clean up repo to reduce size

[aaravgarg]

there are too many large files (eg. photos & videos) in the repo

these need to be stored in a better and optimized way coz rn the repo takes forever to clone

[jagdish-15]

Hello @aaravgarg,

I've worked out how to significantly reduce the repository size using Git LFS with the following file patterns being tracked:

*.jpeg filter=lfs diff=lfs merge=lfs -text
*.jpg filter=lfs diff=lfs merge=lfs -text
*.png filter=lfs diff=lfs merge=lfs -text
*.mp4 filter=lfs diff=lfs merge=lfs -text
*.bin filter=lfs diff=lfs merge=lfs -text
*.mov filter=lfs diff=lfs merge=lfs -text
*.dill filter=lfs diff=lfs merge=lfs -text
*.gif filter=lfs diff=lfs merge=lfs -text
*.elf filter=lfs diff=lfs merge=lfs -text

I have reworked the entire git history to apply this properly. Without history rework, there is no effective way to reduce the repository size.
As a result, the clone size has been reduced from 607 MB to a mere 132 MB.

Here is proof of the cloning process output:

Cloning into 'omi-lfs-clean'...
remote: Enumerating objects: 66576, done.
remote: Counting objects: 100% (6/6), done.
remote: Compressing objects: 100% (6/6), done.
remote: Total 66576 (delta 0), reused 1 (delta 0), pack-reused 66570 (from 1)
Receiving objects: 100% (66576/66576), 132.80 MiB | 6.39 MiB/s, done.
Resolving deltas: 100% (44659/44659), done.
Updating files: 100% (3464/3464), done.

---

I’ve also updated the documentation (contribution.mdx) to include how to clone the repo without downloading all LFS files by default, by adding the following at the end of the file:

Cloning Repository Without Downloading All LFS Files

Because some files tracked by Git LFS can be very large, cloning the full repo may take a long time or consume lots of bandwidth.

To avoid downloading all large files immediately, you can clone the repository with only LFS pointer files (placeholders) using:

GIT_LFS_SKIP_SMUDGE=1 git clone https://github.com/BasedHardware/Omi.git

Then, when you need the actual large files, run:

git lfs pull

To download only specific large files or directories, use:

git lfs pull -I "path/to/large/file1" -I "path/to/large/dir/*"

Replace "path/to/large/file1" and "path/to/large/dir/*" with the actual paths you want.

This keeps your initial clone lightweight and lets you fetch only the LFS files you need.

---

All these changes are hosted in a non-fork repository because GitHub does not allow pushing rewritten history and new LFS objects to a public fork.

Here is the cleaned and restructured repository:
🔗 https://github.com/jagdish-15/omi-lfs-clean

Unfortunately, GitHub doesn’t allow PRs from non-fork repos!

Also note: Accepting this history rewrite means all existing forks will need to be re-synced, and contributors will have to:

• Re-clone the repo OR
• Rebase and clean their local clones
• Force push their changes (if needed) after cleanup

Let me know how we can go ahead with this! I'm happy to coordinate or adjust as needed.