Open
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
calvinp0
force-pushed
the
crest_adapter
branch
from
d6a83b4 to
c231a43
Compare
calvinp0
force-pushed
the
crest_adapter
branch
2 times, most recently
from
3e88c36 to
7674f5f
Compare
calvinp0
force-pushed
the
crest_adapter
branch
3 times, most recently
from
9c25f3d to
51368be
Compare
calvinp0
force-pushed
the
crest_adapter
branch
2 times, most recently
from
9be935e to
eab7647
Compare
MockAdapter writes YAML output (not real ESS logs), so determine_ess_status falsely reports convergence failure. Patching _parse_ess_error to return None in TestRunTask and TestWorkerLoop skips the ESS check for these mockter-based tests.
cluster_tsgs() merges methods into method_sources list, not into the method string. Updated test_cluster_tsgs to expect the representative's original method and execution_time. Updated test_as_dict and test_from_dict to include method_sources in expected output.
Updated Arkane validation logic to provide more granular feedback when atom energy corrections (AEC) or bond additivity corrections (BAC) are missing. Key changes include: - Added `check_arkane_aec` to verify atom energy corrections independently when BAC is disabled. - Enhanced `check_arkane_bacs` to specifically identify and log whether AEC, BAC, or both are missing from the RMG database. - Improved log reporting to distinguish between PBAC and MBAC types.
Updated the energy correction retrieval logic to perform independent fuzzy matching for AEC and BAC keys. This ensures corrections can be retrieved even if they are stored under slightly different level-of-theory definitions in the RMG database (e.g., one including the software attribute and the other not). Key changes include: - Modified `_get_energy_corrections` to find `aec_key` and `bac_key` independently. - Added specific search ranges for PBAC and MBAC sections within the database files. - Updated the `get_qm_corrections.py` script interface to handle separate keys for atom and bond corrections. Remove unused import "get_arkane_model_chemistry"
Updated the logic for determining the Arkane level of theory to provide logging regarding its source and integrated AEC validation when BAC is not used. Key changes include: - Added logging to identify if the Arkane level of theory was explicitly set or inferred from the composite method or single point level. - Integrated a call to `check_arkane_aec` to verify atom energy corrections when `bac_type` is not specified.
Updated the logic for retrieving QM corrections to handle separate keys for atom energy corrections (AEC) and bond additivity corrections (BAC). This ensures that corrections can be resolved independently if they are stored under different level-of-theory definitions in the RMG database. Key changes include: - Updated the script to process `aec_key` and `bac_key` independently. - Maintained backward compatibility by using `matched_key` as a fallback for `aec_key`. - Modified BAC retrieval to utilize the dedicated `bac_key`.
Updated Arkane validation logic to provide more granular feedback when atom energy corrections (AEC) or bond additivity corrections (BAC) are missing. Key changes include: - Added `check_arkane_aec` to verify atom energy corrections independently when BAC is disabled. - Enhanced `check_arkane_bacs` to specifically identify and log whether AEC, BAC, or both are missing from the RMG database. - Improved log reporting to distinguish between PBAC and MBAC types. Enhance Arkane AEC and BAC validation logging and error handling Updated the validation logic for Arkane energy corrections to handle potential input errors and provide more granular feedback when entries are missing. Key changes include: - Added exception handling to catch cases where Arkane quantum corrections data cannot be loaded from the RMG database. - Improved logging to specifically identify the matched BAC key and correction type (PBAC or MBAC). - Added a general warning message when no matching Arkane entry is found for a given level of theory.
Updated the energy correction retrieval logic to perform independent fuzzy matching for AEC and BAC keys. This ensures corrections can be retrieved even if they are stored under slightly different level-of-theory definitions in the RMG database (e.g., one including the software attribute and the other not). Key changes include: - Modified `_get_energy_corrections` to find `aec_key` and `bac_key` independently. - Added specific search ranges for PBAC and MBAC sections within the database files. - Updated the `get_qm_corrections.py` script interface to handle separate keys for atom and bond corrections. Remove unused import "get_arkane_model_chemistry"
Updated the logic for determining the Arkane level of theory to provide logging regarding its source and integrated AEC validation when BAC is not used. Key changes include: - Added logging to identify if the Arkane level of theory was explicitly set or inferred from the composite method or single point level. - Integrated a call to `check_arkane_aec` to verify atom energy corrections when `bac_type` is not specified.
