rond-with-you/internal/web/city_agg_test.go

package web

import (
	"encoding/json"
	"net/http/httptest"
	"testing"
)

// 地图上的城市圈气泡（/api/cities）必须「先按地点聚合、再按城市汇总」。
//
// 旧写法把聚合打在 places ⋈ visits 上，于是每个地点的 visit_count / dwell_minutes
// 会被按它的到访条数重复累加，count(*) 又被当成「地点数」——实测最大那个城市显示
// 「到访 72661 次 · 568 个地点」，真值是 568 次 / 86 个地点。这里按同一套口径手写一遍对账。
func TestCityAggCountsOncePerPlace(t *testing.T) {
	f := newMovementFixture(t)
	ctx, pg, dsID := f.ctx, f.pg, f.dsID

	rec := httptest.NewRecorder()
	f.srv.apiCityAgg(rec, httptest.NewRequest("GET", "/api/cities", nil).WithContext(ctx))
	if rec.Code != 200 {
		t.Fatalf("取城市聚合失败: code=%d body=%s", rec.Code, rec.Body.String())
	}
	var payload struct {
		Cities []struct {
			City   string `json:"city"`
			Visits int    `json:"visits"`
			Places int    `json:"places"`
			Dwell  int64  `json:"dwell"`
			First  string `json:"first,omitempty"`
			Last   string `json:"last,omitempty"`
		} `json:"cities"`
	}
	if err := json.Unmarshal(rec.Body.Bytes(), &payload); err != nil {
		t.Fatalf("解析城市聚合失败: %v", err)
	}
	if len(payload.Cities) == 0 {
		t.Skip("夹具没有带城市的到访")
	}

	// 手写一份同一口径的对照（判据独立写，不复用 stats.SoloVisitSQL）：
	// 不含已并入 / 子段，也不含「家 / 工作」地点——与 apiCityAgg 保持一致。
	rows, err := pg.QueryContext(ctx, `SELECT p.city, count(*), count(DISTINCT p.id),
			COALESCE(sum(LEAST(COALESCE(v.duration_min,0),10080)),0),
			to_char(min(v.arrival), 'YYYY-MM-DD'), to_char(max(v.arrival), 'YYYY-MM-DD')
		FROM places p JOIN visits v ON v.place_id = p.id AND v.dataset_id = p.dataset_id
		WHERE p.dataset_id = $1 AND p.city IS NOT NULL AND p.city <> '' AND p.lat IS NOT NULL
		  AND NOT EXISTS (SELECT 1 FROM entity_raw er
			WHERE er.dataset_id=v.dataset_id AND er.entity='ZVISIT' AND er.src_pk=v.src_pk
			  AND (er.raw->>'ZMERGEDTO' IS NOT NULL OR er.raw->>'ZPARENT' IS NOT NULL))
		  AND NOT EXISTS (SELECT 1 FROM visits vh WHERE vh.place_id = p.id
			AND (vh.is_home OR vh.is_work))
		GROUP BY p.city`, dsID)
	if err != nil {
		t.Fatal(err)
	}
	defer rows.Close()
	type want struct {
		visits, places int
		dwell          int64
		first, last    string
	}
	truth := map[string]want{}
	for rows.Next() {
		var city string
		var w want
		if err := rows.Scan(&city, &w.visits, &w.places, &w.dwell, &w.first, &w.last); err != nil {
			t.Fatal(err)
		}
		truth[city] = w
	}
	if err := rows.Err(); err != nil {
		t.Fatal(err)
	}

	for _, c := range payload.Cities {
		w, ok := truth[c.City]
		if !ok {
			t.Errorf("城市「%s」不该出现（手写 SQL 里没有）", c.City)
			continue
		}
		if c.Visits != w.visits || c.Places != w.places || c.Dwell != w.dwell {
			t.Errorf("城市「%s」聚合对不上：到访 %d/%d、地点 %d/%d、停留 %d/%d",
				c.City, c.Visits, w.visits, c.Places, w.places, c.Dwell, w.dwell)
		}
		// 首末到访必须按「可见到访的到达时间」算，不能拿 places 的全量列值
		if c.First != w.first || c.Last != w.last {
			t.Errorf("城市「%s」首末到访对不上：%s~%s，应为 %s~%s（按可见到访算）",
				c.City, c.First, c.Last, w.first, w.last)
		}
	}
	// 保底的不变量：地点数不可能多于到访数（旧写法里这两个数就换位了）
	for _, c := range payload.Cities {
		if c.Places > c.Visits {
			t.Errorf("城市「%s」的地点数 %d 多于到访数 %d，说明又按到访条数重复累加了", c.City, c.Places, c.Visits)
		}
	}

