule says or Signed-off-by: WaylandYang <wayland0916@gmail.com> * The rule editor writes blocks Signed-off-by: WaylandYang <wayland0916@gmail.com> * The or record takes the number dev left open Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* A rule may read what a rule concluded, written down Signed-off-by: WaylandYang <wayland0916@gmail.com> * A rule reads what a rule concluded Signed-off-by: WaylandYang <wayland0916@gmail.com> * The chaining record takes the number dev left open Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com>
…n down (#523) * A Chinese name does not stop the queue Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * One knowledge base adjudicates one at a time Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A table keeps its rows Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A stated figure is written down Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The ontology offers what a class can have Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A value finds its subject the way a relation does Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A role is a fact, and a spacer column is not Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A blank header row gives way to the real one Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A list of names is a list of facts Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A bench that asks whether the document made it into the graph Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Extraction waits for the ontology it is about to search Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * cargo fmt Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com>
* A rule computes what it concludes, written down Signed-off-by: WaylandYang <wayland0916@gmail.com> * Aggregation is refused for what it claims, not for retirement Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com>
* An operand and a conclusion can be an arithmetic tree Signed-off-by: WaylandYang <wayland0916@gmail.com> * A rule computes what it concludes Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* The schema diagram draws the biggest classes, and reads lighter Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The class count has no stepper Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> (cherry picked from commit fe5ac50119a38553a947557e87aec8eb045a9ebc) * The name you point at reads like the name you picked Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The node you picked keeps its mark while you point at it Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A hovered name is painted once Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* The ontology reads as three tables Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The rail switches between the diagram and the tables Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* A fact row leads with the fact Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Evidence expands to something you can read Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* A fact row reads as a sentence Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> (cherry picked from commit 8b3a6cd5f83532e9da405851829c5b3133191631) * A floating panel does not borrow the colour behind it Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The class legend opens like every other panel Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A switched-off class is struck through, not just faded Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* Record source-scoped RSS summaries Signed-off-by: danwood <118035379+Floating-Y@users.noreply.github.com> * Scope RSS summaries to the source being listed Signed-off-by: danwood <118035379+Floating-Y@users.noreply.github.com> * Record 0026 becomes 0033 Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A source list counts only the listed source Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: danwood <118035379+Floating-Y@users.noreply.github.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* fix(store): wake workers when jobs are queued Publish transactional PostgreSQL notifications for new and immediately requeued jobs so idle workers do not wait for the polling interval. Closes #517 Signed-off-by: RKS <rajesh.sharma@owasp.org> * The idle wait is one function, and it sleeps when the listener fails Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A queued job wakes an idle worker, and not before its row Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: RKS <rajesh.sharma@owasp.org> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* The web app gets a unit-test runner Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A query stays fresh for as long as its key says Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* The ontology page renders before the ontology arrives Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The graph page waits the same way the ontology does Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* Selecting a class goes straight to its highlight Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A property with no classes does not dim the whole diagram Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* The chat loop runs on rig Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * rig comes without its provider clients Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The prompt says what the base is Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0042 records that the chat loop is a runner with hooks Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* A period is when a fact holds, and a unit leaves the value it was read from Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A period is neither subject nor entity Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Shape checks read structure, not words Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * An empty cell with a period after it is still empty Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A