Merged in feature/part (pull request #113)

Size Import Parts

* parts

* think

* workingpart

* textout
This commit is contained in:
Michael McGuinness
2025-04-08 12:44:09 +00:00
parent 60812aba91
commit edc50c7510
21 changed files with 873 additions and 66 deletions
+30 -10
View File
@@ -11,28 +11,48 @@ import (
"github.com/aws/aws-sdk-go-v2/service/textract"
"github.com/aws/aws-sdk-go-v2/service/textract/types"
"github.com/jackc/pgx/v5/pgtype"
)
func (s *Service) triggerExtract(ctx context.Context, clean *repository.Currentcleanentry) error {
version := build.GetVersionUnixTimestamp()
return s.cfg.ExecuteDBTransaction(ctx, func(ctx context.Context, q *repository.Queries) error {
triggerId, err := s.cfg.GetDBQueries().AddDocumentTextTrigger(ctx, &repository.AddDocumentTextTriggerParams{
Cleanid: clean.ID,
Version: version,
key := objectstore.BucketKey{
CreatedAt: time.Now().UTC(),
Location: objectstore.TextTextract,
ClientID: clean.Clientid,
}
currentPart, err := q.GetTextractOutputCurrentPart(ctx, &repository.GetTextractOutputCurrentPartParams{
Clientid: &clean.Clientid,
Querydate: pgtype.Timestamptz{
Time: key.CreatedAt,
Valid: true,
},
})
if err != nil {
return err
}
key := objectstore.BucketKey{
CreatedAt: time.Now().UTC(),
Location: objectstore.TextTextract,
ClientID: clean.Clientid,
EntityID: triggerId,
}
keyStr := key.String()
part := s.cfg.GetDirectoryPart(currentPart.Part, currentPart.Count)
key.Part = &part
triggerId, err := s.cfg.GetDBQueries().AddDocumentTextTrigger(ctx, &repository.AddDocumentTextTriggerParams{
Cleanid: clean.ID,
Version: version,
Createdat: pgtype.Timestamp{
Time: key.CreatedAt,
Valid: true,
},
Part: *key.Part,
})
if err != nil {
return err
}
key.EntityID = triggerId
keyStr := key.String()
jobTag := triggerId.String()
res, err := s.cfg.GetTextractClient().StartDocumentTextDetection(ctx, &textract.StartDocumentTextDetectionInput{