feat: Winkeyer
This commit is contained in:
@@ -0,0 +1,44 @@
|
||||
package adif
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"testing"
|
||||
"unicode/utf8"
|
||||
)
|
||||
|
||||
// Loggers (notably Log4OM's UDP/exported ADIF) sometimes declare a field
|
||||
// length as the CHARACTER count instead of the UTF-8 byte count, truncating
|
||||
// multibyte values mid-rune. The parser must recover the full value.
|
||||
func TestCharCountLengthRepair(t *testing.T) {
|
||||
cases := []struct{ name, wantQTH, wantName, adi string }{
|
||||
{
|
||||
name: "latin",
|
||||
wantQTH: "Tóalmás", // 7 chars / 9 bytes, declared 7
|
||||
wantName: "Laci Budai",
|
||||
adi: "<EOH>\n<CALL:5>HA5XY<QTH:7>Tóalmás<NAME:10>Laci Budai<EOR>\n",
|
||||
},
|
||||
{
|
||||
name: "cyrillic",
|
||||
wantQTH: "Дзержинск", // 9 chars / 18 bytes, declared 9
|
||||
wantName: "Александр Чайка", // 15 chars / 29 bytes, declared 15
|
||||
adi: "<EOH>\n<CALL:6>UA3TFS<NAME:15>Александр Чайка<QTH:9>Дзержинск<EOR>\n",
|
||||
},
|
||||
}
|
||||
for _, c := range cases {
|
||||
t.Run(c.name, func(t *testing.T) {
|
||||
var got Record
|
||||
if err := Parse(strings.NewReader(c.adi), func(r Record) error { got = r; return nil }); err != nil {
|
||||
t.Fatalf("parse: %v", err)
|
||||
}
|
||||
if got["qth"] != c.wantQTH {
|
||||
t.Errorf("qth = %q, want %q", got["qth"], c.wantQTH)
|
||||
}
|
||||
if got["name"] != c.wantName {
|
||||
t.Errorf("name = %q, want %q", got["name"], c.wantName)
|
||||
}
|
||||
if !utf8.ValidString(got["name"]) || !utf8.ValidString(got["qth"]) {
|
||||
t.Errorf("result not valid UTF-8: name=%q qth=%q", got["name"], got["qth"])
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
+19
-4
@@ -27,6 +27,13 @@ type Exporter struct {
|
||||
// AppName / AppVersion populate the ADIF header comments. Optional.
|
||||
AppName string
|
||||
AppVersion string
|
||||
|
||||
// IncludeAppFields controls whether application-specific fields (ADIF
|
||||
// "APP_<programid>_<name>" tags, e.g. Log4OM's APP_LOG4OM_* or our own
|
||||
// OpsLog extensions) are written. Leave false for a clean standard-ADIF
|
||||
// export destined for another logger; set true for a full OpsLog→OpsLog
|
||||
// round-trip that preserves everything.
|
||||
IncludeAppFields bool
|
||||
}
|
||||
|
||||
// ExportFile creates path (overwriting if it exists) and writes every QSO.
|
||||
@@ -70,7 +77,7 @@ func (e *Exporter) Export(ctx context.Context, w io.Writer) (int, error) {
|
||||
|
||||
count := 0
|
||||
err := e.Repo.IterateAll(ctx, func(q qso.QSO) error {
|
||||
writeRecord(bw, q)
|
||||
writeRecord(bw, q, e.IncludeAppFields)
|
||||
count++
|
||||
return nil
|
||||
})
|
||||
@@ -84,7 +91,8 @@ func (e *Exporter) Export(ctx context.Context, w io.Writer) (int, error) {
|
||||
func SingleRecordADIF(q qso.QSO) string {
|
||||
var b strings.Builder
|
||||
bw := bufio.NewWriter(&b)
|
||||
writeRecord(bw, q)
|
||||
// Uploads target other services — keep it standard (no app-specific tags).
