Align daily active metric with active log

internal/storage/active_reconcile.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"database/sql"
 	"log/slog"
+	"strings"
 	"time"
 
 	"go.mongodb.org/mongo-driver/v2/bson"
@@ -16,9 +17,16 @@ import (
 const activeLogCollection = "user_daily_active_log"
 
 type activeLogRow struct {
-	UserID    int64  `bson:"userId"`
+	UserID              int64  `bson:"userId"`
-	ActiveDay string `bson:"activeDate"`
+	ActiveDay           string `bson:"activeDate"`
-	SysOrigin string `bson:"sysOrigin"`
+	SysOrigin           string `bson:"sysOrigin"`
+	RegisterCountryCode string `bson:"registerCountryCode"`
+}
+
+type activeLogEvent struct {
+	UserID    int64
+	SysOrigin string
+	Country   dashboard.Country
 }
 
 type activeDayStats struct {
@@ -107,7 +115,7 @@ func (r *MySQLRepository) RebuildActiveRetentionRange(ctx context.Context, db *m
 	})
 }
 
-func (r *MySQLRepository) loadActiveLogs(ctx context.Context, db *mongo.Database, start, end time.Time) (map[string]map[int64]struct{}, error) {
+func (r *MySQLRepository) loadActiveLogs(ctx context.Context, db *mongo.Database, start, end time.Time) (map[string]map[int64]activeLogEvent, error) {
 	filter := bson.D{
 		{Key: "activeDate", Value: bson.D{{Key: "$gte", Value: start.Format("2006-01-02")}, {Key: "$lt", Value: end.Format("2006-01-02")}}},
 		{Key: "sysOrigin", Value: bson.D{{Key: "$in", Value: r.cfg.Dashboard.SysOrigins}}},
@@ -116,6 +124,7 @@ func (r *MySQLRepository) loadActiveLogs(ctx context.Context, db *mongo.Database
 		{Key: "userId", Value: 1},
 		{Key: "activeDate", Value: 1},
 		{Key: "sysOrigin", Value: 1},
+		{Key: "registerCountryCode", Value: 1},
 	})
 	if r.cfg.Active.BatchLimit > 0 {
 		findOptions.SetLimit(int64(r.cfg.Active.BatchLimit))
@@ -126,7 +135,7 @@ func (r *MySQLRepository) loadActiveLogs(ctx context.Context, db *mongo.Database
 	}
 	defer cursor.Close(ctx)
 
-	events := make(map[string]map[int64]struct{})
+	events := make(map[string]map[int64]activeLogEvent)
 	for cursor.Next(ctx) {
 		var row activeLogRow
 		if err := cursor.Decode(&row); err != nil {
@@ -137,15 +146,27 @@ func (r *MySQLRepository) loadActiveLogs(ctx context.Context, db *mongo.Database
 		}
 		users := events[row.ActiveDay]
 		if users == nil {
-			users = make(map[int64]struct{})
+			users = make(map[int64]activeLogEvent)
 			events[row.ActiveDay] = users
 		}
-		users[row.UserID] = struct{}{}
+		users[row.UserID] = activeLogEvent{
+			UserID:    row.UserID,
+			SysOrigin: strings.TrimSpace(row.SysOrigin),
+			Country:   r.activeLogCountry(row.RegisterCountryCode),
+		}
 	}
 	return events, cursor.Err()
 }
 
-func collectActiveUserIDs(events map[string]map[int64]struct{}) []int64 {
+func (r *MySQLRepository) activeLogCountry(registerCountryCode string) dashboard.Country {
+	code := strings.TrimSpace(registerCountryCode)
+	if code == "" {
+		code = r.cfg.Dashboard.UnknownCountryCode
+	}
+	return dashboard.Country{Code: code, Name: code}
+}
+
+func collectActiveUserIDs(events map[string]map[int64]activeLogEvent) []int64 {
 	seen := make(map[int64]struct{})
 	for _, users := range events {
 		for userID := range users {
@@ -159,6 +180,14 @@ func collectActiveUserIDs(events map[string]map[int64]struct{}) []int64 {
 	return out
 }
 
+func activeUserSet(events map[int64]activeLogEvent) map[int64]struct{} {
+	out := make(map[int64]struct{}, len(events))
+	for userID := range events {
+		out[userID] = struct{}{}
+	}
+	return out
+}
+
 func (r *MySQLRepository) loadUserCountriesByID(ctx context.Context, ids []int64) (map[int64]dashboard.UserCountry, error) {
 	result := make(map[int64]dashboard.UserCountry)
 	for start := 0; start < len(ids); start += bulkInsertChunkSize {
@@ -198,32 +227,43 @@ WHERE user_id IN `+inPlaceholders(end-start), args...)
 }
 
 func (r *MySQLRepository) replaceActiveRetentionDay(ctx context.Context, tx *sql.Tx, day time.Time,
-	events map[string]map[int64]struct{}, users map[int64]dashboard.UserCountry) error {
+	events map[string]map[int64]activeLogEvent, users map[int64]dashboard.UserCountry) error {
 	dayKey := day.Format("2006-01-02")
 	statsByOrigin := make(map[string]map[string]*activeDayStats)
-	for userID := range events[dayKey] {
+	for userID, event := range events[dayKey] {
-		user, ok := users[userID]
+		sysOrigin := strings.TrimSpace(event.SysOrigin)
-		if !ok || user.IsDeleted || !r.allowedOrigin(user.SysOrigin) {
+		if !r.allowedOrigin(sysOrigin) {
 			continue
 		}
-		stats := statsByOrigin[user.SysOrigin]
+		country := event.Country
+		if country.Code == "" {
+			if user, ok := users[userID]; ok && user.Country.Code != "" {
+				country = user.Country
+			} else {
+				country = dashboard.Country{Code: r.cfg.Dashboard.UnknownCountryCode, Name: r.cfg.Dashboard.UnknownCountryCode}
+			}
+		}
+		if country.Name == "" {
+			country.Name = country.Code
+		}
+		stats := statsByOrigin[sysOrigin]
 		if stats == nil {
 			stats = make(map[string]*activeDayStats)
-			statsByOrigin[user.SysOrigin] = stats
+			statsByOrigin[sysOrigin] = stats
 		}
-		stat := stats[user.Country.Code]
+		stat := stats[country.Code]
 		if stat == nil {
 			stat = &activeDayStats{active: make(map[string]map[int64]struct{})}
-			stats[user.Country.Code] = stat
+			stats[country.Code] = stat
 		}
-		active := stat.active[user.Country.Name]
+		active := stat.active[country.Name]
 		if active == nil {
 			active = make(map[int64]struct{})
-			stat.active[user.Country.Name] = active
+			stat.active[country.Name] = active
 		}
 		active[userID] = struct{}{}
 	}
-	retentionByOrigin, err := r.retentionStatsForDay(ctx, day, events[dayKey], users)
+	retentionByOrigin, err := r.retentionStatsForDay(ctx, day, activeUserSet(events[dayKey]), users)
 	if err != nil {
 		return err
 	}