	// 把某个地点**最晚的那次到访**标成「已并入」：该城市的最近到访必须跟着退到更早那条。
	// 这一条专治「偷懒直接取 places.last_visit_at」——那是全量列值，不会跟着退。
	// （光比对相同口径还不够：被并入的行大多落在父段区间内，两种算法在夹具数据上结果一样，
	//  不做这一步就是一条假绿的断言。）
	var placeID int64
	var latestVisit int
	if err := pg.QueryRowContext(ctx, `SELECT p.id, v.src_pk
		FROM visits v JOIN places p ON p.id = v.place_id
		WHERE v.dataset_id=$1 AND p.city IS NOT NULL AND p.city <> '' AND p.lat IS NOT NULL
		  AND NOT EXISTS (SELECT 1 FROM visits vh WHERE vh.place_id = p.id AND (vh.is_home OR vh.is_work))
		  AND (SELECT count(*) FROM visits v2 WHERE v2.place_id = p.id) >= 2
		ORDER BY v.arrival DESC LIMIT 1`, dsID).Scan(&placeID, &latestVisit); err != nil {
		t.Skipf("挑不出「同一地点有多条到访」的城市：%v", err)
	}
	var cityName string
	if err := pg.QueryRowContext(ctx, `SELECT city FROM places WHERE id=$1`, placeID).Scan(&cityName); err != nil {
		t.Fatal(err)
	}
	lastBefore := ""
	lastAfterWant := ""
	for _, c := range payload.Cities {
		if c.City == cityName {
			lastBefore = c.Last
		}
	}
	// 真值（按城市取，标记前后各来一次）：与上面那条对账同一套口径，只是限定到某个城市
	truthLast := func() string {
		var s string
		if err := pg.QueryRowContext(ctx, `SELECT to_char(max(v.arrival),'YYYY-MM-DD')
			FROM places p JOIN visits v ON v.place_id = p.id AND v.dataset_id = p.dataset_id
			WHERE p.dataset_id = $1 AND p.city = $2 AND p.city <> '' AND p.lat IS NOT NULL
			  AND NOT EXISTS (SELECT 1 FROM entity_raw er
				WHERE er.dataset_id=v.dataset_id AND er.entity='ZVISIT' AND er.src_pk=v.src_pk
				  AND (er.raw->>'ZMERGEDTO' IS NOT NULL OR er.raw->>'ZPARENT' IS NOT NULL))
			  AND NOT EXISTS (SELECT 1 FROM visits vh WHERE vh.place_id = p.id
				AND (vh.is_home OR vh.is_work))`, dsID, cityName).Scan(&s); err != nil {
			t.Fatal(err)
		}
		return s
	}
	lastBeforeWant := truthLast()
	if lastBeforeWant != lastBefore {
		t.Fatalf("标记前该城市的最近到访就对不上（%s vs %s），用例前提不成立", lastBefore, lastBeforeWant)
	}
	if _, err := pg.ExecContext(ctx, `UPDATE entity_raw
		SET raw = jsonb_set(raw, '{ZMERGEDTO}', to_jsonb($3::int), true)
		WHERE dataset_id=$1 AND entity='ZVISIT' AND src_pk=$2`, dsID, latestVisit, latestVisit); err != nil {
		t.Fatal(err)
	}
	lastAfterWant = truthLast()
	if lastAfterWant == lastBeforeWant {
		t.Skipf("城市「%s」里最晚那次不是唯一的最晚记录，换一个再测", cityName)
	}

	rec2 := httptest.NewRecorder()
	f.srv.apiCityAgg(rec2, httptest.NewRequest("GET", "/api/cities", nil).WithContext(ctx))
	if rec2.Code != 200 {
		t.Fatalf("取城市聚合失败: code=%d", rec2.Code)
	}
	var after struct {
		Cities []struct {
			City string `json:"city"`
			Last string `json:"last,omitempty"`
		} `json:"cities"`
	}
	if err := json.Unmarshal(rec2.Body.Bytes(), &after); err != nil {
		t.Fatalf("解析城市聚合失败: %v", err)
	}
	gotLast := ""
	for _, c := range after.Cities {
		if c.City == cityName {
			gotLast = c.Last
		}
	}
	if gotLast == lastBeforeWant {
		t.Errorf("把 #%d 标成已并入后，城市「%s」的最近到访仍是 %s（应退到 %s）—— 说明取的是 places 的全量列值而不是可见到访",
			latestVisit, cityName, gotLast, lastAfterWant)
	}
	if gotLast != lastAfterWant {
		t.Errorf("城市「%s」的最近到访应为 %s，实际 %s", cityName, lastAfterWant, gotLast)
	}
}