Attention

June 2024 Status Update: Removing DataPipes and DataLoader V2

We are re-focusing the torchdata repo to be an iterative enhancement of torch.utils.data.DataLoader. We do not plan on continuing development or maintaining the [DataPipes] and [DataLoaderV2] solutions, and they will be removed from the torchdata repo. We’ll also be revisiting the DataPipes references in pytorch/pytorch. In release torchdata==0.8.0 (July 2024) they will be marked as deprecated, and in 0.9.0 (Oct 2024) they will be deleted. Existing users are advised to pin to torchdata==0.8.0 or an older version until they are able to migrate away. Subsequent releases will not include DataPipes or DataLoaderV2. Please reach out if you suggestions or comments (please use this issue for feedback)

ParagraphAggregator

class torchdata.datapipes.iter.ParagraphAggregator(source_datapipe: ~IterDataPipe[~Tuple[str, ~T_co]], joiner: ~Callable = <function _default_line_join>)

Aggregates lines of text from the same file into a single paragraph (functional name: lines_to_paragraphs). Specifically, this accepts a DataPipe consisting of tuples of a file name and a line. For each tuple, it checks if the file name matches the file name from the previous tuple. If yes, it joins the current line with existing paragraph. If the file names do not match, the existing paragraph is yielded and a new paragraph starts.

Parameters:

• source_datapipe – a DataPipe with tuples of a file name and a line

• joiner – a function that joins a list of lines together

Example

>>> from torchdata.datapipes.iter import IterableWrapper
>>> source_dp = IterableWrapper(
>>>                 [("file1", "Line1"), ("file1", "Line2"), ("file2", "Line2,1"), ("file2", "Line2,2"), ("file2", "Line2,3")]
>>>             )
>>> para_agg_dp = source_dp.lines_to_paragraphs(joiner=lambda ls: " ".join(ls))
>>> list(para_agg_dp)
[('file1', 'Line1 Line2'), ('file2', 'Line2,1 Line2,2 Line2,3')]