fenix/webarchive
1
0
Fork 0
mirror of https://github.com/derfenix/webarchive.git synced 2026-03-11 21:35:34 +03:00
Code Issues Packages Projects Releases Wiki Activity

Fix page meta retrieve

Browse Source
This commit is contained in:
derFenix
2023-11-16 22:22:48 +03:00
parent 3147a0b683
commit e27fdabf78
2 changed files with 48 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
adapters/processors/processors.go
View File

@@ -129,8 +129,29 @@ func (p *Processors) GetMeta(ctx context.Context, url string) (entity.Meta, erro
return entity.Meta{}, fmt.Errorf("parse response body: %w", err)
}
var fc *html.Node
for fc = htmlNode.FirstChild; fc != nil && fc.Data != "html"; fc = fc.NextSibling {
}
if fc == nil {
return entity.Meta{}, fmt.Errorf("failed to find html tag")
}
fc = fc.NextSibling
if fc == nil {
return entity.Meta{}, fmt.Errorf("failed to find html tag")
}
for fc = fc.FirstChild; fc != nil && fc.Data != "head"; fc = fc.NextSibling {
fmt.Println(fc.Data)
}
if fc == nil {
return entity.Meta{}, fmt.Errorf("failed to find html tag")
}
meta := entity.Meta{}
getMetaData(htmlNode, &meta)
getMetaData(fc, &meta)
meta.Encoding = encodingFromHeader(response.Header)
return meta, nil