feat(product-analytics): most used property values for the agent (#26985

)
PostHog · Dec 18, 2024 · 3e95fdc · 3e95fdc
1 parent 2f9e896
commit 3e95fdc
Show file tree

Hide file tree

Showing 5 changed files with 253 additions and 20 deletions.
diff --git a/frontend/src/queries/schema.json b/frontend/src/queries/schema.json
@@ -5630,6 +5630,12 @@
                     "$ref": "#/definitions/HogQLQueryModifiers",
                     "description": "Modifiers used when performing the query"
                 },
+                "properties": {
+                    "items": {
+                        "type": "string"
+                    },
+                    "type": "array"
+                },
                 "response": {
                     "$ref": "#/definitions/EventTaxonomyQueryResponse"
                 }

diff --git a/frontend/src/queries/schema.ts b/frontend/src/queries/schema.ts
@@ -2420,6 +2420,7 @@ export type EventTaxonomyResponse = EventTaxonomyItem[]
 export interface EventTaxonomyQuery extends DataNode<EventTaxonomyQueryResponse> {
     kind: NodeKind.EventTaxonomyQuery
     event: string
+    properties?: string[]
 }
 
 export type EventTaxonomyQueryResponse = AnalyticsQueryResponseBase<EventTaxonomyResponse>

diff --git a/posthog/hogql_queries/ai/event_taxonomy_query_runner.py b/posthog/hogql_queries/ai/event_taxonomy_query_runner.py
@@ -50,8 +50,9 @@ def calculate(self):
         )
 
     def to_query(self) -> ast.SelectQuery | ast.SelectSetQuery:
-        query = parse_select(
-            """
+        if not self.query.properties:
+            return parse_select(
+                """
                 SELECT
                     key,
                     -- Pick five latest distinct sample values.
@@ -64,10 +65,23 @@ def to_query(self) -> ast.SelectQuery | ast.SelectSetQuery:
                 ORDER BY total_count DESC
                 LIMIT 500
             """,
-            placeholders={"from_query": self._get_subquery(), "filter": self._get_omit_filter()},
-        )
+                placeholders={"from_query": self._get_subquery(), "filter": self._get_omit_filter()},
+            )
 
-        return query
+        return parse_select(
+            """
+                SELECT
+                    key,
+                    arraySlice(arrayDistinct(groupArray(value)), 1, 5) AS values,
+                    count(DISTINCT value) AS total_count
+                FROM {from_query}
+                GROUP BY key
+                LIMIT 500
+            """,
+            placeholders={
+                "from_query": self._get_subquery(),
+            },
+        )
 
     def _get_omit_filter(self):
         """
@@ -107,21 +121,72 @@ def _get_omit_filter(self):
 
     def _get_subquery_filter(self) -> ast.Expr:
         date_filter = parse_expr("timestamp >= now() - INTERVAL 30 DAY")
-        filter_expr = ast.And(
-            exprs=[
-                date_filter,
-                ast.CompareOperation(
-                    left=ast.Field(chain=["event"]),
-                    right=ast.Constant(value=self.query.event),
-                    op=ast.CompareOperationOp.Eq,
-                ),
-            ]
-        )
-        return filter_expr
+        filter_expr: list[ast.Expr] = [
+            date_filter,
+            ast.CompareOperation(
+                left=ast.Field(chain=["event"]),
+                right=ast.Constant(value=self.query.event),
+                op=ast.CompareOperationOp.Eq,
+            ),
+        ]
+
+        if self.query.properties:
+            filter_expr.append(
+                ast.Or(
+                    exprs=[
+                        ast.CompareOperation(
+                            left=ast.Field(chain=["properties", prop]),
+                            op=ast.CompareOperationOp.NotEq,
+                            right=ast.Constant(value=""),
+                        )
+                        for prop in self.query.properties
+                    ]
+                )
+            )
+
+        return ast.And(exprs=filter_expr)
 
     def _get_subquery(self) -> ast.SelectQuery:
-        query = parse_select(
-            """
+        if self.query.properties:
+            query = parse_select(
+                """
+                    SELECT
+                        key,
+                        value,
+                        count() as count
+                    FROM (
+                        SELECT
+                            {props} as kv
+                        FROM
+                            events
+                        WHERE {filter}
+                    )
+                    ARRAY JOIN kv.1 AS key, kv.2 AS value
+                    WHERE value != ''
+                    GROUP BY key, value
+                    ORDER BY count DESC
+                """,
+                placeholders={
+                    "props": ast.Array(
+                        exprs=[
+                            ast.Tuple(
+                                exprs=[
+                                    ast.Constant(value=prop),
+                                    ast.Call(
+                                        name="JSONExtractString",
+                                        args=[ast.Field(chain=["properties"]), ast.Constant(value=prop)],
+                                    ),
+                                ]
+                            )
+                            for prop in self.query.properties
+                        ]
+                    ),
+                    "filter": self._get_subquery_filter(),
+                },
+            )
+        else:
+            query = parse_select(
+                """
                 SELECT
                     JSONExtractKeysAndValues(properties, 'String') as kv
                 FROM
@@ -130,7 +195,7 @@ def _get_subquery(self) -> ast.SelectQuery:
                 ORDER BY timestamp desc
                 LIMIT 100
             """,
-            placeholders={"filter": self._get_subquery_filter()},
-        )
+                placeholders={"filter": self._get_subquery_filter()},
+            )
 
         return cast(ast.SelectQuery, query)
diff --git a/posthog/hogql_queries/ai/test/test_event_taxonomy_query_runner.py b/posthog/hogql_queries/ai/test/test_event_taxonomy_query_runner.py
@@ -251,3 +251,163 @@ def test_limit(self):
 
