snapshot of https://github.com/mattermost/mattermost.git (master, 2026-09-16; blobs > 131072B omitted)

server/platform/services/docextractor/documents.go

@@ -0,0 +1,69 @@
+// Copyright (c) 2015-present Mattermost, Inc. All Rights Reserved.
+// See LICENSE.txt for license information.
+
+package docextractor
+
+import (
+	"context"
+	"errors"
+	"io"
+	"path"
+	"strings"
+
+	"code.sajari.com/docconv/v2"
+
+	"github.com/mattermost/mattermost/server/v8/channels/utils"
+)
+
+type documentExtractor struct{}
+
+var doconvConverterByExtensions = map[string]func(io.Reader) (string, map[string]string, error){
+	"doc":  docconv.ConvertDoc,
+	"docx": docconv.ConvertDocx,
+	"pptx": docconv.ConvertPptx,
+	"odt":  docconv.ConvertODT,
+	"html": func(r io.Reader) (string, map[string]string, error) { return docconv.ConvertHTML(r, true) },
+	// Temporarily disabled to avoid crashes on malicious .pages files
+	// "pages": docconv.ConvertPages,
+	"rtf": docconv.ConvertRTF,
+	"pdf": docconv.ConvertPDF,
+}
+
+func (de *documentExtractor) Name() string {
+	return "documentExtractor"
+}
+
+func (de *documentExtractor) Match(filename string) bool {
+	extension := strings.TrimPrefix(path.Ext(filename), ".")
+	_, ok := doconvConverterByExtensions[extension]
+	return ok
+}
+
+func (de *documentExtractor) Extract(_ context.Context, filename string, r io.ReadSeeker, maxFileSize int64) (out string, outErr error) {
+	defer func() {
+		if r := recover(); r != nil {
+			out = ""
+			outErr = errors.New("error extracting document text")
+		}
+	}()
+
+	extension := strings.TrimPrefix(path.Ext(filename), ".")
+	converter, ok := doconvConverterByExtensions[extension]
+	if !ok {
+		return "", errors.New("unknown converter")
+	}
+
+	// Bound how much data the converter is allowed to read so a small upload
+	// cannot expand into an unbounded amount of in-memory work.
+	var reader io.Reader = r
+	if maxFileSize > 0 {
+		reader = utils.NewLimitedReaderWithError(r, maxFileSize)
+	}
+
+	text, _, err := converter(reader)
+	if err != nil {
+		return "", err
+	}
+
+	return text, nil
+}