sql-code-graph 1.34.4__py3-none-any.whl → 1.35.1__py3-none-any.whl

{sql_code_graph-1.34.4.dist-info → sql_code_graph-1.35.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sql-code-graph
-Version: 1.34.4
+Version: 1.35.1
 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-1.34.4.dist-info → sql_code_graph-1.35.1.dist-info}/RECORD

@@ -1,8 +1,8 @@
-sqlcg/__init__.py,sha256=jFojpr6ZojswTJKMfPZEdxqOfiqZPTpuB2Cb9aa1RM0,116
+sqlcg/__init__.py,sha256=VahDFz44BfVysbx4bUBilcYdqMPT5hNDCMqoIkgonw0,116
 sqlcg/__main__.py,sha256=1YoFLcqEgTwYq1J3TbUwpkdG0zeeLIf2fJvwWI-CLFU,109
 sqlcg/cli/__init__.py,sha256=W8fD0LpMq2xm_5WKGNMvJh2WBL1ho5E8hUeAqXQYT1g,28
 sqlcg/cli/coverage.py,sha256=Xm9ITzZDHv2mJ70Q5jCacVuhDStVrE3gq12_-Ypvtd8,43823
-sqlcg/cli/main.py,sha256=tlFvl-l9hWfZBP9dO2svZXvLDbh7T5731CFyctuVqDI,2485
+sqlcg/cli/main.py,sha256=K7CDIJ1tz7kxySGLeuKGZoTinucUrhINCxs_zieMmsg,2526
 sqlcg/cli/commands/__init__.py,sha256=oSHtr6VD-jNubOjuCQyZj2tBppjMEpQDh-IGQ8of9eA,30
 sqlcg/cli/commands/analyze.py,sha256=lC7QeBKljWVpKQ-x8BOF5Ok6y4jd329U1aBGrp159OA,31046
 sqlcg/cli/commands/catalog.py,sha256=xCIWp_9xcDPHIMUB0xrZR3-G0WMvKbfq4yMwe2sDT2g,12670
@@ -16,9 +16,10 @@ sqlcg/cli/commands/mcp.py,sha256=drh5udcroZFddmJAG3q9JWxekDqTO9H4xC-oIqiis-Y,148
 sqlcg/cli/commands/reindex.py,sha256=6j1QL3TjrsnWaFSIKvpGJ3LF8CoCjAbue0UFtmk9FS0,22704
 sqlcg/cli/commands/report.py,sha256=JU0qjyMxwOukE7bN3XvvIzOI7zMg_Gsnvk_8F6pKNpA,4915
 sqlcg/cli/commands/uninstall.py,sha256=WrA1FnINxnd6mmE4-_QBK0aHBnibstJeAT8swnQKG4M,8962
+sqlcg/cli/commands/viz.py,sha256=2GAP6YSNuFXmWM8D8IohkaO5GJ3nj2KLdA0kHLx137M,3779
 sqlcg/cli/commands/watch.py,sha256=_OvUQ4m_Pp2lAzI_f9UX2Gwb56zbsaSC4Tj8h51uciA,2806
 sqlcg/core/__init__.py,sha256=dXvLWpbQ72f5CM6sKSvBDnEGqGuIZaN5MmHyD8Vf1aA,154
-sqlcg/core/config.py,sha256=QxhIiF_let_ablMa-WttGbcAlCCDIpT4lR4tUVymIEI,19091
+sqlcg/core/config.py,sha256=dOQUQCsMGgT8Vp8gIfLdmMasB2IozGvHpAsjftCyDXw,20552
 sqlcg/core/duckdb_backend.py,sha256=zPsxiO2_qcE6axVZu8P3gUNHnJ1cgCCFv1XXy_t_LEw,42298
 sqlcg/core/freshness.py,sha256=Cm-w7jACPrt8TIstZpCzDJpYsHCiBo63ni5js8wCGbg,6248
 sqlcg/core/graph_db.py,sha256=3iZ1J32cEn7_cVhNA1tLiNuJHCyatN4F40HXLLfmwHY,10815
@@ -43,7 +44,7 @@ sqlcg/lineage/schema_resolver.py,sha256=iXt6LYF6UVWsGUpcfbmjmGn9wCgXl721lTGf_8Aa
