_MultiSourceReader

class scikitplot.corpus._MultiSourceReader(readers)

Chains multiple DocumentReader instances into one stream.

Returned by DocumentReader.create when more than one source is supplied, and by DocumentReader.from_manifest.

Also acts as a context manager ensuring temporary directories from from_url() downloads are cleaned up on exit.

Parameters:

readerslist[DocumentReader]

Ordered list of sub-readers. Documents are yielded in order.

Parameters:

readers (list[DocumentReader])

Notes

Context manager usage (ensures temp-file cleanup):

with DocumentReader.create(
    "https://iris.who.int/.../content",
    Path("report.pdf"),
) as reader:
    docs = list(reader.get_documents())

Duck-typed interface — exposes get_documents() matching DocumentReader so it works anywhere a single reader is accepted.

Examples

>>> from pathlib import Path
>>> import scikitplot.corpus._readers
>>> reader = DocumentReader.create(Path("a.txt"), Path("b.pdf"))
>>> type(reader).__name__
'_MultiSourceReader'
>>> docs = list(reader.get_documents())

close()

Release temporary directories created by from_url() downloads.

Each sub-reader that downloaded a file has a _from_url_tmp_dir attribute set by DocumentReader.from_url. This method deletes those directories. Called automatically when used as a context manager; call manually otherwise.

Return type:

None

get_documents()

Yield all documents from all sub-readers in order.

Yields:

CorpusDocument

Documents from each sub-reader, chained sequentially.

Return type:

Generator[Any, None, None]

property n_readers: int

Number of constituent readers.

Returns:

int

len(self.readers).

readers: list[DocumentReader]