media().writeData() filename formatting

hmmm · October 17, 2020, 2:07pm

Just wondering where I can find information about how filenames are formatted by media().writeData(). I understand the value it returns is a filename, so I assume it sanitises the string somehow. More to the point, is there a way to sanitise a string the same way without writing to media?

kelciour · October 18, 2020, 9:02pm

It’s all in the source code

If it’s just a raw data, it’ll be formatted with its hash as its filename.

github.com

ankitects/anki/blob/1aa29f2e79e8451a7abc90feac02ea9b602fb129/qt/aqt/editor.py#L812-L817


# ext should include dot
def _addPastedImage(self, data: bytes, ext: str) -> str:
    # hash and write
    csum = checksum(data)
    fname = "{}-{}{}".format("paste", csum, ext)
    return self._addMediaFromData(fname, data)

github.com

ankitects/anki/blob/603210149c830a9c59f60fd27ccd7d62e4076af9/pylib/anki/utils.py#L220-L223


def checksum(data: Union[bytes, str]) -> str:
    if isinstance(data, str):
        data = data.encode("utf-8")
    return sha1(data).hexdigest()

Otherwise, it’s quite complicated. Everyting is done in Rust and at the moment, as far I can see, is not exposed and can’t be used with Python.

Take a look at rslib/src/media/files.rs. The whole file with tests in the end is pretty much about filename formatting.

github.com

ankitects/anki/blob/3f403040db128617cf86ce76d0f12ec4c938e236/rslib/src/media/files.rs#L151-L153


/// Write desired_name into folder, renaming if existing file has different content.
/// Returns the used filename.
pub fn add_data_to_folder_uniquely<'a, P>(

In Anki 2.1.15 it was done in Python. Take a look at

github.com

ankitects/anki/blob/2.1.15/anki/media.py#L148


        return True

# Adding media
##########################################################################
# opath must be in unicode

def addFile(self, opath):
    with open(opath, "rb") as f:
        return self.writeData(opath, f.read())

def writeData(self, opath, data, typeHint=None):
    # if fname is a full path, use only the basename
    fname = os.path.basename(opath)

    # if it's missing an extension and a type hint was provided, use that
    if not os.path.splitext(fname)[1] and typeHint:
        # mimetypes is returning '.jpe' even after calling .init(), so we'll do
        # it manually instead
        typeMap = {
            "image/jpeg": ".jpg",
            "image/png": ".png",

hmmm · October 24, 2020, 12:16am

Thanks for the comprehensive and considerate response !

Topic		Replies	Views
Bug on Anki 2.1.60 (awesometts and customize shortcuts?) Add-ons	0	269	June 5, 2023
2.1.46 Release Candidate Beta Testing	9	554	August 4, 2021
A rusty regex. Has the Browser RegEx search changed? Beta Testing	3	383	February 1, 2023
Colpkg can't import due to non-UTF-8 encoding? Anki Desktop	3	623	May 1, 2023
Getting raw field data in custom TTS implementation Development	7	471	May 1, 2023

media().writeData() filename formatting

Related Topics