 sqlcg/metrics/__init__.py,sha256=hLJ6wm4St8qqYwKh3o9QG7lcEt1BEYM31ccqO9tGpIg,133
 sqlcg/metrics/store.py,sha256=KuDtxvyAgug9_KtiSCpvgKM2VZM7VSaI3D11uMLjJJk,10604
 sqlcg/parsers/__init__.py,sha256=AamA8wBbDZV9_zEtZCI4Hyen5UAVKHmBwjTghTt2PZE,785
-sqlcg/parsers/ansi_parser.py,sha256=Iayv93KDOs3qKrQlrPcpQRd4_K8dEsGHOaTY076e-GY,38032
+sqlcg/parsers/ansi_parser.py,sha256=RX6eVj7gt1qmsHNJLAF_a4jyW3RCI5W2oF4rd53cKNg,39336
 sqlcg/parsers/base.py,sha256=Q2X_qN7DmGFU2oA6Z9fcxyjpkeb6lOsia4BcwPvwNLo,97806
 sqlcg/parsers/bigquery_parser.py,sha256=mOnWTfXB_Dp4JwFE1PVYOB6CDPf5nYE0Dea8kJCl9uQ,2827
 sqlcg/parsers/postgres_parser.py,sha256=lYfUpQY6j4Qm7ndXBtXbgPoGzYqYddWt5YeFnWKdA6I,946
@@ -65,7 +66,13 @@ sqlcg/utils/__init__.py,sha256=--iqt5ThTXmT8Wz7da8hs3n0zDfYPl8P-z5OgRJ_77E,154
 sqlcg/utils/hashing.py,sha256=H25-sYfxHKb3_IERFnHyAIYNiXN470Oqo5sJT_D3YOA,438
 sqlcg/utils/ignore.py,sha256=6__DfiYgynI2b8nH52O2BzI1EZcZ_LRazhqEOJVGGcE,1955
 sqlcg/utils/logging.py,sha256=u0fCmYsLj9o81vawm3xZTHaw68GQYVm7JxG-gP81u8A,840
-sql_code_graph-1.34.4.dist-info/METADATA,sha256=tzuVGUb49jlzRUkFgV28LwMdvIR0h83TsN9zpFHFA0A,17791
-sql_code_graph-1.34.4.dist-info/WHEEL,sha256=mffPy8wBnZQn2VnJUU5jE99KsxaSfiyMHV9Yt0aLVxs,87
-sql_code_graph-1.34.4.dist-info/entry_points.txt,sha256=Wfe49sVzV9p4eVFGo5RxcV-frr3HOP0yzzst8JBxQLQ,46
-sql_code_graph-1.34.4.dist-info/RECORD,,
+sqlcg/viz/__init__.py,sha256=ha-RMeqhhOoAsDQOfOjCcO0mnDKOU8MgHlK_Yxs4O8k,455
+sqlcg/viz/data.py,sha256=deLWOZBgewM1x-kk1NhUnt0mswyt0nzqSevHzBCOHHc,5899
+sqlcg/viz/render.py,sha256=BINkGbJbbb_iqhrkN795RaQsdg8nqCiJtsEFF1yo22Y,2737
+sqlcg/viz/tags.py,sha256=6zRnGlHjuGmEeB6yN1uhzm8rqL7ZGoyL1Ki7jI5oM6A,5368
+sqlcg/viz/assets/force-graph.min.js,sha256=jNdYdDdrYiUdUlElxRkolPBt30rstQk2q15Q32VVdzc,177272
+sqlcg/viz/assets/template.html,sha256=9_j-mvo1ZxwgiJPDdVrNmca37dTrTjjYVd3977u-DxE,12294
+sql_code_graph-1.35.1.dist-info/METADATA,sha256=YOc2OzDrdadgZdpmY3QhwueUwScKj39svzKO4dgd9Q0,17791
+sql_code_graph-1.35.1.dist-info/WHEEL,sha256=mffPy8wBnZQn2VnJUU5jE99KsxaSfiyMHV9Yt0aLVxs,87
+sql_code_graph-1.35.1.dist-info/entry_points.txt,sha256=Wfe49sVzV9p4eVFGo5RxcV-frr3HOP0yzzst8JBxQLQ,46
+sql_code_graph-1.35.1.dist-info/RECORD,,

