ben/coze-studio
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

feat(infra): integrate PaddleOCR's PP-StructureV3 as a document parser backend (#714)

Browse Source
This commit is contained in:
Lin Manhui
2025-08-13 16:37:42 +08:00
committed by GitHub
parent 708a6ed0c0
commit 6b60c07c22
30 changed files with 657 additions and 174 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
backend/infra/impl/document/parser/builtin/parse_image.go
View File

@@ -31,7 +31,7 @@ import (
"github.com/coze-dev/coze-studio/backend/types/errno"
)
func parseImage(config *contract.Config, model chatmodel.BaseChatModel) parseFn {
func ParseImage(config *contract.Config, model chatmodel.BaseChatModel) ParseFn {
return func(ctx context.Context, reader io.Reader, opts ...parser.Option) (docs []*schema.Document, err error) {
options := parser.GetCommonOptions(&parser.Options{}, opts...)
doc := &schema.Document{
@@ -76,14 +76,14 @@ func parseImage(config *contract.Config, model chatmodel.BaseChatModel) parseFn
output, err := model.Generate(ctx, []*schema.Message{input})
if err != nil {
return nil, fmt.Errorf("[parseImage] model generate failed: %w", err)
return nil, fmt.Errorf("[ParseImage] model generate failed: %w", err)
}
doc.Content = output.Content
case contract.ImageAnnotationTypeManual:
// do nothing
default:
return nil, fmt.Errorf("[parseImage] unknown image annotation type=%d", config.ParsingStrategy.ImageAnnotationType)
return nil, fmt.Errorf("[ParseImage] unknown image annotation type=%d", config.ParsingStrategy.ImageAnnotationType)
}
return []*schema.Document{doc}, nil