|
||||
writeRecord(bw, q, false)
|
||||
bw.Flush()
|
||||
return b.String()
|
||||
}
|
||||
@@ -93,7 +101,7 @@ func SingleRecordADIF(q qso.QSO) string {
|
||||
// Empty fields are omitted. MODE/SUBMODE are massaged so a "promoted"
|
||||
// mode (e.g. FT4 stored without a parent) is exported as the canonical
|
||||
// pair MODE=MFSK SUBMODE=FT4 — round-trips cleanly with strict loggers.
|
||||
func writeRecord(bw *bufio.Writer, q qso.QSO) {
|
||||
func writeRecord(bw *bufio.Writer, q qso.QSO, includeApp bool) {
|
||||
// --- Core ---
|
||||
writeField(bw, "CALL", q.Callsign)
|
||||
|
||||
@@ -218,8 +226,15 @@ func writeRecord(bw *bufio.Writer, q qso.QSO) {
|
||||
writeField(bw, "NOTES", q.Notes)
|
||||
|
||||
// --- Extras (unpromoted ADIF fields preserved verbatim) ---
|
||||
// In standard mode we drop application-specific tags (APP_*) so the file
|
||||
// stays portable to other loggers; in full mode they're kept for a
|
||||
// lossless OpsLog round-trip.
|
||||
for k, v := range q.Extras {
|
||||
writeField(bw, strings.ToUpper(k), v)
|
||||
tag := strings.ToUpper(k)
|
||||
if !includeApp && strings.HasPrefix(tag, "APP_") {
|
||||
continue
|
||||
}
|
||||
writeField(bw, tag, v)
|
||||
}
|
||||
|
||||
bw.WriteString("<EOR>\n")
|
||||
|
||||
+90
-1
@@ -20,6 +20,7 @@ import (
|
||||
type ImportResult struct {
|
||||
Total int `json:"total"` // records found in the file
|
||||
Imported int `json:"imported"` // successfully inserted
|
||||
Updated int `json:"updated"` // existing QSOs refreshed (update-duplicates mode)
|
||||
Skipped int `json:"skipped"` // dropped (missing required fields)
|
||||
Duplicates int `json:"duplicates"` // matched an existing or earlier-in-file QSO
|
||||
DuplicateSamples []string `json:"duplicate_samples"` // up to maxDuplicateSamples human-readable IDs of skipped duplicates
|
||||
@@ -36,6 +37,19 @@ type Importer struct {
|
||||
Repo *qso.Repo
|
||||
BatchSize int // 0 → 500
|
||||
SkipDuplicates bool // when true, records matching an existing or earlier-in-file QSO are skipped; otherwise all are inserted
|
||||
// UpdateDuplicates, when true, takes precedence over SkipDuplicates:
|
||||
// a record matching an existing QSO MERGES its non-empty fields onto
|
||||
// that QSO (refreshes QSL/confirmation statuses on re-sync) instead of
|
||||
// being skipped or re-inserted.
|
||||
UpdateDuplicates bool
|
||||
// Enrich, when set, is called on each parsed QSO before dedup/insert.
|
||||
// Used to recompute country / zones from cty.dat so a bad COUNTRY in the
|
||||
// source file (common with contest loggers) is corrected on the way in.
|
||||
Enrich func(*qso.QSO)
|
||||
// OnProgress, when set, is called periodically with (processed, total)
|
||||
// record counts so the UI can show a progress bar. total is an estimate
|
||||
// from counting <EOR> tags up front.
|
||||
OnProgress func(processed, total int)
|
||||
}
|
||||
|
||||
// ImportFile reads the file at path and imports it into the repo. The
|
||||
@@ -62,6 +76,14 @@ func (im *Importer) ImportFile(ctx context.Context, path string) (ImportResult,
|
||||
// Windows-1252 (É is one byte 0xC9). Pre-decoding to UTF-8 would make É
|
||||
// two bytes, and the parser reading 7 bytes after the tag would chop the
|
||||
// É in half → "YAOUND" + an orphan 0xC3 byte → "YAOUND�" after JSON.
