[ecommerce] Add a new simple metric aggr: geo_bounds (#1360)

Co-authored-by: trzysiek <[email protected]>

Author: trzysiek

docs/public/docs/limitations.md

@@ -38,7 +38,7 @@ Currently supported:
   including: `boolean`, `match`, `match phrase`, `multi-match`, `query string`, `nested`, `match all`, `exists`, `prefix`, `range`, `term`, `terms`, `wildcard`
 - most popular [Aggregations](https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations.html),
   including: `avg`, `cardinality`, `max`, `min`, `percentile ranks`, `percentiles`, `stats`, `sum`, `top hits`, `top metrics`, `value count`,
-  `date histogram`, `date range`, `filter`, `filters`, `histogram`, `range`, `singificant terms`, `terms`, `ip prefix`, `ip range`, `geohash_grid`
+  `date histogram`, `date range`, `filter`, `filters`, `histogram`, `range`, `singificant terms`, `terms`, `ip prefix`, `ip range`, `geo_bounds`, `geohash_grid`
 
 Which as a result allows you to run Kibana/OSD queries and dashboards on data residing in ClickHouse/Hydrolix.
 

platform/model/README.md

@@ -18,7 +18,7 @@ More info: https://www.elastic.co/guide/en/elasticsearch/reference/current/searc
  Boxplot                   |          :x:           | Composite                    | :white_check_mark: | Bucket selector        |        :x:         |
  Cardinality               |   :white_check_mark:   | Date histogram               | :white_check_mark: | Bucket sort            |        :x:         |
  Extended stats            | :white_check_mark:[^1] | Date range                   | :white_check_mark: | Change point           |        :x:         |
- Geo-bounds                |          :x:           | Diversified sampler          |        :x:         | Cumulative cardinality |        :x:         |
+ Geo-bounds                |   :white_check_mark:   | Diversified sampler          |        :x:         | Cumulative cardinality |        :x:         |
  Geo-centroid              |          :x:           | Filter                       | :white_check_mark: | Cumulative sum         | :white_check_mark: |
  Geo-line                  |          :x:           | Filters                      | :white_check_mark: | Derivative             | :white_check_mark: |
  Cartesian-bounds          |          :x:           | Frequent item sets           |        :x:         | Extended stats bucket  |        :x:         |

platform/model/metrics_aggregations/geo_bounds.go

@@ -0,0 +1,49 @@
+// Copyright Quesma, licensed under the Elastic License 2.0.
+// SPDX-License-Identifier: Elastic-2.0
+package metrics_aggregations
+
+import (
+	"context"
+	"github.com/QuesmaOrg/quesma/platform/logger"
+	"github.com/QuesmaOrg/quesma/platform/model"
+)
+
+type GeoBounds struct {
+	ctx context.Context
+}
+
+func NewGeoBounds(ctx context.Context) GeoBounds {
+	return GeoBounds{ctx: ctx}
+}
+
+func (query GeoBounds) AggregationType() model.AggregationType {
+	return model.MetricsAggregation
+}
+
+func (query GeoBounds) TranslateSqlResponseToJson(rows []model.QueryResultRow) model.JsonMap {
+	if len(rows) == 0 {
+		logger.ErrorWithCtx(query.ctx).Msg("GeoBounds: expected at least one row in the result")
+		return model.JsonMap{}
+	}
+	if len(rows[0].Cols) < 4 {
+		logger.ErrorWithCtx(query.ctx).Msgf("GeoBounds: expected at least 4 columns in the result, got: %v", rows[0].Cols)
+		return model.JsonMap{}
+	}
+
+	return model.JsonMap{
+		"bounds": model.JsonMap{
+			"top_left": model.JsonMap{
+				"lon": rows[0].Cols[0].Value,
+				"lat": rows[0].Cols[1].Value,
+			},
+			"bottom_right": model.JsonMap{
+				"lat": rows[0].Cols[2].Value,
+				"lon": rows[0].Cols[3].Value,
+			},
+		},
+	}
+}
+
+func (query GeoBounds) String() string {
+	return "geo_bounds"
+}

platform/parsers/elastic_query_dsl/aggregation_parser.go

@@ -158,6 +158,13 @@ func (cw *ClickhouseQueryTranslator) tryMetricsAggregation(queryMap QueryMap) (m
 		}, true
 	}
 
+	if geoBounds, exists := queryMap["geo_bounds"]; exists {
+		return metricsAggregation{
+			AggrType: "geo_bounds",
+			Fields:   []model.Expr{cw.parseFieldField(geoBounds, "geo_bounds")},
+		}, true
+	}
+
 	return metricsAggregation{}, false
 }
 