corporate form is left to identity, not to a word list Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A shape check keeps what it cannot tell apart Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* 0041 records that a name is a claim about an entity Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0041 settles that a name is a value fact and never a node Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The identity bench measures merges and splits in both arrival orders Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The identity bench pins an anchor to the fact its sentence is about Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0041 records what cut 1 measured and where it revised decisions 2 and 5 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* Schema columns carry primary-key, foreign-key and nullability flags (#502 cut 1) explore_mappings tells the model 'table X has these columns' and currently the model has no way to tell an ID column from a measure on a wide table (the 80-column TPC-H case in #502: proposal coverage 25%, the seven-query core revenue metric dropped because nothing in the schema said which column was the key). Adds four flags to SchemaColumn and surfaces them in the schema text the explorer prompt receives: * is_primary_key single-column PRIMARY KEY only; composite members stay false so the prompt does not mistake them * is_foreign_key the column itself is an FK * references_table <schema>.<table> the FK points at * nullable false for any NOT NULL column (PKs included) This cut lands Postgres end-to-end (catalog query + mapping consumer + three live tests gated behind UTOPIA_TEST_DATABASE_URL). MySQL, Trino, Snowflake and Databricks get the new struct fields with default-false so the workspace compiles, but their catalog queries are deferred to per-engine cuts (#502 cuts 2..5). Databricks intentionally stays is_primary_key=false across the board because Unity Catalog does not register PK/FK. Tests added (env-gated, all pass against postgres://localhost:1517): * a_primary_key_column_is_marked_and_not_null * a_single_column_foreign_key_carries_its_target * a_composite_primary_key_member_stays_unmarked The prompt format adds ' -- PK' / ' -- FK ->schema.table' / ' -- NOT NULL' markers after the existing comment, preserving line order so existing tests stay green. Signed-off-by: rollroyces <rollroyces@users.noreply.github.com> * Keys come from pg_catalog, and a lookup that fails leaves the columns Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: rollroyces <rollroyces@users.noreply.github.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: rollroyces <rollroyces@users.noreply.github.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> Co-authored-by: WaylandYang <145302500+WaylandYang@users.noreply.github.com>
…670) * 0041 records that a name is a claim about an entity Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0041 settles that a name is a value fact and never a node Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The identity bench measures merges and splits in both arrival orders Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The identity bench pins an anchor to the fact its sentence is about Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0041 records what cut 1 measured and where it revised decisions 2 and 5 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A name is a fact, and the extractor reports the other names it reads Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A name another entity already has sends the pair to the adjudicator Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The adjudicator sees an entity's other names Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A name already declared for another entity is not an alias Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A name keeps to its own path: decided pairs stay decided, the attribute stays builtin, a memory waits for a nod Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A name lookup stays inside its base Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0041 carries the cut-1 numbers measured on DeepSeek-V3 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* Schema columns carry PK/FK on MySQL/MariaDB (#502 cut 2) follows #671 (now merged with revisions). The struct change lands in #671 — this PR adds the MySQL implementation only, against the new shape (is_primary_key: bool, references_table: Option<String>). Drops the two fields (#671 cut: is_foreign_key, nullable) that the maintainer explicitly removed for not being used by the explore prompts. keys() helper mirrors postgres.rs's pattern: * single-column PRIMARY KEY from information_schema.statistics with the same composite-PK guard Postgres uses (index column count = 1) * single-column FOREIGN KEY from information_schema.key_column_usage with ORDINAL_POSITION = 1 + composite-FK guard (a column appears in key_column_usage once per FK constraint it's part of; without the composite guard, both members of a composite FK get marked) * UNION ALL on a five-column shape with NULL placeholders on the PK side, because sqlx's query_as wants a uniform tuple. PK and FK rows are distinguished in the merge loop by whether the FK target columns are NULL * Graceful degradation on the same line as #671: a failing keys() returns an empty Keys + a tracing::warn, leaving the columns without key marks rather than failing the whole fetch_schema call. The read-only role test exercises this on MySQL — same permission issue as PG (information_schema constraints are empty for SELECT-only users), same fix shape. fetch_schema keeps its existing CAST(... AS CHAR) treatment — MySQL 8.0 binary protocol reports VARCHAR columns in information_schema as VARBINARY, sqlx strict-typed decoding refuses to read them as String. Three live tests gated on UTOPIA_TEST_MYSQL_URL, mirroring the maintainer's three Postgres tests: * a_single_column_key_is_marked_and_a_composite_one_is_not_mysql — single-column PK and FK both marked; both members of a composite PK and a composite FK stay unmarked * a_foreign_key_points_at_its_own_target_mysql — self-referencing FK resolves correctly; cross-database FK resolves correctly; the PK on a table whose FK name collides with another table's PK stays marked as PK (not as FK) * a_read_only_login_still_sees_the_keys_mysql — a SELECT-only role still gets the PK/FK marks. information_schema reads succeed for SELECT roles (per #671 followup), but the test gracefully skips if CREATE USER / GRANT fail. Fx test harness mirrors postgres.rs's: per-test schema with random uuid suffix (CREATE DATABASE here — MySQL's 'schema' is 'database'), no Mutex for parallel safety, SET FOREIGN_KEY_CHECKS = 0 around cleanup so cross-database FK cleanup doesn't fight the drop order. Verified locally against mysql:8.0 on 127.0.0.1:13306 with UTOPIA_TEST_MYSQL_URL set: 269 passed, 1 failed (pre-existing a_live_server_answers_with_typed_values needs the sales.orders fixture from #316; not a regression). My new tests all pass; no regressions to existing tests. Signed-off-by: rollroyces <rollroyces@users.noreply.github.com> * MySQL keys are counted once per view, and the tests leave nothing behind Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: rollroyces <rollroyces@users.noreply.github.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: rollroyces <rollroyces@users.noreply.github.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
The entity panel's 'blocked' key is asked alongside derived, facts, and same_name, all of which already carry as_of through entity_detail. blocked_for_entity hard-coded 'v.status = ''open''', so a violation detected in March and closed in April still haunted the rewound panel. Wire as_of through blocked_for_entity and the entity panel callsite. violation_open_at is the same shape as derived_held_at / facts_held_at, with a different clock (detected_at / decided_at). Write paths (decide_violation, retract_from_violation, etc.) keep the existing status filter — 0019. Test (DB-backed, UTOPIA_DATABASE_URL-gated): - still_open: detected 3 月, never resolved - resolved: detected 3 月 20, resolved 4 月 15 回看 Mira 的 blocked： - 当下：still_open 一条 - 2 月：都未检出 → 0 条 - 3 月 25 日：resolved 在 decided_at 之前还开着 → 2 条 - 5 月：resolved 在 decided_at 之后消失 → still_open 一条 The two violations cross over between 3 月 25 日 and 5 月 — without as_of the same two rows always return just still_open today. cargo check -p utopia-store -p utopia-server clean. cargo fmt clean. cargo clippy -p utopia-store -p utopia-server --all-targets -- -D warnings clean. cargo test -p utopia-store --test blocked_for_entity_respects_as_of passes; the existing a_contradiction_points_upstream test still passes after passing None to the new signature. Signed-off-by: Royce <roycelam@umich.edu> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: WaylandYang <145302500+WaylandYang@users.noreply.github.com>
* Same-name peers follow the recording axis (#307) The same-name list on the entity panel is asked from the panel that already forwards as_of to entity_detail. same_name_peers hard-coded 'e.merged_into IS NULL', so a Zhang Wei merged in March stayed invisible at every as_of, and the panel could only ever report today's same-name collisions. Wire as_of through same_name_peers and the entity panel callsite. The write-side callers (update_entity, which always means today) pass None explicitly so their behaviour is unchanged. Test (DB-backed, UTOPIA_DATABASE_URL-gated): - zhang_wei_a: 1 月记下事实，3 月合并进 zhang_wei_b - zhang_wei_b: 1 月记下事实，从不合并（查询的锚点） - zhang_wei_c: 5 月记下事实，从不合并 回看 b 的同名： - 当下：c 一条 - 2 月：a、c 两条 - 4 月：c 一条 - 6 月：c 一条 a 在 2 月回到同名列、3 月合并后消失——没有 as_of 时 a 永远不在。 c 一直都在（5 月创建，在所有 ≥ 5 月的时刻数 1 条）。 cargo check -p utopia-store -p utopia-server clean. cargo fmt clean. cargo clippy -p utopia-store -p utopia-server --all-targets -- -D warnings clean. cargo test -p utopia-store --test same_name_peers_respects_as_of passes. Signed-off-by: Royce <roycelam@umich.edu> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> * A merged peer is counted by who held its facts at the time Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: Royce <roycelam@umich.edu> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: WaylandYang <145302500+WaylandYang@users.noreply.github.com>
…act format (#680) * A chunk reads the opening of its document Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A date keeps the contract format when a literal keeps its units Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The opening is the document's first live chunk, and a quote from it is not evidence Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * An obligation belongs to the agreement that imposes it Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
The point-search box on the graph page already passes as_of from
the time slider, but search_entities hard-coded node_sql(None, None)
in its degree subquery: a fact corrected in March still counts
toward a node's degree at every as_of, so the dropdown's numbers
disagree with the canvas they're meant to be picking from.
Thread as_of through graph::search_entities. NULL is unchanged
(today's degree); Some(t) hands the subquery the held_at predicate
plus a bound parameter. Tool callers (chat / MCP) keep passing
None so their behaviour is unchanged.
Test (DB-backed, UTOPIA_DATABASE_URL-gated):
- alpha: 1 live fact, no retraction. degree stays 1 at every as_of.
- beta: Jan fact, retracted March, new fact recorded May.
Today degree=1; Feb degree=1 (old fact alive); Apr degree=0
(retracted + not yet re-recorded); Jun degree=1.
The two cross over between Apr and Apr — without as_of the same
two rows always return degree=1 each.
cargo check -p utopia-store -p utopia-server clean. cargo fmt
clean. cargo clippy -p utopia-store -p utopia-server --all-targets
-- -D warnings clean. cargo test -p utopia-store --test
search_entities_degree_respects_as_of passes.
