ec1f89217a
Full Mattermost server source with integrated Community Enterprise features. Includes vendor directory for offline/air-gapped builds. Structure: - enterprise-impl/: Enterprise feature implementations - enterprise-community/: Init files that register implementations - enterprise/: Bridge imports (community_imports.go) - vendor/: All dependencies for offline builds Build (online): go build ./cmd/mattermost Build (offline/air-gapped): go build -mod=vendor ./cmd/mattermost 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
51 lines
1.7 KiB
Go
51 lines
1.7 KiB
Go
// Copyright (c) 2015-present Mattermost, Inc. All Rights Reserved.
|
|
// See LICENSE.txt for license information.
|
|
|
|
package docextractor
|
|
|
|
import (
|
|
"io"
|
|
|
|
"github.com/mattermost/mattermost/server/public/shared/mlog"
|
|
)
|
|
|
|
// ExtractSettings defines the features enabled/disable during the document text extraction.
|
|
type ExtractSettings struct {
|
|
ArchiveRecursion bool
|
|
MMPreviewURL string
|
|
MMPreviewSecret string
|
|
}
|
|
|
|
// Extract extract the text from a document using the system default extractors
|
|
func Extract(logger mlog.LoggerIFace, filename string, r io.ReadSeeker, settings ExtractSettings) (string, error) {
|
|
return ExtractWithExtraExtractors(logger, filename, r, settings, []Extractor{})
|
|
}
|
|
|
|
// ExtractWithExtraExtractors extract the text from a document using the provided extractors beside the system default extractors.
|
|
func ExtractWithExtraExtractors(logger mlog.LoggerIFace, filename string, r io.ReadSeeker, settings ExtractSettings, extraExtractors []Extractor) (string, error) {
|
|
enabledExtractors := &combineExtractor{
|
|
logger: logger,
|
|
}
|
|
for _, extraExtractor := range extraExtractors {
|
|
enabledExtractors.Add(extraExtractor)
|
|
}
|
|
enabledExtractors.Add(&documentExtractor{})
|
|
enabledExtractors.Add(&pdfExtractor{})
|
|
|
|
if settings.ArchiveRecursion {
|
|
enabledExtractors.Add(&archiveExtractor{SubExtractor: enabledExtractors})
|
|
} else {
|
|
enabledExtractors.Add(&archiveExtractor{})
|
|
}
|
|
|
|
if settings.MMPreviewURL != "" {
|
|
enabledExtractors.Add(newMMPreviewExtractor(settings.MMPreviewURL, settings.MMPreviewSecret, pdfExtractor{}))
|
|
}
|
|
enabledExtractors.Add(&plainExtractor{})
|
|
|
|
if enabledExtractors.Match(filename) {
|
|
return enabledExtractors.Extract(filename, r)
|
|
}
|
|
return "", nil
|
|
}
|