|
||||
// ValueDecoderFor returns the per-field byte decoder appropriate for a raw
|
||||
// ADIF payload: identity when it's valid UTF-8, otherwise a Windows-1252
|
||||
// decoder. Exposed so non-file ingest paths (UDP auto-log from Log4OM /
|
||||
// JTAlert) transcode accented NAME/QTH fields the same way file import does.
|
||||
func ValueDecoderFor(data []byte) func([]byte) string {
|
||||
return pickValueDecoder(data)
|
||||
}
|
||||
|
||||
func pickValueDecoder(data []byte) func([]byte) string {
|
||||
if utf8.Valid(data) {
|
||||
return nil // identity
|
||||
@@ -97,6 +119,15 @@ func (im *Importer) importBytes(ctx context.Context, data []byte) (ImportResult,
|
||||
res := ImportResult{}
|
||||
batch := make([]qso.QSO, 0, im.BatchSize)
|
||||
|
||||
// Up-front record-count estimate (count <EOR> tags, case-insensitive) so
|
||||
// the UI progress bar has a denominator. Cheap single scan.
|
||||
total := countEOR(data)
|
||||
reportProgress := func(force bool) {
|
||||
if im.OnProgress != nil && (force || res.Total%200 == 0) {
|
||||
im.OnProgress(res.Total, total)
|
||||
}
|
||||
}
|
||||
|
||||
// One upfront query for every existing dedup key — cheaper than N
|
||||
// per-record EXISTS calls. The same map gets new keys appended as we
|
||||
// import so duplicates inside the file are caught too. Loaded
|
||||
@@ -107,6 +138,16 @@ func (im *Importer) importBytes(ctx context.Context, data []byte) (ImportResult,
|
||||
return res, fmt.Errorf("load dedupe keys: %w", err)
|
||||
}
|
||||
|
||||
// Update-duplicates mode needs the existing row's ID per key so it can
|
||||
// fetch, merge and write it back. Loaded only when needed (extra query).
|
||||
var keyIDs map[string]int64
|
||||
if im.UpdateDuplicates {
|
||||
keyIDs, err = im.Repo.DedupeKeyIDs(ctx)
|
||||
if err != nil {
|
||||
return res, fmt.Errorf("load dedupe ids: %w", err)
|
||||
}
|
||||
}
|
||||
|
||||
flush := func() error {
|
||||
if len(batch) == 0 {
|
||||
return nil
|
||||
@@ -119,6 +160,7 @@ func (im *Importer) importBytes(ctx context.Context, data []byte) (ImportResult,
|
||||
|
||||
err = ParseWithDecoder(bytes.NewReader(data), decode, func(rec Record) error {
|
||||
res.Total++
|
||||
reportProgress(false)
|
||||
q, ok := recordToQSO(rec)
|
||||
if !ok {
|
||||
res.Skipped++
|
||||
@@ -128,6 +170,9 @@ func (im *Importer) importBytes(ctx context.Context, data []byte) (ImportResult,
|
||||
}
|
||||
return nil
|
||||
}
|
||||
if im.Enrich != nil {
|
||||
im.Enrich(&q)
|
||||
}
|
||||
key := qso.DedupeKey(q.Callsign, q.QSODate.UTC().Format("2006-01-02T15:04"), q.Band, q.Mode)
|
||||
if _, dup := seen[key]; dup {
|
||||
res.Duplicates++
|
||||
@@ -138,6 +183,28 @@ func (im *Importer) importBytes(ctx context.Context, data []byte) (ImportResult,
|
||||
q.QSODate.UTC().Format("2006-01-02 15:04"),
|
||||
q.Band, q.Mode))
|
||||
}
|
||||
if im.UpdateDuplicates {
|
