more refactoring with glm

quietlight

Apr 30, 2026, 9:19 PM

2HAQZPV377VV26SMPSXSZR6CL7SS2GTNPR5COIAPN47NLJILRQGAC

Dependencies

• [2] QVIGQOQZ more work on utils/ with glm
• [3] LBWQJEDH minor refactor and more tests for utils/
• [4] KZKLAINJ run out of space on nest, cleaned out
• [5] LQLC7S3A trying gemini: Inconsistent Standards in @utils/ refactoring

Change contents

• replacement in utils/wav_metadata.go at line 142
  [3.29717]→[3.29717:29787](∅→∅)
  − // Returns (sampleRate, duration, error). Does not parse INFO chunks.
  [3.29717]

  [3.29787]
  + // Returns (sampleRate, duration, error). Delegates to parseWAVFromBytes and

  + // extracts just the fields needed for batch processing.
• replacement in utils/wav_metadata.go at line 145
  [3.29869]→[3.29869:31391](∅→∅)
  − if len(data) < 44 {

  − return 0, 0, fmt.Errorf("file too small to be valid WAV")

  − }

  −

  − // Verify RIFF header

  − if string(data[0:4]) != "RIFF" {

  − return 0, 0, fmt.Errorf("not a valid WAV file (missing RIFF header)")

  − }

  −

  − // Verify WAVE format

  − if string(data[8:12]) != "WAVE" {

  − return 0, 0, fmt.Errorf("not a valid WAV file (missing WAVE format)")

  − }

  −

  − var channels, bitsPerSample int

  −

  − // Parse chunks - stop after finding data chunk

  − offset := 12

  − for offset < len(data)-8 {

  − chunkID := string(data[offset : offset+4])

  − chunkSize := int(binary.LittleEndian.Uint32(data[offset+4 : offset+8]))

  − offset += 8

  −

  − switch chunkID {

  − case "fmt ":

  − // Parse format chunk

  − if chunkSize >= 16 && offset+16 <= len(data) {

  − channels = int(binary.LittleEndian.Uint16(data[offset+2 : offset+4]))

  − sampleRate = int(binary.LittleEndian.Uint32(data[offset+4 : offset+8]))

  − bitsPerSample = int(binary.LittleEndian.Uint16(data[offset+14 : offset+16]))

  − }

  −

  − case "data":

  − // Found data chunk - calculate duration and return

  − if sampleRate > 0 && channels > 0 && bitsPerSample > 0 {

  − bytesPerSample := bitsPerSample / 8

  − bytesPerSecond := sampleRate * channels * bytesPerSample

  − if bytesPerSecond > 0 {

  − duration = float64(chunkSize) / float64(bytesPerSecond)

  − return sampleRate, duration, nil

  − }

  − }

  − return 0, 0, fmt.Errorf("invalid WAV: fmt chunk missing or corrupt before data chunk")

  − }

  −

  − // Move to next chunk (word-aligned)

  − offset += chunkSize

  − if chunkSize%2 != 0 {

  − offset++

  − }
  [3.29869]

  [3.31391]
  + metadata, err := parseWAVFromBytes(data)

  + if err != nil {

  + return 0, 0, err
• replacement in utils/wav_metadata.go at line 149
  [3.31394]→[3.31394:31565](∅→∅)
  −

  − // Data chunk not found within 4KB - file may have large INFO chunks

  − return 0, 0, fmt.Errorf("data chunk not found in first 4KB (try ParseWAVHeader for full parsing)")
  [3.31394]

  [3.31565]
  + return metadata.SampleRate, metadata.Duration, nil
• edit in utils/resample.go at line 9
  [3.74657]→[3.74657:74854](∅→∅)
  − return samples

  − }

  −

  − // Calculate ratio: toRate/fromRate (e.g., 16000/250000 = 0.064)

  − ratio := float64(toRate) / float64(fromRate)

  − newLen := int(float64(len(samples)) * ratio)

  − if newLen <= 0 {
• edit in utils/resample.go at line 10
  [3.74871]→[3.74871:75351](∅→∅)
  − }

  −

  − result := make([]float64, newLen)

  −

  − for i := range newLen {

  − // Source index in original samples (floating point)

  − srcIdx := float64(i) / ratio

  − idx0 := int(srcIdx)

  − idx1 := idx0 + 1

  −

  − // Clamp to valid range

  − if idx0 >= len(samples) {

  − idx0 = len(samples) - 1

  − }

  − if idx1 >= len(samples) {

  − idx1 = len(samples) - 1

  − }

  −

  − // Linear interpolation between adjacent samples

  − frac := srcIdx - float64(idx0)

  − result[i] = samples[idx0]*(1-frac) + samples[idx1]*frac
• replacement in utils/resample.go at line 11
  [3.75354]→[3.75354:75370](∅→∅)
  −

  − return result
  [3.75354]

  [3.75370]
  + // speed = fromRate/toRate: e.g. 250000/16000 = 15.625 (skip samples to downsample)

  + return Resample(samples, float64(fromRate)/float64(toRate))
• edit in utils/mapping.go at line 4
  [3.89580]→[3.89580:89596](∅→∅)
  − "database/sql"
• edit in utils/mapping.go at line 89
  [3.92000]→[3.4817:4962](∅→∅)
  − }

  −

  − // DBQueryer is an interface satisfied by *sql.DB and *sql.Tx

  − type DBQueryer interface {

  − Query(query string, args ...any) (*sql.Rows, error)
• replacement in utils/mapping.go at line 94
  [3.92217]→[3.4963:4983](∅→∅)
  − queryer DBQueryer,
  [3.92217]

  [3.92230]
  + queryer DB,
• replacement in utils/file_import.go at line 124
  [3.138445]→[3.138445:138596](∅→∅)
  − // DBQueryable is an interface satisfied by both *sql.DB and *sql.Tx

  − // for running duplicate hash checks against either.

  − type DBQueryable interface {
  [3.138445]

  [3.138596]
  + // DB is an interface satisfied by both *sql.DB and *sql.Tx.

  + // Used throughout utils for database queries that must work with either.

  + type DB interface {

  + Query(query string, args ...any) (*sql.Rows, error)
• replacement in utils/file_import.go at line 134
  [3.138841]→[3.138841:138944](∅→∅)
  − func CheckDuplicateHash(q DBQueryable, hash string) (existingID string, isDuplicate bool, err error) {
  [3.138841]

  [3.138944]
  + func CheckDuplicateHash(q DB, hash string) (existingID string, isDuplicate bool, err error) {
• replacement in utils/cluster_import.go at line 377
  [3.184587]→[3.184587:184637](∅→∅),[3.184637]→[2.902:981](∅→∅),[2.981]→[3.184698:184729](∅→∅),[3.184698]→[3.184698:184729](∅→∅)
  − var exists bool

  − err = tx.QueryRowContext(ctx,

  − "SELECT EXISTS(SELECT 1 FROM file WHERE xxh64_hash = ? AND active = true)",

  − fd.Hash,

  − ).Scan(&exists)

  −
  [3.184587]

  [3.184729]
  + _, isDuplicate, err := CheckDuplicateHash(tx, fd.Hash)
• replacement in utils/cluster_import.go at line 387
  [3.184927]→[3.184927:184941](∅→∅)
  − if exists {
  [3.184927]

  [3.184941]
  + if isDuplicate {