Skip to content

Include dependencies needed for the new sequence embedding models #97

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

Merged
julenmendieta merged 2 commits into from
Jun 24, 2026
Merged

Include dependencies needed for the new sequence embedding models #97

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
06121fd
Include dependencies needed for the new sequence embedding models
julenmendieta Jun 24, 2026
d4487b1
Solve CI issue
julenmendieta Jun 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 11 additions & 0 deletions .changeset/torch-cuda-ablang2-sceptr.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
---
'@platforma-open/milaboratories.runenv-python-3.12.10-torch-cuda': minor
---

Add antibody (AbLang2) and TCR (SCEPTR) specialist embedding libraries plus rdkit to the torch-cuda run environment.

- `ablang2==0.2.1` — AbLang2 antibody language model (pulls einops + rotary-embedding-torch)
- `sceptr==1.2.0` — SCEPTR paired-chain TCR model (pulls libtcrlm + tidytcells + pandas + blosum)
- `rdkit==2026.3.3` — cheminformatics toolkit for the PeptideCLM-2 amino-acid→SMILES conversion (pulls Pillow)
Comment thread
vadimpiven marked this conversation as resolved.

All install conflict-free on top of the existing pins (transformers 4.53.2, polars-lts-cpu 1.33.1, numpy 2.2.6, pyarrow 21.0.0, torch 2.7.0); `pip check` clean.
7 changes: 7 additions & 0 deletions checker/whitelists/linux-x64.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -185,6 +185,9 @@
"pynvjitlink_cu12-0.7.0-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl": {
"pynvjitlink._nvjitlinklib": "libcuda.so.1: cannot open shared object file: No such file or directory"
},
"rdkit-2026.3.3-cp312-cp312-manylinux_2_28_x86_64.whl": {
"rdkit.Chem.Draw.rdMolDraw2D": "libXrender.so.1: cannot open shared object file: No such file or directory"
},
"scipy-1.*.whl": {
"scipy.linalg._matfuncs_sqrtm_triu": "cannot import name 'within_block_loop' from partially initialized module",
"scipy.stats._unuran.unuran_wrapper": "module 'numpy.random.bit_generator' has no attribute 'SeedlessSequence'"
Expand All @@ -202,6 +205,10 @@
"torch-2.9.1+cpu-cp312-cp312-manylinux_2_28_x86_64.whl": {
"functorch._C": "initialization failed"
},
"torch-2.12.1+cu126-cp312-cp312-manylinux_2_28_x86_64.whl": {
"functorch._C": "initialization failed",
"torch.lib.libcaffe2_nvrtc": "libcuda.so.1: cannot open shared object file"
},
"torchaudio-2.7.0+cpu-cp312-cp312-manylinux_2_28_x86_64.whl": {
"torchaudio.lib._torchaudio_sox": "libtorchaudio_sox.so: cannot open shared object file: No such file or directory",
"torchaudio.lib.libtorchaudio_sox": "libsox.so: cannot open shared object file: No such file or directory",
Expand Down
5 changes: 4 additions & 1 deletion python-3.12.10-torch-cuda/config.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,10 @@
"transformers==4.53.2",
"polars-lts-cpu==1.33.1",
"numpy==2.2.6",
"pyarrow==21.0.0"
"pyarrow==21.0.0",
"ablang2==0.2.1",
"sceptr==1.2.0",
"rdkit==2026.3.3"
Comment thread
julenmendieta marked this conversation as resolved.
Comment on lines +13 to +15

@greptile-apps greptile-apps Bot Jun 24, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Transitive dependencies not pinned

ablang2, sceptr, and rdkit each pull in several transitive packages (einops, rotary-embedding-torch, libtcrlm, tidytcells, pandas, blosum, Pillow) that are not pinned here. A future pip resolve could silently upgrade any of them and introduce an ABI or API mismatch with the pinned numpy==2.2.6 or torch==2.7.0 in this same environment. Consider adding explicit pins for the key transitives — especially pandas (which has historically broken APIs across minor versions) — either directly in this array or via a constraints.txt / lock file.

Prompt To Fix With AI 
This is a comment left during a code review.
Path: python-3.12.10-torch-cuda/config.json
Line: 13-15

Comment:
**Transitive dependencies not pinned**

`ablang2`, `sceptr`, and `rdkit` each pull in several transitive packages (`einops`, `rotary-embedding-torch`, `libtcrlm`, `tidytcells`, `pandas`, `blosum`, `Pillow`) that are not pinned here. A future `pip` resolve could silently upgrade any of them and introduce an ABI or API mismatch with the pinned `numpy==2.2.6` or `torch==2.7.0` in this same environment. Consider adding explicit pins for the key transitives — especially `pandas` (which has historically broken APIs across minor versions) — either directly in this array or via a `constraints.txt` / lock file.

How can I resolve this? If you propose a fix, please make it concise.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

Fix in Claude Code

],
"platformSpecific": {
"linux-x64": {
Expand Down
Loading