lingniu-vehicle-ingest/go/vehicle-gateway/cmd/stats-backfill/customer_hydrogen.go

package main

import (
	"context"
	"database/sql"
	"fmt"
	"log/slog"
	"math"
	"sort"
	"strings"
	"time"

	"lingniu-vehicle-ingest/go/vehicle-gateway/internal/envelope"
	"lingniu-vehicle-ingest/go/vehicle-gateway/internal/stats"
)

const customerHydrogenAlgorithm = "MODE_INTERVAL_V1"

type customerHydrogenDay struct {
	VIN, Date, Protocol, SourceKey             string
	HydrogenKM, ObservedKM, ClassifiedKM       float64
	Samples, KnownSamples, Intervals, Rejected int64
	First, Last                                time.Time
	PreviousKM                                 float64
}

func addCustomerHydrogenSample(days map[string]*customerHydrogenDay, s stats.MetricSample) {
	if s.Protocol != envelope.ProtocolGB32960 && s.Protocol != envelope.ProtocolYutongMQTT {
		return
	}
	if math.IsNaN(s.TotalMileageKM) || math.IsInf(s.TotalMileageKM, 0) || s.TotalMileageKM < 0 {
		return
	}
	key := s.VIN + "/" + s.StatDate + "/" + string(s.Protocol) + "/" + s.SourceKey
	d := days[key]
	if d == nil {
		d = &customerHydrogenDay{VIN: s.VIN, Date: s.StatDate, Protocol: string(s.Protocol), SourceKey: s.SourceKey, First: s.EventTime, Last: s.EventTime, PreviousKM: s.TotalMileageKM, Samples: 1}
		if s.PureHydrogenModeKnown {
			d.KnownSamples = 1
		}
		days[key] = d
		return
	}
	// Do not count duplicate or out-of-order reports as extra evidence.
	if !s.EventTime.After(d.Last) {
		return
	}
	elapsed := s.EventTime.Sub(d.Last)
	delta := s.TotalMileageKM - d.PreviousKM
	d.Samples++
	if s.PureHydrogenModeKnown {
		d.KnownSamples++
	}
	if delta < -0.2 || delta > math.Max(0.5, elapsed.Hours()*130) || elapsed > 15*time.Minute {
		d.Rejected++
	} else {
		delta = math.Max(0, delta)
		d.ObservedKM += delta
		d.Intervals++
		if s.PureHydrogenModeKnown {
			d.ClassifiedKM += delta
			if s.PureHydrogenActive {
				d.HydrogenKM += delta
			}
		}
	}
	d.Last = s.EventTime
	d.PreviousKM = s.TotalMileageKM
}

// This path reads raw work-mode evidence and only writes the dedicated customer
// monitor projection. It never resets/reprojects the existing GPS/daily tables.
func runCustomerHydrogenBackfill(ctx context.Context, mysqlDB, td *sql.DB, cfg config) error {
	return runCustomerHydrogenBackfillWithLease(ctx, mysqlDB, td, cfg, "")
}

