8000 Create Debian/Ubuntu package dataset · Issue #4 · LLNL/dapper · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

Create Debian/Ubuntu package dataset #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
nightlark opened this issue Dec 5, 2024 · 8 comments
Closed

Create Debian/Ubuntu package dataset #4

nightlark opened this issue Dec 5, 2024 · 8 comments
Assignees
@monwen @swest50

Comments

@nightlark
Copy link
Collaborator
nightlark commented Dec 5, 2024 •
edited
Loading

Create a package dataset for Debian/Ubuntu that maps file names to the package(s) that could have installed the files. Relates to #5 and #8 for determining how file names should be normalized to use as a "key" for lookups. We should also consider how we may want to split up the dataset into smaller chunks based on how it will be used (e.g. only includes, only binary files, etc).

Some potential sources of data for this are:

  • SQLite database from linux-package-analyzer
  • Contents-amd64 gzipped file from http://security.ubuntu.com/ubuntu/dists/ for noble or oracular (sources also may have interesting info to add); Debian equivalent is https://ftp.debian.org/debian/dists/stable-updates/main/
    • The package name given isn't the best one (usually contains version information in its name) -- in general the source package name is a better option that is more recognizable; though there are exceptions like libstdc++-12-dev comes from the gcc-12 source package (a few special cases for some well-known libraries like that may be needed.
@monwen monwen self-assigned this Dec 5, 2024
@nightlark
Copy link
Collaborator Author

Update/write script to output results as a sqlite database with an index generated on the file name for fast lookups.

@nightlark
Copy link
Collaborator Author

1/16 - regenerate dataset using normalized file name function for quick lookup file name key (ran into issue with a haskell file name normalization corner case).

@nightlark
Copy link
Collaborator Author

1/23 - Steven updated dataset using the updated/fixed normalization function (fixed in #52)

@nightlark
Copy link
Collaborator Author
nightlark commented Feb 13, 2025 •
edited
Loading

2/13 - Steven generated sqlite database from Debian bookworm Contents file. About 25% of files in Debian don't appear Ubuntu -- not sure if it's Ubuntu lagging behind or choosing not to pull certain packages. Which of the packages those files belong to don't appear in Ubuntu?

  • 3/6 - most "missing" packages seem to be dev related; maybe just lagging behind in synchronizing

Ubuntu has 4x the number of files in Debian.

@nightlark
Copy link
Collaborator Author

Once #6 is done, upload current Ubuntu dataset, consider also uploading Debian dataset.

@nightlark
Copy link
Collaborator Author
nightlark commented Mar 20, 2025 •
edited
Loading

Name ubuntu/debian dataset releases with major version of distro -- ubuntu-24.04.db or ubuntu-noble.db

@nightlark
Copy link
Collaborator Author

@nightlark: create the repos for this on HF and Zenodo.

@nightlark
Copy link
Collaborator Author

Steven uploaded datasets to 3 debian (bookworm, bullseye, buster) and 3 ubuntu (jammy, focal, noble) versions to HuggingFace (https://huggingface.co/dapper-datasets).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@monwen monwen

@swest50 swest50

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

When branches are created from issues, their pull requests are automatically linked.

data_scraping-dev LLNL/dapper
3 participants
@nightlark @monwen @swest50
0