||||
if id, ok := keyIDs[key]; ok {
|
||||
existing, gerr := im.Repo.GetByID(ctx, id)
|
||||
if gerr != nil {
|
||||
if len(res.Errors) < maxErrors {
|
||||
res.Errors = append(res.Errors,
|
||||
fmt.Sprintf("record %d (%s): load existing: %v", res.Total, q.Callsign, gerr))
|
||||
}
|
||||
return nil
|
||||
}
|
||||
qso.MergeNonZero(&existing, q)
|
||||
if uerr := im.Repo.Update(ctx, existing); uerr != nil {
|
||||
if len(res.Errors) < maxErrors {
|
||||
res.Errors = append(res.Errors,
|
||||
fmt.Sprintf("record %d (%s): update: %v", res.Total, q.Callsign, uerr))
|
||||
}
|
||||
return nil
|
||||
}
|
||||
res.Updated++
|
||||
}
|
||||
return nil
|
||||
}
|
||||
if im.SkipDuplicates {
|
||||
return nil
|
||||
}
|
||||
@@ -159,9 +226,28 @@ func (im *Importer) importBytes(ctx context.Context, data []byte) (ImportResult,
|
||||
if err := flush(); err != nil {
|
||||
return res, err
|
||||
}
|
||||
reportProgress(true) // final 100%
|
||||
return res, nil
|
||||
}
|
||||
|
||||
// countEOR estimates the record count by counting case-insensitive <EOR>
|
||||
// tags. Used only to give the import progress bar a denominator.
|
||||
func countEOR(data []byte) int {
|
||||
n := 0
|
||||
for i := 0; i+4 <= len(data); i++ {
|
||||
if data[i] != '<' {
|
||||
continue
|
||||
}
|
||||
if (data[i+1] == 'e' || data[i+1] == 'E') &&
|
||||
(data[i+2] == 'o' || data[i+2] == 'O') &&
|
||||
(data[i+3] == 'r' || data[i+3] == 'R') &&
|
||||
(i+4 < len(data) && (data[i+4] == '>' || data[i+4] == ':')) {
|
||||
n++
|
||||
}
|
||||
}
|
||||
return n
|
||||
}
|
||||
|
||||
// adifPromoted lists every lowercase ADIF tag that maps to a promoted column.
|
||||
// Anything not in this set ends up in Extras.
|
||||
var adifPromoted = stringSet(
|
||||
@@ -178,7 +264,7 @@ var adifPromoted = stringSet(
|
||||
"age", "lat", "lon", "rig", "ant",
|
||||
// QSL
|
||||
"qsl_sent", "qsl_rcvd",
|
||||
"qslsdate", "qslrdate", "qsl_via", "qslmsg", "qslmsg_rcvd",
|
||||
"qslsdate", "qslrdate", "qsl_via", "qsl_sent_via", "qsl_rcvd_via", "qslmsg", "qslmsg_rcvd",
|
||||
"lotw_qsl_sent", "lotw_qsl_rcvd", "lotw_qslsdate", "lotw_qslrdate",
|
||||
"eqsl_qsl_sent", "eqsl_qsl_rcvd", "eqsl_qslsdate", "eqsl_qslrdate",
|
||||
"clublog_qso_upload_date", "clublog_qso_upload_status",
|
||||
@@ -300,6 +386,9 @@ func recordToQSO(rec Record) (qso.QSO, bool) {
|
||||
q.QSLSentDate = rec["qslsdate"]
|
||||
q.QSLRcvdDate = rec["qslrdate"]
|
||||
q.QSLVia = rec["qsl_via"]
|
||||
if q.QSLVia == "" { // many loggers (Log4OM) write QSL_SENT_VIA instead
|
||||
q.QSLVia = rec["qsl_sent_via"]
|
||||
}
|
||||
q.QSLMsg = rec["qslmsg"]
|
||||
q.QSLMsgRcvd = rec["qslmsg_rcvd"]
|
||||
q.LOTWSent = rec["lotw_qsl_sent"]
|
||||
|
||||
@@ -16,6 +16,7 @@ import (
|
||||
"io"
|
||||
"strconv"
|
||||
"strings"
|
||||
"unicode/utf8"
|
||||
)
|
||||
|
||||
// Record is a single ADIF record. Keys are lowercased field names.
