luciano/markitdown
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

allow plugin

Browse source
This commit is contained in:
rong-xyz 2025-04-23 09:23:59 +00:00
parent 5213b8d22e
commit 87738fd782
2 changed files with 34 additions and 35 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
README.md
View file

@ -23,9 +23,9 @@ While markitdown is a useful tool, its returned content is too text-focused, whi
Install directly from GitHub: Install directly from GitHub:
```bash ```bash
pip install git+https://github.com/pathintegral-institute/markitup.git pip install git+https://github.com/pathintegral-institute/markitup.git@main#subdirectory=packages/markitup
``` ```
```bash ```bash
uv add git+https://github.com/pathintegral-institute/markitup.git uv add git+https://github.com/pathintegral-institute/markitup.git@main#subdirectory=packages/markitup
``` ```

63
packages/markitup/src/markitup/_markitup.py
View file

@ -35,45 +35,44 @@ class MarkItUp:
def __init__( def __init__(
self, self,
config: Config = Config(), config: Config = Config(),
plugins: Optional[Dict[str, DocumentConverter]] = None,
): ):
self.config = config self.config = config
self.accepted_categories = ["text", "image", "audio", "pdf", "docx", "pptx", "xlsx", "xls", "csv", "html"]
self.converters = {
"text": PlainTextConverter,
"image": ImageConverter,
"audio": AudioConverter,
"pdf": PdfConverter,
"docx": DocxConverter,
"pptx": PptxConverter,
"xlsx": XlsxConverter,
"xls": XlsConverter,
"csv": CsvConverter,
"html": HtmlConverter,
}
if plugins:
for plugin_name, converter in plugins.items():
self.converters[plugin_name] = converter
def convert(self, stream: BinaryIO, file_name: str) -> Dict[DocumentConverterResult, StreamInfo]: def convert(self, stream: BinaryIO, file_name: str) -> Dict[DocumentConverterResult, StreamInfo]:
stream_info: StreamInfo = self._get_stream_info(stream, file_name) stream_info: StreamInfo = self._get_stream_info(stream, file_name)
# Deal with unsupported file types # Deal with unsupported file types
try: try:
match stream_info.category: if stream_info.category in self.converters.keys():
case "text": converter = self.converters[stream_info.category](config=self.config)
return PlainTextConverter(config=self.config).convert(stream, stream_info), stream_info return converter.convert(stream, stream_info), stream_info
case "pptx": else:
return PptxConverter(config=self.config).convert(stream, stream_info), stream_info match stream_info.category:
case "pdf": case "ppt":
return PdfConverter(config=self.config).convert(stream, stream_info), stream_info raise UnsupportedFormatException(
case "audio": ".ppt files are not supported, try .pptx instead")
return AudioConverter(config=self.config).convert(stream, stream_info), stream_info case "doc":
case "xlsx": raise UnsupportedFormatException(
return XlsxConverter(config=self.config).convert(stream, stream_info), stream_info ".doc files are not supported, try .docx instead")
case "xls": case "other":
return XlsConverter(config=self.config).convert(stream, stream_info), stream_info raise UnsupportedFormatException(
case "csv": f"{stream_info.magic_type} files are not supported")
return CsvConverter(config=self.config).convert(stream, stream_info), stream_info
case "docx":
return DocxConverter(config=self.config).convert(stream, stream_info), stream_info
case "image":
return ImageConverter(config=self.config).convert(stream, stream_info), stream_info
case "html":
return HtmlConverter(config=self.config).convert(stream, stream_info), stream_info
case _:
match stream_info.category:
case "ppt":
raise UnsupportedFormatException(
".ppt files are not supported, try .pptx instead")
case "doc":
raise UnsupportedFormatException(
".doc files are not supported, try .docx instead")
case "other":
raise UnsupportedFormatException(
f"{stream_info.magic_type} files are not supported")
except FailedConversionAttempt: except FailedConversionAttempt:
raise FileConversionException( raise FileConversionException(
f"Failed to convert file of type {stream_info.magic_type}") f"Failed to convert file of type {stream_info.magic_type}")