playlink/internal/api/export.go

package api

import (
	"encoding/csv"
	"encoding/json"
	"fmt"
	"net/http"
	"net/url"
	"regexp"
	"strconv"
	"strings"
	"time"

	"github.com/cdnutter/playlink/internal/engine"
	"github.com/cdnutter/playlink/internal/probe"
	"github.com/cdnutter/playlink/internal/store"
)

var kindLabels = map[string]string{
	store.KindPath:      "same path",
	store.KindPathCase:  "path differs in case",
	store.KindExtension: "other format",
	store.KindFuzzy:     "similar title",
	store.KindElsewhere: "same top folder, other subfolder",
	store.KindApprox:    "approximate match",
	store.KindNone:      "no match",
}

var exportHeader = []string{
	"status", "match_type", "ignored", "recommendation",
	"slave_path", "master_path", "link_target", "reason",
	"slave_size_bytes", "master_size_bytes",
	"slave_codec", "slave_lossless", "slave_bitrate_kbps", "slave_sample_rate_hz", "slave_bit_depth", "slave_duration_s",
	"master_codec", "master_lossless", "master_bitrate_kbps", "master_sample_rate_hz", "master_bit_depth", "master_duration_s",
	"better_copy", "duration_difference_s", "notes", "error",
	"slave_modified", "master_modified", "checked_at",
}

// csvSafe neutralises cells a spreadsheet would evaluate as a formula (a file
// named "=HYPERLINK(...).mp3" must stay text).
func csvSafe(s string) string {
	if s != "" && strings.ContainsRune("=+-@\t\r", rune(s[0])) {
		return "'" + s
	}
	return s
}

func recommendation(m store.Match, v probe.Verdict) string {
	switch {
	case m.Ignored:
		return "ignored"
	case m.Status == store.StatusLinked:
		return "done"
	case m.Status == store.StatusIdentical:
		return "link (identical)"
	case m.Status == store.StatusDifferent || m.Status == store.StatusSimilar:
		switch v.Better {
		case "master":
			return "upgrade to master copy"
		case "slave":
			return "review: master is lower quality"
		}
		return "review"
	case m.Status == store.StatusError:
		return "fix error"
	}
	return ""
}

func probeCols(raw json.RawMessage) []string {
	if len(raw) == 0 {
		return make([]string, 6)
	}
	var p probe.Info
	if json.Unmarshal(raw, &p) != nil || p.Error != "" {
		return make([]string, 6)
	}
	num := func(n int) string {
		if n == 0 {
			return ""
		}
		return strconv.Itoa(n)
	}
	bitrate := ""
	if p.Bitrate > 0 {
		bitrate = strconv.FormatInt((p.Bitrate+500)/1000, 10)
	}
	dur := ""
	if p.Duration > 0 {
		dur = strconv.FormatFloat(p.Duration, 'f', 1, 64)
	}
	return []string{p.Codec, strconv.FormatBool(p.Lossless), bitrate, num(p.SampleRate), num(p.BitDepth), dur}
}

func nsTime(ns int64) string {
	if ns == 0 {
		return ""
	}
	return time.Unix(0, ns).UTC().Format(time.RFC3339)
}

var unsafeName = regexp.MustCompile(`[^A-Za-z0-9._-]+`)

// handleExport streams the pair's scan results as CSV. It honours the same
// status and search filters as the Review list.
func (s *Server) handleExport(w http.ResponseWriter, r *http.Request) {
	id, _ := idParam(r)
	p, err := s.st.Pair(r.Context(), id)
	if err != nil {
		opErr(w, err)
		return
	}
	q := r.URL.Query()
	f := store.MatchFilter{Status: q.Get("status"), Query: q.Get("q")}

	scope := f.Status
	if scope == "" {
		scope = "all"
	}
	slug := strings.Trim(unsafeName.ReplaceAllString(p.Name, "-"), "-")
	if slug == "" {
		slug = fmt.Sprintf("pair%d", p.ID)
	}
	name := fmt.Sprintf("playlink-%s-%s-%s.csv", slug, scope, time.Now().Format("20060102-1504"))
	w.Header().Set("Content-Type", "text/csv; charset=utf-8")
	w.Header().Set("Content-Disposition", fmt.Sprintf(`attachment; filename="%s"; filename*=UTF-8''%s`, name, url.PathEscape(name)))

	// A UTF-8 byte order mark makes Excel read non-ASCII names correctly.
	_, _ = w.Write([]byte{0xEF, 0xBB, 0xBF})
	cw := csv.NewWriter(w)
	_ = cw.Write(exportHeader)
	rows := 0
	err = s.st.ExportMatches(r.Context(), id, f, func(m store.Match) error {
		var v probe.Verdict
		_ = json.Unmarshal(m.Verdict, &v)
		target := ""
		if m.RelMaster != "" && m.Status != store.StatusLinked {
			target = engine.LinkTarget(m.RelSlave, m.RelMaster, false)
		}
		size := func(n int64) string {
			if n == 0 {
				return ""
			}
			return strconv.FormatInt(n, 10)
		}
		delta := ""
		if v.DurationDelta > 0 {
			delta = strconv.FormatFloat(v.DurationDelta, 'f', 1, 64)
		}
		rec := []string{
			m.Status, kindLabels[m.Kind], strconv.FormatBool(m.Ignored), recommendation(m, v),
			csvSafe(m.RelSlave), csvSafe(m.RelMaster), csvSafe(target), csvSafe(m.Reason),
			size(m.SlaveSize), size(m.MasterSize),
		}
		rec = append(rec, probeCols(m.SlaveProbe)...)
		rec = append(rec, probeCols(m.MasterProbe)...)
		better := v.Better
		if better == "unknown" {
			better = ""
		}
		rec = append(rec, better, delta, csvSafe(strings.Join(v.Notes, "; ")), csvSafe(m.Note),
			nsTime(m.SlaveMtime), nsTime(m.MasterMtime), time.Unix(m.UpdatedAt, 0).UTC().Format(time.RFC3339))
		rows++
		if err := cw.Write(rec); err != nil {
			return err
		}
		if rows%500 == 0 {
			cw.Flush()
		}
		return cw.Error()
	})
	cw.Flush()
	if err != nil {
		// Headers are already sent; the truncated file is the only signal left.
		s.log.Warn("csv export interrupted", "pair", id, "rows", rows, "err", err)
		return
	}
	s.log.Info("csv export", "pair", p.Name, "filter", scope, "rows", rows)
}