Signed-off-by: Royce <roycelam@umich.edu>
Signed-off-by: WaylandYang <wayland0916@gmail.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: WaylandYang <wayland0916@gmail.com>
Co-authored-by: WaylandYang <145302500+WaylandYang@users.noreply.github.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* Add single sign-on (OIDC) A narrow OIDC authorization-code flow (PKCE, one-use state, nonce, issuer/audience and RS256 verification), off by default and enabled only once all four UTOPIA_OIDC_* variables are set. Accounts must be explicitly linked by an administrator on a new Administration > Single sign-on page; a subject that has never been linked is refused at callback rather than silently provisioned. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0136c4gq3BDppPhcWcRdy5MJ Signed-off-by: Nafeeur Rahman <hello@nafeeur.nyc> * An identity is linked by its owner, and a refusal says why Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: Nafeeur Rahman <hello@nafeeur.nyc> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: WaylandYang <145302500+WaylandYang@users.noreply.github.com>
* A value that arrives late takes its place in history Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A value with no start ends where the next known one begins Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A timeline is placed whole, under one lock, whatever the order its values arrive Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A deadline relative to an event is kept as written and still ends the one before it Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Rule 8a points a relative deadline at rule 10's flag Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * An end the engine drew moves with what follows it Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A row on two timelines ends at the earlier side, and the engine keeps what people and texts said Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A chained revert brings facts home, a large deletion locks by predicate, and a read start follows its anchor Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* A date the text spells out is read as a date Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The interval ends and literal check read written dates, with tests Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Every prompt rule has its own number Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* A Databricks schema read falls back to the system information_schema, and a failed sync stops hiding the engine's error Signed-off-by: Qishang Zhong <zhongqishang@gmail.com> * A schema read falls back only when the table is missing, and an empty fallback is a failure Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: Qishang Zhong <zhongqishang@gmail.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* Connector timestamps carry an instant precision beside day #610 提案的第一刀。一个来源可以在 config 里写 "precision": "instant"， 连接器就把 created_at / updated_at 写成完整 RFC 3339（与 #351 后 change_line 用的同一形式），下游 facts.valid_from_precision 跟着变成 "second"。缺省仍是 day，旧来源的形状不变。 四件事： - Source::precision() 读 config.precision，缺省 day，typo 不翻转 - time_text::world 接受 Some("instant")，输出与无精度路径一致但语义 不同——给抽取器的精度字段是 second，不是 day - github_issues::render_with 接 precision 参数，render() 保持 day 精度 作为旧别名 - ingest_sources::sync_github_issues 把 source.precision() 传给 render_with 测试：721 个全过，3 个新加（time_text 的 instant 路径与跨午夜回归， models 的 precision 默认与 typo 政策，github_issues 的 instant 精度 端到端——含注释/历史/抬头三个时间点）。 设计稿：.roadmap-proposals/instant-precision.md 留待维护者回复的 5 个问题：默认、迁移、typo 政策、doc_time 单独、 JQL 游标。 Signed-off-by: rollroyces <rollroyces@users.noreply.github.com> * GitHub and Jira write their timestamps to the second, with no setting Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: rollroyces <rollroyces@users.noreply.github.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: rollroyces <rollroyces@users.noreply.github.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> Co-authored-by: WaylandYang <145302500+WaylandYang@users.noreply.github.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
…hing (#696) * A confident merge waits for no one, and a succession clashes with nothing Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A clash one side already had is not the merge's Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* The per-chunk ontology list keeps to the prompt budget Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * Test strings spell their line breaks Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0006 records that the per-chunk list keeps to the budget Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* An agreement keeps its identity through its amendments Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * A name that raises doubt is settled by what the second look finds Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The governance bench waits out a deferred extraction Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * 0025 records what the second look settles Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> * The settled shape has a test Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* Say which test and save each Models card's note refers to Signed-off-by: Aniruddha Adak <aniruddhaadak80@users.noreply.github.com> * A Models card with unsaved changes says so instead of testing the saved one Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: Aniruddha Adak <aniruddhaadak80@users.noreply.github.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Aniruddha Adak <aniruddhaadak80@users.noreply.github.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com> Co-authored-by: WaylandYang <145302500+WaylandYang@users.noreply.github.com>
…ine (#715) * The extraction progress counts each document once, not once per pipeline Signed-off-by: Qishang Zhong <zhongqishang@gmail.com> * The progress counts are tested against live, deleted and failed documents Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> --------- Signed-off-by: Qishang Zhong <zhongqishang@gmail.com> Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Signed-off-by: WaylandYang <wayland0916@gmail.com> Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
9 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merges
devintomain(139+ commits since #478). Includes, among others, the chat loop on rig (#548), names as facts (#670), the temporal engine rework (#679), OIDC single sign-on (#684), written dates (#689), second-precision connector timestamps (#691), the Databricks schema fallback (#702), reasoning-model replies (#707), clearer Models cards (#708), the entity names panel (#711) and the extraction progress fix (#715).Not included: #699 (held after review), #710 (record being rewritten), #712 and #713.
No tag in this release.
🤖 Generated with Claude Code