Extract text directly from file-object / file-content rather than using filename

[jrkkfst]

Maybe this is already possible?

How would i go about to extract text from the content of a file, rather than reading the file itself? Background is that using the upload component in Dash, one gets the content of the file rather than a pointer to the file location.

Perhaps, by accessing one of the internal functions in textract and specifying an extension this is already possible?

[traverseda]

Unfortunately there's no way to do that, as textract launches external commands (most notably pdftotext) to process files sometimes.