iris-vector-graph 2.4.5__tar.gz → 2.4.6__tar.gz

{iris_vector_graph-2.4.5 → iris_vector_graph-2.4.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: iris-vector-graph
-Version: 2.4.5
+Version: 2.4.6
 Summary: Native graph database layer for InterSystems IRIS — openCypher, hybrid vector search, temporal property graph, RDF export, SHACL validation
 Project-URL: Homepage, https://github.com/intersystems-community/iris-vector-graph
 Project-URL: Documentation, https://github.com/intersystems-community/iris-vector-graph/tree/main/docs

{iris_vector_graph-2.4.5 → iris_vector_graph-2.4.6}/iris_src/src/Graph/KG/TemporalIndex.cls

@@ -164,6 +164,38 @@ ClassMethod QueryWindow(source As %String, predicate As %String, tsStart As %Int
 	Return result _ "]"
 }
 
+ClassMethod QueryWindowSources(predicate As %String, tsStart As %Integer, tsEnd As %Integer) As %String
+{
+	Set tsStart = +tsStart
+	Set tsEnd = +tsEnd
+	Kill dedup
+	Set ts = tsStart - 1
+	While 1 {
+		Set ts = $Order(^KG("tout", ts))
+		If ts = "" Quit
+		If ts > tsEnd Quit
+		Set s = ""
+		While 1 {
+			Set s = $Order(^KG("tout", ts, s))
+			If s = "" Quit
+			If predicate = "" {
+				Set dedup(s) = 1
+			} Else {
+				If $Data(^KG("tout", ts, s, predicate)) Set dedup(s) = 1
+			}
+		}
+	}
+	Set result = "[", first = 1, src = ""
+	For {
+		Set src = $Order(dedup(src))
+		Quit:(src = "")
+		If 'first { Set result = result _ "," }
+		Set first = 0
+		Set result = result _ """" _ src _ """"
+	}
+	Return result _ "]"
+}
+
 ClassMethod GetVelocity(nodeId As %String, windowSec As %Integer = 300) As %Integer
 {
 	Set now = $ZTimeStamp

iris_vector_graph-2.4.6/iris_src/src/Graph/KG/kgNodeEmbeddings.cls

@@ -0,0 +1,12 @@
+Class Graph.KG.kgNodeEmbeddings Extends %Persistent [ DdlAllowed, Final, SqlTableName = kg_NodeEmbeddings ]
+{
+
+Property id As %String(COLLATION = "EXACT", MAXLEN = 256) [ Required, SqlColumnNumber = 2, SqlFieldName = id ];
+
+Property emb As %Library.Vector(DATATYPE = "double", LEN = 384) [ SqlColumnNumber = 3, SqlFieldName = emb ];
+
+Property metadata As %Library.DynamicObject [ SqlColumnNumber = 4, SqlFieldName = metadata ];
+
+Index IDKEY On id [ IdKey, PrimaryKey, Unique ];
+
+}

iris_vector_graph-2.4.6/iris_src/src/Graph/KG/kgNodeEmbeddingsoptimized.cls

@@ -0,0 +1,12 @@
+Class Graph.KG.kgNodeEmbeddingsoptimized Extends %Persistent [ DdlAllowed, Final, SqlTableName = kg_NodeEmbeddings_optimized ]
+{
+
+Property id As %String(COLLATION = "EXACT", MAXLEN = 256) [ Required, SqlColumnNumber = 2, SqlFieldName = id ];
+
+Property emb As %Library.Vector(DATATYPE = "double", LEN = 384) [ SqlColumnNumber = 3, SqlFieldName = emb ];
+
+Property metadata As %Library.DynamicObject [ SqlColumnNumber = 4, SqlFieldName = metadata ];
+
+Index IDKEY On id [ IdKey, PrimaryKey, Unique ];
+
+}

{iris_vector_graph-2.4.5 → iris_vector_graph-2.4.6}/iris_vector_graph/_engine/temporal.py

@@ -231,6 +231,43 @@ class TemporalMixin:
         )
         return json.loads(str(result))
 
+    def get_window_sources(
+        self,
+        predicate: str,
+        ts_start: int,
+        ts_end: int,
+    ) -> list[str]:
+        """Return distinct source node IDs with at least one edge for a
+        predicate over a time window, without materializing per-edge JSON.
+
+        Unlike ``get_edges_in_window`` (which returns every matching edge —
+        source, target, timestamp, weight — and requires the caller to
+        de-duplicate sources in Python), this walks ``^KG("tout")`` only down
+        to the source level per timestamp bucket and returns each matching
+        source once. Use this when you only need "which sources fired this
+        predicate in this window" (e.g. fleet/tenant enumeration) and would
+        otherwise be discarding target/timestamp/weight from a full window
+        scan.
+
+        Args:
+            predicate: Edge type to filter on. Empty string matches any
+                predicate (fastest — skips the per-source predicate check).
+            ts_start: Window start as Unix timestamp (inclusive).
+            ts_end: Window end as Unix timestamp (inclusive).
+
+        Returns:
+            list[str]: Distinct source node IDs, each returned once even if
+                they have multiple matching edges in the window.
+        """
+        result = self._iris_obj().classMethodValue(
+            "Graph.KG.TemporalIndex",
+            "QueryWindowSources",
+            predicate,
+            ts_start,
+            ts_end,
+        )
+        return json.loads(str(result))
+
     def get_distinct_count(
         self,
         source: str,

