fix: support mdx files close #11557 (#11565)

Signed-off-by: yihong0618 <zouzou0208@gmail.com>
This commit is contained in:
yihong
2024-12-12 13:37:56 +08:00
committed by GitHub
parent e565ecdaef
commit 36cb25b341
5 changed files with 7 additions and 6 deletions

View File

@@ -103,7 +103,7 @@ class ExtractProcessor:
extractor = ExcelExtractor(file_path)
elif file_extension == ".pdf":
extractor = PdfExtractor(file_path)
elif file_extension in {".md", ".markdown"}:
elif file_extension in {".md", ".markdown", ".mdx"}:
extractor = (
UnstructuredMarkdownExtractor(file_path, unstructured_api_url, unstructured_api_key)
if is_automatic
@@ -141,7 +141,7 @@ class ExtractProcessor:
extractor = ExcelExtractor(file_path)
elif file_extension == ".pdf":
extractor = PdfExtractor(file_path)
elif file_extension in {".md", ".markdown"}:
elif file_extension in {".md", ".markdown", ".mdx"}:
extractor = MarkdownExtractor(file_path, autodetect_encoding=True)
elif file_extension in {".htm", ".html"}:
extractor = HtmlExtractor(file_path)