scripte/archivmail
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(PROJ-44): ocr_chars-Spalte + SetOCRResult-Helper

Browse Source 
DB-Schema bekommt eine idempotente ocr_chars BIGINT-Spalte (Default 0).
SetOCRResult schreibt status und chars atomar; GetOCRMeta liest beide
mit COALESCE-Defaults. Der OCR-Worker ersetzt jeden SetOCRStatus-Call
durch SetOCRResult und uebergibt die extrahierte Zeichenzahl bei 'done'.
This commit is contained in:
sysops
2026-05-10 22:20:46 +02:00
parent 7be73c1041
commit 5078830469
3 changed files with 70 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
internal/ocr/worker.go
+10 -9
View File
@@ -137,26 +137,26 @@ func (w *Worker) process(ctx context.Context, job Job) {
logger := w.logger.With("mail_id", job.MailID, "tenant_id", job.TenantID)
if w.store.OCREnabled(ctx, job.TenantID) == false {
_ = w.store.SetOCRStatus(ctx, job.MailID, "disabled")
_ = w.store.SetOCRResult(ctx, job.MailID, "disabled", 0)
return
}
raw, err := w.store.Load(job.MailID)
if err != nil {
logger.Warn("ocr worker: load failed", "err", err)
_ = w.store.SetOCRStatus(ctx, job.MailID, "failed")
_ = w.store.SetOCRResult(ctx, job.MailID, "failed", 0)
return
}
pm, err := mailparser.Parse(raw)
if err != nil {
logger.Warn("ocr worker: parse failed", "err", err)
_ = w.store.SetOCRStatus(ctx, job.MailID, "failed")
_ = w.store.SetOCRResult(ctx, job.MailID, "failed", 0)
return
}
if len(pm.Attachments) == 0 {
_ = w.store.SetOCRStatus(ctx, job.MailID, "skipped")
_ = w.store.SetOCRResult(ctx, job.MailID, "skipped", 0)
return
}
@@ -183,7 +183,7 @@ func (w *Worker) process(ctx context.Context, job Job) {
}
if processed == 0 {
_ = w.store.SetOCRStatus(ctx, job.MailID, "skipped")
_ = w.store.SetOCRResult(ctx, job.MailID, "skipped", 0)
return
}
@@ -191,15 +191,16 @@ func (w *Worker) process(ctx context.Context, job Job) {
updater, ok := idx.(index.AttachmentTextUpdater)
if !ok {
logger.Warn("ocr worker: indexer does not support AttachmentTextUpdater — text dropped")
_ = w.store.SetOCRStatus(ctx, job.MailID, "failed")
_ = w.store.SetOCRResult(ctx, job.MailID, "failed", 0)
return
}
if err := updater.UpdateAttachmentText(job.MailID, combined.String()); err != nil {
logger.Warn("ocr worker: index update failed", "err", err)
_ = w.store.SetOCRStatus(ctx, job.MailID, "failed")
_ = w.store.SetOCRResult(ctx, job.MailID, "failed", 0)
return
}
_ = w.store.SetOCRStatus(ctx, job.MailID, "done")
logger.Info("ocr worker: indexed", "attachments", processed, "chars", combined.Len())
chars := int64(combined.Len())
_ = w.store.SetOCRResult(ctx, job.MailID, "done", chars)
logger.Info("ocr worker: indexed", "attachments", processed, "chars", chars)
}