|
||||
@@ -83,6 +84,17 @@ func parseWith(r io.Reader, decodeValue func([]byte) string, fn func(Record) err
|
||||
if _, err := io.ReadFull(br, val); err != nil {
|
||||
return fmt.Errorf("read field %s: %w", name, err)
|
||||
}
|
||||
// Repair character-count lengths. The ADIF spec says LENGTH is a
|
||||
// byte count, but some loggers (notably Log4OM's UDP "ADIF
|
||||
// message") write the CHARACTER count instead. For UTF-8 values
|
||||
// with accented chars that truncates mid-rune — e.g. "<QTH:7>
|
||||
// Tóalmás" is 9 bytes but says 7, leaving an orphan byte that
|
||||
// renders as "Tóalm�". When we're in UTF-8 mode (no Windows-1252
|
||||
// decoder) and the naive byte read isn't valid UTF-8, keep reading
|
||||
// until the value holds `length` whole runes (or the next tag).
|
||||
if decodeValue == nil && !utf8.Valid(val) {
|
||||
val = extendToRunes(br, val, length)
|
||||
}
|
||||
if headerDone && name != "" {
|
||||
if decodeValue != nil {
|
||||
rec[name] = decodeValue(val)
|
||||
@@ -94,6 +106,37 @@ func parseWith(r io.Reader, decodeValue func([]byte) string, fn func(Record) err
|
||||
}
|
||||
}
|
||||
|
||||
// extendToRunes recovers a value whose declared length was a character count
|
||||
// rather than a byte count. `have` holds the first `wantRunes` BYTES of the
|
||||
// value, which turned out to be invalid UTF-8 (a multibyte rune was cut). We
|
||||
// append bytes from br until the value holds `wantRunes` complete runes — or
|
||||
// until the next '<' (start of the following tag) / EOF, so we never cross
|
||||
// into another field. Capped so a genuinely-corrupt value can't run away.
|
||||
func extendToRunes(br *bufio.Reader, have []byte, wantRunes int) []byte {
|
||||
const maxExtra = 8 // at most ~4 extra bytes/rune for the few cut runes
|
||||
limit := len(have) + maxExtra*wantRunes + maxExtra
|
||||
for len(have) < limit {
|
||||
// Stop only when the value is complete UTF-8 (no partial trailing
|
||||
// rune) AND holds enough runes. Checking utf8.RuneCount alone is a
|
||||
// trap: a trailing orphan lead byte (e.g. the D0 of a cut Cyrillic
|
||||
// "а") counts as one rune, so the loop would stop one continuation
|
||||
// byte short → "Чайк�". Requiring utf8.Valid forces us to read it.
|
||||
if utf8.Valid(have) && utf8.RuneCount(have) >= wantRunes {
|
||||
break
|
||||
}
|
||||
b, err := br.ReadByte()
|
||||
if err != nil {
|
||||
break
|
||||
}
|
||||
if b == '<' {
|
||||
_ = br.UnreadByte() // belongs to the next tag — leave it
|
||||
break
|
||||
}
|
||||
have = append(have, b)
|
||||
}
|
||||
return have
|
||||
}
|
||||
|
||||
// parseSpec splits "callsign:5", "callsign:5:S" or "eor" into name and length.
|
||||
// name is lowercased; length is 0 for control tags or when missing.
|
||||
func parseSpec(spec string) (name string, length int) {
|
||||
|
||||
Reference in New Issue
Block a user