sqlcg/__init__.py

@@ -1,5 +1,5 @@
 """SQL Code Graph - SQL lineage and dependency analysis tool."""
 
-__version__ = "1.34.4"
+__version__ = "1.35.1"
 
 __all__ = ["__version__"]

sqlcg/cli/commands/viz.py

@@ -0,0 +1,102 @@
+"""The sqlcg viz command — generate a self-contained graph-explorer HTML.
+
+Plan: plan/sprints/feature_graph_viz.md §"CLI signature".
+
+Queries the live graph (via the server-aware ``run_read_routed`` seam),
+attaches per-table label facets (tags/jobs) from BYO CSVs, and writes ONE
+self-contained ``.html`` with the force-graph library inlined — openable on
+Windows via ``file://``. Touches no parser/indexer/lineage code.
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import typer
+from rich.console import Console
+
+from sqlcg.core.config import get_viz_schemas
+from sqlcg.utils.logging import getLogger
+from sqlcg.viz.data import build_viz_data
+from sqlcg.viz.render import DEFAULT_VIZ_OUT, render_html
+from sqlcg.viz.tags import FacetMap, load_facet, resolve_labels
+
+logger = getLogger(__name__)
+console = Console()
+
+
+def _facet_payload(facet: FacetMap) -> dict:
+    """Shape a FacetMap for baking: ``{labels, colors}``."""
+    return {"labels": facet.labels, "colors": facet.colors}
+
+
+def viz_cmd(
+    tags: Path | None = typer.Option(  # noqa: B008
+        None,
+        "--tags",
+        help=(
+            "BYO tags CSV (pattern,label[,color]; header required). The 'tag' "
+            "facet drives the legend swatches (color + filter). Patterns are "
+            "fnmatch globs (*, ?, [seq]) matched case-insensitively against the "
+            "qualified table name. Many-to-many. Omit to hide the tag legend."
+        ),
+    ),
+    jobs: Path | None = typer.Option(  # noqa: B008
+        None,
+        "--jobs",
+        help=(
+            "BYO jobs CSV, same shape as --tags. The 'job' facet drives the job "
+            "dropdown (filter only). Omit to leave the dropdown empty."
+        ),
+    ),
+    out: Path | None = typer.Option(  # noqa: B008
+        None,
+        "--out",
+        help=f"Output HTML path (default: {DEFAULT_VIZ_OUT} in the current directory).",
+    ),
+    config_dir: Path = typer.Option(  # noqa: B008
+        Path("."),
+        "--config-dir",
+        help="Directory searched for .sqlcg.toml ([sqlcg.viz] schemas).",
+    ),
+) -> None:
+    """Generate a self-contained graph-explorer HTML from the live graph.
+
+    Reads the graph via the server-aware routed-read path (never a direct DB
+    open). The emitted file inlines the force-graph library plus the baked node/
+    edge/adjacency data and the tag/job facets — no external resources — so it
+    opens by double-click on Windows.
+
+    The client-side filter composes schema ∩ kind ∩ tag (multi-select per facet,
+    union within a facet, intersection across) plus a job dropdown; an edge
+    renders only when both endpoints are visible. Node kinds table/view/temp are
+    ON by default, cte/derived OFF (switchable).
+    """
+    out_path = out if out is not None else Path(DEFAULT_VIZ_OUT)
+
+    schemas = get_viz_schemas(config_dir)
+
+    data = build_viz_data()
+
+    tag_facet = load_facet(tags, "tag") if tags is not None else FacetMap(name="tag")
+    job_facet = load_facet(jobs, "job") if jobs is not None else FacetMap(name="job")
+
+    # Attach resolved facet labels to every node (m2m glob resolution).
+    for node in data["nodes"]:
+        node["tags"] = resolve_labels(node["id"], tag_facet)
+        node["jobs"] = resolve_labels(node["id"], job_facet)
+
+    html = render_html(
+        data,
+        _facet_payload(tag_facet),
+        _facet_payload(job_facet),
+        schemas,
+    )
+    out_path.write_text(html, encoding="utf-8")
+
+    schema_count = len(schemas) if schemas else len({n["schema"] for n in data["nodes"]})
+    console.print(
+        f"Wrote {out_path} — {len(data['nodes'])} nodes, "
+        f"{len(data['links'])} edges, {len(tag_facet.labels)} tags, "
+        f"{len(job_facet.labels)} jobs, {schema_count} schemas."
+    )

sqlcg/cli/main.py

@@ -16,6 +16,7 @@ from sqlcg.cli.commands import (
     reindex,
     report,
     uninstall,
+    viz,
     watch,
 )
 
@@ -84,6 +85,7 @@ app.command("reindex")(reindex.reindex_cmd)
 app.command("watch")(watch.watch_cmd)
 app.command("gain")(gain.gain_cmd)
 app.command("report")(report.report_cmd)
+app.command("viz")(viz.viz_cmd)
 app.command("install")(install.install_cmd)
 app.command("uninstall")(uninstall.uninstall_cmd)
 

sqlcg/core/config.py

@@ -151,6 +151,44 @@ def get_noise_filter_patterns(path: Path) -> list[str]:
     return default_patterns
 
