ocr is working

This commit is contained in:
m
2026-09-16 12:44:48 +02:00
parent 050434325a
commit 906d48702c
21 changed files with 595 additions and 33 deletions
+47
View File
@@ -3,6 +3,7 @@ package repository
import (
"database/sql"
"errors"
"time"
)
// OcrJobRow maps ocr_jobs.
@@ -18,6 +19,9 @@ type OcrJobRow struct {
// ErrJobNotFound marks an OCR job absent or owned by another device.
var ErrJobNotFound = errors.New("ocr job not found")
// ErrNoQueuedJobs marks an OCR queue drained (no job to claim).
var ErrNoQueuedJobs = errors.New("no queued ocr jobs")
type OcrJobs struct{ DB *sql.DB }
func (o *OcrJobs) Create(jobID, deviceID, fileID string) error {
@@ -62,6 +66,49 @@ func (o *OcrJobs) TouchProcessing(deviceID, jobID string) error {
return err
}
// ClaimNext atomically picks the oldest queued job (FIFO) and moves it to
// `processing`. Safe for concurrent workers: `FOR UPDATE SKIP LOCKED` blocks
// the row as part of the same statement. Empty queue → ErrNoQueuedJobs.
func (o *OcrJobs) ClaimNext() (OcrJobRow, error) {
var row OcrJobRow
err := o.DB.QueryRow(
`WITH next AS (
SELECT job_id FROM ocr_jobs
WHERE status = 'queued'
ORDER BY created_at, job_id
LIMIT 1
FOR UPDATE SKIP LOCKED
)
UPDATE ocr_jobs
SET status = 'processing', started_at = NOW()
FROM next
WHERE ocr_jobs.job_id = next.job_id
RETURNING ocr_jobs.job_id, ocr_jobs.file_id, ocr_jobs.device_id, ocr_jobs.status`,
).Scan(&row.ID, &row.FileID, &row.DeviceID, &row.Status)
if err == sql.ErrNoRows {
return OcrJobRow{}, ErrNoQueuedJobs
}
if err != nil {
return OcrJobRow{}, err
}
return row, nil
}
// ResetStaleProcessing requeues jobs stuck in `processing` (worker crash /
// processus redémarré) et plus vieux que `olderThan`. Retourne le nb de jobs
// requeued.
func (o *OcrJobs) ResetStaleProcessing(olderThan time.Duration) (int64, error) {
res, err := o.DB.Exec(
`UPDATE ocr_jobs SET status = 'queued', started_at = NULL
WHERE status = 'processing' AND started_at < $1`,
time.Now().Add(-olderThan),
)
if err != nil {
return 0, err
}
return res.RowsAffected()
}
func (o *OcrJobs) Complete(deviceID, jobID, text string) error {
_, err := o.DB.Exec(
`UPDATE ocr_jobs SET status = 'done', text = NULLIF($3, ''), started_at = COALESCE(started_at, NOW()), completed_at = NOW()