Index office documents (#340)

👌
This commit is contained in:
demig00d
2024-01-20 14:01:43 +03:00
committed by GitHub
parent b86bbf8407
commit e618d4ca47
3 changed files with 40 additions and 3 deletions

View File

@@ -13,6 +13,7 @@ import {
isFileFromDataloomPlugin,
isFileImage,
isFilePDF,
isFileOffice,
isFilePlaintext,
isFilenameIndexable,
logDebug,
@@ -106,6 +107,15 @@ async function getAndMapIndexedDocument(
content = await extractor.extractText(file)
}
// ** Office document **
else if (
isFileOffice(path) &&
settings.officeIndexing &&
extractor?.canFileBeExtracted(path)
) {
content = await extractor.extractText(file)
}
// ** Unsupported files **
else if (isFilenameIndexable(path)) {
content = file.path