sql-code-graph 1.37.1__tar.gz → 1.38.1__tar.gz

{sql_code_graph-1.37.1 → sql_code_graph-1.38.1}/CHANGELOG.md

@@ -1,3 +1,23 @@
+## v1.38.1 (2026-06-15)
+
+> **No re-index needed.** `sqlcg viz` template change only; regenerate the
+> artifact with `sqlcg viz --out table_graph.html`.
+
+### Fix
+
+- **viz highlight is click-only**: hovering a node no longer highlights its 1-hop
+  lineage edges (it flickered the edges on every mouse move). The neighborhood
+  highlight is now driven solely by the click-pin — click a node to trace, click
+  again / background / Esc to clear. Hover still shows the node label.
+- **viz interaction-v2 lifecycle abandoned**: the gate-then-stay-live lifecycle
+  (warmup gating + `autoPauseRedraw(false)` + bounded `cooldownTicks` + a
+  "laying out…" cue) was reverted after live testing on the ~6,553-node DWH graph
+  — `autoPauseRedraw(false)` forced a perpetual full repaint that pegged the tab,
+  and the bounded-cooldown stop froze the nodes. The viz is back on plain
+  force-graph defaults (engine settles, repaint auto-pauses, zoom/pan/drag always
+  on). See [`plan/sprints/viz_interaction_at_scale.md`](plan/sprints/viz_interaction_at_scale.md)
+  (postmortem).
+
 ## v1.25.0 (2026-06-13)
 
 > **Re-index recommended (no schema change).** `SCHEMA_VERSION` unchanged.

{sql_code_graph-1.37.1 → sql_code_graph-1.38.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sql-code-graph
-Version: 1.37.1
+Version: 1.38.1
 Summary: SQL code graph analyzer and lineage tracer
 Project-URL: Homepage, https://github.com/Warhorze/sql-code-graph
 Project-URL: Repository, https://github.com/Warhorze/sql-code-graph

{sql_code_graph-1.37.1 → sql_code_graph-1.38.1}/docs/cli.md

@@ -473,11 +473,19 @@ and consumers that read the table only via COLUMN_LINEAGE / STAR_SOURCE
 one-hop consumer completeness ``find_table_usages`` already has. Stays
 one-hop (NOT transitive — that is ``downstream``'s job).
 
+With ``--through-temps`` the consumer set is augmented with the downstream
+**temp-collapsed** closure: physical tables reachable from this table only
+by threading through ``kind IN ('cte','derived','temp')`` intermediates,
+computed from the structural COLUMN_LINEAGE hops (read-only; it never
+mutates the graph). This recovers consumers — e.g. the star-into-temp
+final target — that the one-hop default and TEMP_INLINE do not surface.
+
 ### Options
 
 | Option | Type | Required | Repeatable | Default | Description |
 | --- | --- | --- | --- | --- | --- |
 | --raw | BOOLEAN | No | No | False | Disable noise filtering on results |
+| --through-temps | BOOLEAN | No | No | False | Also report consumers reachable only THROUGH a temp/cte/derived intermediate (table-grain temp-collapsed closure). Off by default — the default impact stays one-hop and does not thread temps. |
 
 ## `sqlcg analyze failures`
 

{sql_code_graph-1.37.1 → sql_code_graph-1.38.1}/plan/sprints/lineage_gold_regression_and_temp_collapse.md

@@ -382,6 +382,30 @@ fallback). Grep-confirm the call site lands in this PR.
 
 ---
 
+### Deviations
+
+#### Deviation 1: intermediate node = `kind IN ('cte','derived','temp')` AND namespaced (`::`)
+- **Reason**: The plan said thread *through* nodes whose `kind IN ('cte','derived','temp')`
+  and terminate at `kind IN ('table','external')`/unknown. Live indexing of the PR-1
+  synthetic fixtures showed the FINAL physical targets — the gold-set endpoints
+  `mart.dim_person`, `mart.f_orders`, `stgmart.f_compet_article` — are stamped
+  `kind='derived'` (they are INSERT/CTAS targets), NOT `'table'`. Threading through every
+  `derived` node would walk past the gold-set finals and never terminate (the recovered
+  edge `raw.src_emp_backup -> mart.dim_person` was returned EMPTY under the literal rule).
+- **Change**: A node is treated as a temp/cte/derived INTERMEDIATE to thread *through* only
+  when its `kind IN ('cte','derived','temp')` **AND** its key is namespaced
+  (`table_qualified LIKE '%::%'`). Namespacing (`<file>::<db>.<name>`, [`base.py:97-98`](../../src/sqlcg/parsers/base.py))
+  is the load-bearing discriminator: true CTE/temp intermediates are always namespaced,
+  whereas a `derived` INSERT/CTAS target is never namespaced and is a genuine endpoint.
+  Endpoints are everything else (non-namespaced, or kind not temp/cte/derived, or NULL).
+  This realises the plan's stated NODE-keyed intent against the actual stored data model.
+- **Impact**: No scope change; the headline acceptance (`raw.src_emp_backup -> mart.dim_person`
+  recovered, 4 gating absences not fabricated) holds and is asserted. The plan's
+  `_PHYSICAL_KIND_VALUES = ('table','external')` constant is therefore not used in the
+  traversal predicate (the namespacing test subsumes it); the temp-kind set
+  `('cte','derived','temp')` is still the authoritative constant, defined once in the module.
+- **Date**: 2026-06-15
+
 ## Versioning
 
 - **Base: master @ `1.37.0`.**