Updated the logic for retrieving QM corrections to handle separate keys for atom energy corrections (AEC) and bond additivity corrections (BAC). This ensures that corrections can be resolved independently if they are stored under different level-of-theory definitions in the RMG database. Key changes include: - Updated the script to process `aec_key` and `bac_key` independently. - Maintained backward compatibility by using `matched_key` as a fallback for `aec_key`. - Modified BAC retrieval to utilize the dedicated `bac_key`. Support backward compatibility for BAC keys in the QM corrections script Updated the logic for retrieving QM corrections to handle legacy input formats where atom and bond corrections are not defined as independent keys. Key changes include: - Maintained backward compatibility by using `matched_key` as a fallback for `bac_key`, mirroring the existing behavior for `aec_key`.
In the Scheduler, only break after conformer troubleshooting if new jobs (conf_opt or conf_sp) are actually running for that species. This ensures the scheduler correctly falls through to the "all conformers done" check if troubleshooting was attempted but failed to launch new tasks.
Don't trigger a resubmission if "fresh" pending tasks (attempt_index == 0) still exist. The presence of fresh tasks indicates that some workers from the initial submission are still queued in the HPC scheduler; these workers will pick up both fresh and retried tasks once they start, making a new submission unnecessary.
DLPNO methods are incompatible with monoatomic species. This update generalizes the previous H-atom-specific check to all monoatomic species, substituting the DLPNO method with its canonical equivalent and logging a warning. It also adds a monoatomic status flag to the scheduler's job metadata.
Generalize the H-atom-specific check to all monoatomic species when using DLPNO methods in Orca, as these methods are incompatible with single-atom systems that lack electron pairs to correlate.
DLPNO methods are incompatible with monoatomic species. This change generalizes the previous hydrogen-specific check to all monoatomic species and automatically falls back to the canonical method by stripping the 'dlpno-' prefix.
When a TS guess fails validation (e.g., NMD check), switch_ts picks the next guess but previously left stale state behind:
1. IRC species from the invalidated guess were never cleaned up. delete_all_species_jobs('TS0') only deletes jobs under the TS0 label, but IRC species like IRC_TS0_1 are separate entries in running_jobs/species_dict/etc. These orphaned species
continued running in parallel with the new guess, potentially interfering with job processing.
2. job_types flags (freq, sp, opt) were never reset. After guess N's freq completed, job_types['freq'] = True carried over to guess N+1, causing the scheduler to skip re-running freq for the new geometry.
3. convergence was never reset to None.
4. The old line self.output[label]['geo'] = ... wrote to the wrong dict level (top-level keys instead of self.output[label]['paths']), making it dead code.
5. Pending pipe batches from the old guess were never discarded.
Passing server_job_ids allows the pipe coordinator to monitor the status of jobs on the server, facilitating job cancellation handling as indicated by the branch name.
Moving task spec reading and scratch directory creation into the try-except block ensures that initialization errors are properly caught. This allows the worker to mark the task as failed instead of leaving it stuck in a RUNNING state.
If the underlying scheduler job for a pipe run is no longer alive, any CLAIMED or RUNNING tasks are immediately marked as orphaned and PENDING tasks are cancelled. This ensures the pipe can reach a terminal state and doesn't hang indefinitely when its workers are lost. .
Use server job IDs to check if a pipe's scheduler job is still in the cluster queue. This allows the coordinator to identify when workers have been lost so that orphaned or stuck tasks can be cleaned up immediately. The implementation includes logic to match both standard and array job ID formats (e.g., Slurm/PBS). Fixes for SLURM
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
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.
Addition of CREST Adapter that complements the heuristic adapter.
This pull request adds support for the CREST conformer and transition state (TS) search method to the ARC project, along with several related improvements and code cleanups. The most important changes include integrating CREST as a TS search adapter, updating configuration and constants, and enhancing the heuristics TS search logic for better provenance tracking and code clarity.
CREST Integration:
JobEnum(arc/job/adapter.py), included CREST in the list of adapters and RMG family mapping, and registered it as a default incore adapter (arc/job/adapters/common.py,arc/job/adapters/ts/__init__.py). [1] [2] [3] [4]arc/job/adapters/ts/crest_test.py).Makefile). [1] [2]Constants and Configuration:
angstrom_to_bohrconversion constant to both Cython and Python constants modules (arc/constants.pxd,arc/constants.py). [1] [2] [3]Heuristics TS Search Enhancements and Refactoring:
arc/job/adapters/ts/heuristics.py). [1] [2] [3] [4]arc/job/adapters/ts/heuristics.py). [1] [2] [3] [4] [5] [6] [7].