 
+def get_viz_schemas(path: Path) -> list[str]:
+    """Get the curated viz schema list from .sqlcg.toml.
+
+    Reads [sqlcg.viz] -> schemas (a list of schema-name strings) from
+    .sqlcg.toml. Used by ``sqlcg viz`` to decide which schemas get a distinct
+    hue / appear as multi-select facets. When the block/key is absent or not a
+    list, returns ``[]`` and the generator falls back to **data-derived**
+    schemas (every distinct schema present in the graph), preserving the
+    zero-config small-repo path::
+
+        [sqlcg.viz]
+        schemas = ["ba", "da", "ia_analytics"]
+
+    This reader mirrors ``get_schema_aliases`` / ``get_noise_filter_patterns``
+    exactly (same .sqlcg.toml open, same ``except Exception: pass`` guard, same
+    fallback shape). It is presentation-only config; it never touches indexing
+    or lineage.
+
+    Args:
+        path: Root directory to search for .sqlcg.toml
+
+    Returns:
+        List of configured schema names. ``[]`` when the key is absent or
+        malformed (caller derives schemas from the data).
+    """
+    config_file = Path(path) / ".sqlcg.toml"
+    if config_file.exists():
+        try:
+            with open(config_file, "rb") as f:
+                config = tomllib.load(f)
+            raw = config.get("sqlcg", {}).get("viz", {}).get("schemas")
+            if isinstance(raw, list):
+                return [s for s in raw if isinstance(s, str)]
+        except Exception:
+            pass
+    return []
+
+
 def get_ignored_tables(path: Path) -> list[str]:
     """Get explicitly-ignored qualified table names from .sqlcg.toml.
 

sqlcg/parsers/ansi_parser.py

@@ -551,25 +551,44 @@ class AnsiParser(SqlParser):
         # measurement artifact (96.1%/4,544 of 4,727 on the DWH corpus).
         can_have_lineage = self._can_have_column_lineage(stmt)
 
+        # Phantom source-node gate (#161): a non-table/non-view CREATE
+        # (SEQUENCE / STAGE / FILE FORMAT / SCHEMA / INDEX / PROCEDURE / ...) has
+        # target=None and no data-flow source. Its only Table node is its own
+        # object name, which _fallback_table_scan below would otherwise scoop into
+        # `sources`, leaking a phantom SqlTable SOURCE node + SELECTS_FROM edge
+        # into the graph. #159 fixed the kind LABEL (-> OTHER); this closes the
+        # source-node half. Keying off stmt.kind (the sqlglot AST value, e.g.
+        # 'SEQUENCE') keeps the gate independent of the _classify/QueryKind path.
+        # CREATE TABLE/VIEW ... AS SELECT keep kind in (TABLE, VIEW) and so retain
+        # their real sources. Real procedure lineage comes from the spliced inner
+        # DML (a separate node the gate never touches), so gating PROCEDURE is
+        # edge-neutral. This single gate feeds all three
+        # `referenced_tables.extend(query_node.sources)` sites
+        # (ansi_parser.py:246, snowflake_parser.py:784 / :944).
+        is_non_table_create = isinstance(stmt, exp.Create) and stmt.kind not in ("TABLE", "VIEW")
+
         # Try to extract table references using scope analysis
         try:
-            # Use pre-built scope if provided, otherwise build it here (fallback)
-            root_scope = scope
-            if root_scope is None:
-                from sqlglot.optimizer.scope import build_scope
-
-                root_scope = build_scope(stmt)
-
-            if root_scope:
-                sources = self._real_tables(root_scope)
+            if is_non_table_create:
+                sources = []
             else:
-                # Fallback to basic table extraction
-                sources = [
-                    r
-                    for s in self._fallback_table_scan(stmt)
-                    if (r := self._apply_table_alias(s)) is not None
-                ]
-                parse_failed = can_have_lineage
+                # Use pre-built scope if provided, otherwise build it here (fallback)
+                root_scope = scope
+                if root_scope is None:
+                    from sqlglot.optimizer.scope import build_scope
+
+                    root_scope = build_scope(stmt)
+
+                if root_scope:
+                    sources = self._real_tables(root_scope)
+                else:
+                    # Fallback to basic table extraction
+                    sources = [
+                        r
+                        for s in self._fallback_table_scan(stmt)
+                        if (r := self._apply_table_alias(s)) is not None
+                    ]
+                    parse_failed = can_have_lineage
         except Exception as exc:
             logger.debug("Failed to build scope for statement %d in %s: %s", stmt_index, path, exc)
             sources = [

sqlcg/viz/__init__.py

@@ -0,0 +1,10 @@
+"""Self-contained graph-visualization generator for ``sqlcg viz``.
+
+Plan: plan/sprints/feature_graph_viz.md
+
+A pure read + render feature: queries the live DuckDB graph (via the
+server-aware ``run_read_routed`` seam), resolves query-anchored edges to
+table->table links, attaches per-table label facets (tags/jobs) from a BYO
+CSV, and emits ONE self-contained HTML file with the force-graph library
+inlined. No parser/indexer/lineage code is touched.
+"""