         response = EventTaxonomyQueryRunner(team=self.team, query=EventTaxonomyQuery(event="event1")).calculate()
         self.assertEqual(len(response.results), 500)
+
+    def test_property_taxonomy_returns_unique_values_for_specified_property(self):
+        _create_person(
+            distinct_ids=["person1"],
+            properties={"email": "[email protected]"},
+            team=self.team,
+        )
+        _create_person(
+            distinct_ids=["person2"],
+            properties={"email": "[email protected]"},
+            team=self.team,
+        )
+
+        _create_event(
+            event="event1",
+            distinct_id="person1",
+            properties={"$host": "us.posthog.com"},
+            team=self.team,
+        )
+
+        for _ in range(10):
+            _create_event(
+                event="event1",
+                distinct_id="person1",
+                properties={"$host": "posthog.com"},
+                team=self.team,
+            )
+
+        for _ in range(3):
+            _create_event(
+                event="event1",
+                distinct_id="person2",
+                properties={"$host": "eu.posthog.com"},
+                team=self.team,
+            )
+
+        response = EventTaxonomyQueryRunner(
+            team=self.team, query=EventTaxonomyQuery(event="event1", properties=["$host"])
+        ).calculate()
+        self.assertEqual(len(response.results), 1)
+        self.assertEqual(response.results[0].property, "$host")
+        self.assertEqual(response.results[0].sample_values, ["posthog.com", "eu.posthog.com", "us.posthog.com"])
+        self.assertEqual(response.results[0].sample_count, 3)
+
+    def test_property_taxonomy_filters_events_by_event_name(self):
+        _create_person(
+            distinct_ids=["person1"],
+            properties={"email": "[email protected]"},
+            team=self.team,
+        )
+        _create_person(
+            distinct_ids=["person2"],
+            properties={"email": "[email protected]"},
+            team=self.team,
+        )
+
+        _create_event(
+            event="event1",
+            distinct_id="person1",
+            properties={"$host": "us.posthog.com", "$browser": "Chrome"},
+            team=self.team,
+        )
+
+        for _ in range(10):
+            _create_event(
+                event="event2",
+                distinct_id="person1",
+                properties={"$host": "posthog.com", "prop": 10},
+                team=self.team,
+            )
+
+        for _ in range(3):
+            _create_event(
+                event="event1",
+                distinct_id="person2",
+                team=self.team,
+            )
+
+        response = EventTaxonomyQueryRunner(
+            team=self.team, query=EventTaxonomyQuery(event="event1", properties=["$host"])
+        ).calculate()
+        self.assertEqual(len(response.results), 1)
+        self.assertEqual(response.results[0].property, "$host")
+        self.assertEqual(response.results[0].sample_values, ["us.posthog.com"])
+        self.assertEqual(response.results[0].sample_count, 1)
+
+    def test_property_taxonomy_handles_multiple_properties_in_query(self):
+        _create_person(
+            distinct_ids=["person1"],
+            properties={"email": "[email protected]"},
+            team=self.team,
+        )
+        _create_person(
+            distinct_ids=["person2"],
+            properties={"email": "[email protected]"},
+            team=self.team,
+        )
+
+        _create_event(
+            event="event1",
+            distinct_id="person1",
+            properties={"$host": "us.posthog.com", "$browser": "Chrome"},
+            team=self.team,
+        )
+
+        for _ in range(5):
+            _create_event(
+                event="event1",
+                distinct_id="person1",
+                properties={"$host": "posthog.com", "prop": 10},
+                team=self.team,
+            )
+
+        for _ in range(3):
+            _create_event(
+                event="event1",
+                distinct_id="person2",
+                team=self.team,
+            )
+
+        response = EventTaxonomyQueryRunner(
+            team=self.team, query=EventTaxonomyQuery(event="event1", properties=["$host", "prop"])
+        ).calculate()
+        self.assertEqual(len(response.results), 2)
+        self.assertEqual(response.results[0].property, "prop")
+        self.assertEqual(response.results[0].sample_values, ["10"])
+        self.assertEqual(response.results[0].sample_count, 1)
+        self.assertEqual(response.results[1].property, "$host")
+        self.assertEqual(response.results[1].sample_values, ["posthog.com", "us.posthog.com"])
+        self.assertEqual(response.results[1].sample_count, 2)
+
+    def test_property_taxonomy_includes_events_with_partial_property_matches(self):
+        _create_person(
+            distinct_ids=["person1"],
+            properties={"email": "[email protected]"},
+            team=self.team,
+        )
+        _create_event(
+            event="event1",
+            distinct_id="person1",
+            properties={"$host": "us.posthog.com"},
+            team=self.team,
+        )
+        _create_event(
+            event="event1",
+            distinct_id="person2",
+            properties={"prop": 10},
+            team=self.team,
+        )
+
+        response = EventTaxonomyQueryRunner(
+            team=self.team, query=EventTaxonomyQuery(event="event1", properties=["$host", "prop"])
+        ).calculate()
+        self.assertEqual(len(response.results), 2)
+        self.assertEqual(response.results[0].property, "prop")
+        self.assertEqual(response.results[0].sample_values, ["10"])
+        self.assertEqual(response.results[0].sample_count, 1)
+        self.assertEqual(response.results[1].property, "$host")
+        self.assertEqual(response.results[1].sample_values, ["us.posthog.com"])
+        self.assertEqual(response.results[1].sample_count, 1)
diff --git a/posthog/schema.py b/posthog/schema.py
@@ -5582,6 +5582,7 @@ class EventTaxonomyQuery(BaseModel):
     modifiers: Optional[HogQLQueryModifiers] = Field(
         default=None, description="Modifiers used when performing the query"
     )
+    properties: Optional[list[str]] = None
     response: Optional[EventTaxonomyQueryResponse] = None