allenai / allenai/cached_path

Support extracting files from RPM archives.

未关闭
#205 1 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看
Feature request
主要语言
Python
星标
47
派生
22
PR 合并指标
30 天内没有已合并 PR

描述

I use cached_path for downloading executable dependencies. This works perfectly on Windows; on Linux however, some dependencies are only available as RPM files. Extracting those is not supported by cached_path yet.

The feature could be implemented using the [rpmfile](https://pypi.org/project/rpmfile/) module. The following steps would be necessary for this:
* Implement a function `is_rpmfile` that tries to parse the file and is missing from the rpm module linked above (that step could (should?) be done in the repo of the above module, which is available at https://github.com/srossross/rpmfile.
* Use the `is_rpmfile` in the checks at https://github.com/allenai/cached_path/blob/0d250ae06ca48b014cd8f5079ca6165beca58311/cached_path/_cached_path.py#L181 and https://github.com/allenai/cached_path/blob/0d250ae06ca48b014cd8f5079ca6165beca58311/cached_path/_cached_path.py#L201
* Add RPM extraction code after the block at https://github.com/allenai/cached_path/blob/0d250ae06ca48b014cd8f5079ca6165beca58311/cached_path/_cached_path.py#L245
* It might also be sensible to include a test verifying the extraction code into https://github.com/allenai/cached_path/blob/main/tests/cached_path_test.py.

I'm currently working around this by explicitly checking for RPM files when using cached_path and extracting the cached RPMs. It would be much cleaner to rely on `cached_path` for this however :)

I could write and PR this code if you're interested. Let me know if that's the case.

贡献指南

打开贡献指南

评估

这个 Issue 还没有评估数据。

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。