wrossmorrow opened a new issue, #25316:
URL: https://github.com/apache/beam/issues/25316
### What would you like to happen?
LZMA compression is standard in python but not one of the strategies in the
`beam.io.{Read,Write}FromText` `PTransform`s.
[openwebtext](https://skylion007.github.io/OpenWebTextCorpus/), for example,
uses this compression. I think this may be a pretty simple change. For example,
I hacked up a naive "shim"
[here](https://github.com/wrossmorrow/efnlp/blob/main/beam/shims/beam-io-with-lzma-filesystem.py)
for use in Dataflow with a custom container by just overwriting
`apache_beam/io/filesystem.py` in the `site-packages`. It's working (a) locally
with decompression and compression (though the output filenames are malformed,
the part schema follows the compression extension) and (b) in a
`DataflowRunner` reading a GCS dump of all the openwebtext `.xz` archives.
(Without this I've been having a hell of a time getting any horizontal scaling
while reading openwebtext.) It may be this simple, but I haven't run any Beam
tests on these minor changes. I will probably do a bit more research int
o that myself.
### Issue Priority
Priority: 2 (default / most feature requests should be filed as P2)
### Issue Components
- [X] Component: Python SDK
- [ ] Component: Java SDK
- [ ] Component: Go SDK
- [ ] Component: Typescript SDK
- [ ] Component: IO connector
- [ ] Component: Beam examples
- [ ] Component: Beam playground
- [ ] Component: Beam katas
- [ ] Component: Website
- [ ] Component: Spark Runner
- [ ] Component: Flink Runner
- [ ] Component: Samza Runner
- [ ] Component: Twister2 Runner
- [ ] Component: Hazelcast Jet Runner
- [ ] Component: Google Cloud Dataflow Runner
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