func runCustomerHydrogenBackfillWithLease(ctx context.Context, mysqlDB, td *sql.DB, cfg config, owner string) error {
	if cfg.Reset {
		return fmt.Errorf("hydrogen-only backfill does not allow reset")
	}
	if cfg.Limit > 0 && !cfg.DryRun {
		return fmt.Errorf("limited scan cannot publish a daily result")
	}
	for _, p := range cfg.Protocols {
		if p != envelope.ProtocolGB32960 && p != envelope.ProtocolYutongMQTT {
			return fmt.Errorf("hydrogen-only backfill accepts GB32960,YUTONG_MQTT only")
		}
	}
	rows, err := queryRawFrames(ctx, td, cfg)
	if err != nil {
		return err
	}
	defer rows.Close()
	days := map[string]*customerHydrogenDay{}
	var scanned, skipped int64
	for rows.Next() {
		scanned++
		row, err := scanRawFrame(rows)
		if err != nil {
			return err
		}
		text := strings.TrimSpace(row.ParsedJSON)
		if isChunkManifest(text) {
			text, err = readChunkedPayload(ctx, td, cfg.TDengineDatabase, row.EventID, "parsed_fields")
			if err != nil {
				return fmt.Errorf("read complete mode evidence: %w", err)
			}
		}
		fields := fieldsForStats(row.Protocol, row.VIN, text)
		if len(fields) == 0 && row.Protocol == envelope.ProtocolYutongMQTT {
			fields = fieldsForStats(row.Protocol, row.VIN, row.RawText)
		}
		samples, err := stats.SamplesFromEnvelope(envelope.FrameEnvelope{EventID: row.EventID, Protocol: row.Protocol, MessageID: row.MessageID, VIN: row.VIN, Phone: row.Phone, DeviceID: row.DeviceID, SourceEndpoint: row.SourceEndpoint, EventTimeMS: row.EventTimeMS, ReceivedAtMS: row.ReceivedAtMS, Fields: fields, ParsedFields: fields, ParseStatus: envelope.ParseOK}, cfg.Location)
		if err != nil {
			return err
		}
		if len(samples) == 0 {
			skipped++
		}
		for _, s := range samples {
			if s.StatDate >= cfg.DateFrom && s.StatDate <= cfg.DateTo {
				addCustomerHydrogenSample(days, s)
			}
		}
		if cfg.ProgressEvery > 0 && scanned%cfg.ProgressEvery == 0 {
			slog.Info("customer hydrogen scan", "scanned", scanned, "source_days", len(days), "skipped", skipped)
		}
	}
	if err = rows.Err(); err != nil {
		return err
	}
	var known, samples int64
	var hydrogen float64
	for _, d := range days {
		known += d.KnownSamples
		samples += d.Samples
		hydrogen += d.HydrogenKM
	}
	slog.Info("customer hydrogen evidence", "scanned", scanned, "samples", samples, "known_mode_samples", known, "source_days", len(days), "hydrogen_km_across_sources", hydrogen, "dry_run", cfg.DryRun)
	if cfg.DryRun {
		return nil
	}
	return publishCustomerHydrogenDays(ctx, mysqlDB, days, cfg.DateFrom, owner)
}

func publishCustomerHydrogenDays(ctx context.Context, mysqlDB *sql.DB, days map[string]*customerHydrogenDay, date, owner string) error {
	// Publish only after a complete successful scan. Upserts are deterministic;
	// failures roll back the entire projection update.
	tx, err := mysqlDB.BeginTx(ctx, nil)
	if err != nil {
		return err
	}
	defer tx.Rollback()
	if owner != "" {
		// Fence publication with the same row lock used to claim jobs. A stale
		// worker must not overwrite results even before its next heartbeat.
		var valid int
		if err = tx.QueryRowContext(ctx, `SELECT 1 FROM platform_customer_hydrogen_job WHERE stat_date=? AND lease_owner=? AND status='running' AND lease_until>NOW(3) FOR UPDATE`, date, owner).Scan(&valid); err != nil {
			return fmt.Errorf("customer hydrogen publication lease lost: %w", err)
		}
	}
	keys := make([]string, 0, len(days))
	for k := range days {
		keys = append(keys, k)
	}
	sort.Strings(keys)
	for start := 0; start < len(keys); start += 200 {
		end := start + 200
		if end > len(keys) {
			end = len(keys)
		}
		values := []string{}
		args := []any{}
		for _, key := range keys[start:end] {
			d := days[key]
			values = append(values, "(?,?,?,?,?,?,?,?,?,?,?,?,?,?,NOW(3))")
			args = append(args, d.VIN, d.Date, d.Protocol, d.SourceKey, d.HydrogenKM, d.ObservedKM, d.ClassifiedKM, d.Samples, d.KnownSamples, d.Intervals, d.Rejected, d.First, d.Last, customerHydrogenAlgorithm)
		}
		_, err = tx.ExecContext(ctx, `INSERT INTO platform_customer_hydrogen_daily(vin,stat_date,protocol,source_key,hydrogen_km,observed_km,classified_km,sample_count,known_sample_count,interval_count,rejected_intervals,first_event_time,last_event_time,algorithm_version,calculated_at) VALUES `+strings.Join(values, ",")+` ON DUPLICATE KEY UPDATE hydrogen_km=VALUES(hydrogen_km),observed_km=VALUES(observed_km),classified_km=VALUES(classified_km),sample_count=VALUES(sample_count),known_sample_count=VALUES(known_sample_count),interval_count=VALUES(interval_count),rejected_intervals=VALUES(rejected_intervals),first_event_time=VALUES(first_event_time),last_event_time=VALUES(last_event_time),algorithm_version=VALUES(algorithm_version),calculated_at=VALUES(calculated_at)`, args...)
		if err != nil {
			return err
		}
	}
	return tx.Commit()
}