Repository navigation
Conversation
…olerance. Skip preview post_process so add_fix configs no longer require in.lammps, warn on atom overlaps, and keep multi-task property posts alive when individual LAMMPS points fail. Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
is_failed_task_result wrongly treated fixture result_task.json (LabeledSystem) as failed because it is not a plain dict, breaking test_gamma.test_compute_lower. Also harden refine path assert, ticket expireIn test, and add local pre-commit smoke. Co-authored-by: Cursor <cursoragent@cursor.com>
Fix: Update Bohrium ticket expireIn and adaptive DFT k-spacing
…tion(hours) - Change parameter name from 'expireIn' to 'expiration' (correct API field) - Change unit from seconds to hours (correct API unit) - Update default from 604800 to 168 (7 days in hours) - Update SKILL.md documentation accordingly
Co-authored-by: Cursor <cursoragent@cursor.com>
fix(skill): correct ticket API param from expireIn(seconds) to expira…
## OpenAPI Sandbox Integration (config.py, account.py) - Add context_type="OpenAPI" branch for Bohrium OpenAPI Sandbox - TiefblueClient uses access_key directly (no email/password needed) - Support dflow_host: lbg-workflow-dflow.dp.tech, namespace="dflow" - Add dispatcher_image, image_address, machine_type fields - Separate validate_project_id_types() logic for OpenAPI vs legacy ## Bug Fix: Interstitial insert_ele (Interstitial.py) - Normalize insert_ele from string to list at lines 38-43 - Prevents "Cu" being iterated as ["C","u"] -> ValueError ## generate_config.py Enhancements - Add build_global_json_sandbox() for OpenAPI Sandbox mode - LAMMPS run_command defaults to "lmp -in in.lammps" (not mpirun) - GPU machine_type auto-selection for deepmd/mace/nep potentials - Backend-conditional image_address (LAMMPS/ABACUS/VASP) - ABACUS INPUT template: relax_nmax=50, cal_stress=1, cal_force=1 - Adaptive kspacing based on system size - insert_ele list normalization in post-processing Tested: Cu FCC 1-atom cell, ABACUS all-properties, LAMMPS+DPA-3.2 GPU all-properties
feat: add OpenAPI Sandbox support and fix critical bugs
integrate validated DPA4 phonolammps runtime
feat: default GPU workflows to NVIDIA L20
Zikkying
changed the title
Remove redundant development environments and test artifacts, flatten the pinned scientific base, and reject runtime filesystems at 35 GB. Preserve model hashes and runtime package versions. CPU audit measured 26.90 GB before model payload; registry build and fresh L20 qualification remain pending.
Retain Miniconda shared libraries required by native LAMMPS while removing its redundant Python environment. The trimmed L20 sandbox measured 27.94 GB and completed FeCoNi single-point and phonoLAMMPS checks. Final published-image qualification remains pending.
Enable pinned tongqi runtime and correct skill execution guidance
Enable DPA4 tongqi OpenAPI after real L20 end-to-end acceptance
Prevent unresolved type maps and document tongqi runtime requirements
Release APEX 1.3.4 runtime images and synchronized skill
Synchronize runtime image documentation and exported skill
Streamline APEX skill workflow and debugging guide
Validate Gamma preview interface composition and fix-xy geometry
Guide MM preview and preflight through the APEX CPU runtime
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
APEX cloud jobs reuse a working image runtime instead of upgrading apex-flow on every run. This PR also adds two-phase melting workflows, staged finite-temperature VASP execution, OpenAPI Sandbox support, and stronger Gamma/phonon/input validation.
1 * NVIDIA L20_48g.Validation
images/dpa4-alloytongqi-l20/pre-snapshot-evidence.json.L20 publication blocker
The tongqi production profile remains
pre_snapshot_only. Bohrium snapshots 178312 and 178321 from the L20 sandbox, and 178333 from a clean CPU assembly, returned status 3 with empty build logs. An immutable image digest, a fresh L20 image qualification, and a real job-SKU smoke are still required. This commit is not a new PyPI or production skill release.Previously published runtime
apex-flow==1.3.2.post1, published from ec6e00f.registry.dp.tech/dptech/dp/native/prod-397953/397637/apex-flow:1.3.2-post1.PIP_NO_INDEX=1.