platform/parsers/elastic_query_dsl/pancake_aggregation_parser_metrics.go

@@ -116,6 +116,21 @@ func generateMetricSelectedColumns(ctx context.Context, metricsAggr metricsAggre
 			result = append(result, model.NewFunction("avgOrNull", lonColumn))
 			result = append(result, model.NewCountFunc())
 		}
+	case "geo_bounds":
+		firstExpr := getFirstExpression()
+		result = make([]model.Expr, 0, 3)
+		if col, ok := firstExpr.(model.ColumnRef); ok {
+			// TODO this is internalPropertyName and should be taken from schema
+			// instead of using util.FieldToColumnEncoder and doing encoding in-place
+			colName := util.FieldToColumnEncoder(col.ColumnName)
+			// TODO we have create columns according to the schema
+			latColumn := model.NewGeoLat(colName)
+			lonColumn := model.NewGeoLon(colName)
+			result = append(result, model.NewFunction("minOrNull", lonColumn))
+			result = append(result, model.NewFunction("argMinOrNull", latColumn, lonColumn))
+			result = append(result, model.NewFunction("minOrNull", latColumn))
+			result = append(result, model.NewFunction("argMinOrNull", lonColumn, latColumn))
+		}
 	default:
 		logger.WarnWithCtx(ctx).Msgf("unknown metrics aggregation: %s", metricsAggr.AggrType)
 		return nil, fmt.Errorf("unknown metrics aggregation %s", metricsAggr.AggrType)
@@ -151,6 +166,9 @@ func (cw *ClickhouseQueryTranslator) generateMetricsType(metricsAggr metricsAggr
 		return metrics_aggregations.NewPercentileRanks(cw.Ctx, metricsAggr.CutValues, metricsAggr.Keyed)
 	case "geo_centroid":
 		return metrics_aggregations.NewGeoCentroid(cw.Ctx)
+	case "geo_bounds":
+		return metrics_aggregations.NewGeoBounds(cw.Ctx)
 	}
+
 	return nil
 }

platform/testdata/kibana_sample_data_ecommerce.go

@@ -2819,4 +2819,118 @@ var KibanaSampleDataEcommerce = []AggregationTestCase{
 			ORDER BY "aggr__join__count" DESC, "aggr__join__key_0" ASC
 			LIMIT 65536`,
 	},
+	{ // [15]
+		TestName: "weird",
+		QueryRequestJson: `
+		{
+			"_source": {
+				"excludes": []
+			},
+			"aggs": {
+				"fitToBounds": {
+					"geo_bounds": {
+						"field": "OriginLocation"
+					}
+				}
+			},
+			"fields": [
+				{
+					"field": "@timestamp",
+					"format": "date_time"
+				},
+				{
+					"field": "timestamp",
+					"format": "date_time"
+				}
+			],
+			"query": {
+				"bool": {
+					"filter": [
+						{
+							"range": {
+								"timestamp": {
+									"format": "strict_date_optional_time",
+									"gte": "2025-03-02T14:16:32.069Z",
+									"lte": "2025-03-09T14:16:32.069Z"
+								}
+							}
+						}
+					],
+					"must": [],
+					"must_not": [],
+					"should": []
+				}
+			},
+			"runtime_mappings": {
+				"hour_of_day": {
+					"script": {
+						"source": "emit(doc['timestamp'].value.getHour());"
+					},
+					"type": "long"
+				}
+			},
+			"script_fields": {},
+			"size": 0,
+			"stored_fields": [
+				"*"
+			],
+			"track_total_hits": false
+		}`,
+		ExpectedResponse: `
+		{
+			"completion_time_in_millis": 1740838900680,
+			"expiration_time_in_millis": 1740838960672,
+			"id": "FnBCYVZTQWtUVEgtVGNiUzFabnFqbVEdUEQ3d19oVkxSMEthNU02NjIwRGpkZzo3MTY5NTM=",
+			"is_partial": false,
+			"is_running": false,
+			"response": {
+				"_shards": {
+					"failed": 0,
+					"skipped": 0,
+					"successful": 1,
+					"total": 1
+				},
+				"aggregations": {
+					"fitToBounds": {
+						"bounds": {
+							"top_left": {
+								"lat": 68.15180202014744,
+								"lon": -122.59799961000681
+							},
+							"bottom_right": {
+								"lat": -37.67330203671008,
+								"lon": 153.11700434423983
+							}
+						}
+					}
+				},
+				"hits": {
+					"hits": [],
+					"max_score": null
+				},
+				"timed_out": false,
+				"took": 8
+			},
+			"start_time_in_millis": 1740838900672
+		}`,
+		ExpectedPancakeResults: []model.QueryResultRow{
+			{Cols: []model.QueryResultCol{
+				model.NewQueryResultCol("metric__fitToBounds_col_0", -122.59799961000681),
+				model.NewQueryResultCol("metric__fitToBounds_col_1", 68.15180202014744),
+				model.NewQueryResultCol("metric__fitToBounds_col_2", -37.67330203671008),
+				model.NewQueryResultCol("metric__fitToBounds_col_3", 153.11700434423983),
+			}},
+		},
+		ExpectedPancakeSQL: `
+			SELECT minOrNull(__quesma_geo_lon("originlocation")) AS
+			  "metric__fitToBounds_col_0",
+			  argMinOrNull(__quesma_geo_lat("originlocation"), __quesma_geo_lon(
+			  "originlocation")) AS "metric__fitToBounds_col_1",
+			  minOrNull(__quesma_geo_lat("originlocation")) AS "metric__fitToBounds_col_2",
+			  argMinOrNull(__quesma_geo_lon("originlocation"), __quesma_geo_lat(
+			  "originlocation")) AS "metric__fitToBounds_col_3"
+			FROM __quesma_table_name
+			WHERE ("timestamp">=fromUnixTimestamp64Milli(1740924992069) AND "timestamp"<=
+			  fromUnixTimestamp64Milli(1741529792069))`,
+	},
 }

platform/testdata/unsupported_requests.go

@@ -361,20 +361,6 @@ var UnsupportedQueriesTests = []UnsupportedQueryTestCase{
 			}
 		}`,
 	},
-	{ // [18]
-		TestName:  "metrics aggregation: geo_bounds",
-		QueryType: "geo_bounds",
-		QueryRequestJson: `
-		{
-			"aggs": {
-				"viewport": {
-					"geo_bounds": {
-						"field": "geometry"
-					}
-				}
-			}
-		}`,
-	},
 	{ // [19]
 		TestName:  "metrics aggregation: geo_line",
 		QueryType: "geo_line",