{iris_vector_graph-2.4.5 → iris_vector_graph-2.4.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "iris-vector-graph"
-version = "2.4.5"
+version = "2.4.6"
 description = "Native graph database layer for InterSystems IRIS — openCypher, hybrid vector search, temporal property graph, RDF export, SHACL validation"
 readme = "README.md"
 license = "MIT"

{iris_vector_graph-2.4.5 → iris_vector_graph-2.4.6}/tests/unit/test_temporal_edges.py

@@ -312,6 +312,33 @@ class TestTemporalPreAggUnit:
         result = engine.get_bucket_group_targets("Rtn:A:P:proc", "CALLED_BY", 0, 9999)
         assert result == []
 
+    # ── get_window_sources unit tests ─────────────────────────────────
+    def test_get_window_sources_returns_list(self):
+        """get_window_sources parses JSON array of distinct source IDs."""
+        engine, mock = self._make_engine()
+        mock.classMethodValue.return_value = '["QG:A:P:g1","QG:B:P:g2"]'
+        result = engine.get_window_sources("COST_ON", 0, 9999)
+        assert isinstance(result, list)
+        assert set(result) == {"QG:A:P:g1", "QG:B:P:g2"}
+
+    def test_get_window_sources_empty(self):
+        engine, mock = self._make_engine()
+        mock.classMethodValue.return_value = '[]'
+        result = engine.get_window_sources("COST_ON", 0, 9999)
+        assert result == []
+
+    def test_get_window_sources_calls_correct_classmethod(self):
+        """Verifies dispatch to QueryWindowSources with (predicate, tsStart, tsEnd)."""
+        engine, mock = self._make_engine()
+        mock.classMethodValue.return_value = '[]'
+        engine.get_window_sources("COST_ON", 100, 200)
+        args = mock.classMethodValue.call_args[0]
+        assert args[0] == "Graph.KG.TemporalIndex"
+        assert args[1] == "QueryWindowSources"
+        assert args[2] == "COST_ON"
+        assert args[3] == 100
+        assert args[4] == 200
+
     # ── ENH-3: docstring test ────────────────────────────────────────
     def test_get_bucket_groups_docstring(self):
         """get_bucket_groups has a docstring documenting all return keys."""
@@ -588,6 +615,55 @@ class TestTemporalAPIGapsE2E:
             f"{self.PREFIX}:ow", "CALLED_BY", now - 100, now + 100)
         assert targets == []
 
+    # ── get_window_sources ─────────────────────────────────────────────
+    def test_get_window_sources_returns_distinct_sources(self):
+        """get_window_sources returns each source once, filtered by predicate."""
+        now = int(time.time())
+        self.engine.bulk_create_edges_temporal([
+            {"s": f"{self.PREFIX}:src1", "p": "COST_ON",
+             "o": "Date:d1", "ts": now, "w": 1.0},
+            {"s": f"{self.PREFIX}:src1", "p": "COST_ON",
+             "o": "Date:d2", "ts": now + 1, "w": 2.0},
+            {"s": f"{self.PREFIX}:src2", "p": "COST_ON",
+             "o": "Date:d1", "ts": now + 2, "w": 1.0},
+            {"s": f"{self.PREFIX}:src3", "p": "OTHER_PRED",
+             "o": "Date:d1", "ts": now + 3, "w": 1.0},
+        ])
+        sources = self.engine.get_window_sources("COST_ON", now - 10, now + 100)
+        assert isinstance(sources, list)
+        assert set(sources) == {f"{self.PREFIX}:src1", f"{self.PREFIX}:src2"}
+        assert sources.count(f"{self.PREFIX}:src1") == 1
+
+    def test_get_window_sources_empty_predicate_matches_any(self):
+        """Empty predicate string matches sources regardless of edge type."""
+        now = int(time.time())
+        self.engine.bulk_create_edges_temporal([
+            {"s": f"{self.PREFIX}:anyp", "p": "SOME_PRED",
+             "o": "Date:d1", "ts": now, "w": 1.0},
+        ])
+        sources = self.engine.get_window_sources("", now - 10, now + 100)
+        assert f"{self.PREFIX}:anyp" in sources
+
+    def test_get_window_sources_empty_when_outside_window(self):
+        """No sources returned when all edges are outside the time window."""
+        now = int(time.time())
+        self.engine.bulk_create_edges_temporal([
+            {"s": f"{self.PREFIX}:ows", "p": "COST_ON",
+             "o": "Date:d1", "ts": now + 10000, "w": 1.0},
+        ])
+        sources = self.engine.get_window_sources("COST_ON", now - 100, now + 100)
+        assert sources == []
+
+    def test_get_window_sources_respects_predicate_filter(self):
+        """A source with only non-matching-predicate edges is excluded."""
+        now = int(time.time())
+        self.engine.bulk_create_edges_temporal([
+            {"s": f"{self.PREFIX}:wrongpred", "p": "LATENCY_ON",
+             "o": "Date:d1", "ts": now, "w": 1.0},
+        ])
+        sources = self.engine.get_window_sources("COST_ON", now - 10, now + 100)
+        assert f"{self.PREFIX}:wrongpred" not in sources
+
     def test_purge_before_removes_old_edges(self):
         now = int(time.time())
         old_ts = now - 7200