deepmodeling / deepmodeling/dpgen
[Feature Request] _select which step to do in `dpgen run`, and manually do some processes_
- Dominant language
- Python
- Stars
- 399
- Forks
- 198
- Avg merge
- 15d 11h
- Merged PRs (30d)
- 44
Description
**Summary**
How about adding a "skip-stage" like parameter in dpgen run param parameters? When `dpgen run` deal with skipped stages, it just waits and checks the data needed by the next step prepared.
**Detailed Description**
I run dpgen script on a totally different machine from where I run FP calculations. I wonder whether I can do the FP calculation and collect the FP data manually, and just give the prepared result to dpgen again.
The fellow is what I do to get such processes:
- Give fake param and machine parameters about FP steps to let the dpgen shutdown at FP step (do stage 6 but not do stage 7)
- Manually collect the POSCAR prepared by dpgen.
- Run FP processes on another machine.
- Filter and collect FP output data.
- Add collected FP data to the system list and build a new folder to run a new `dpgen run` job.
I have to do such, because 1) I use FP software that dpgen does not support now, and 2) the machine I do FP can not automatically communicate with where I run dpgen script.
I wonder whether `dgen run` can add a param parameter to control which step to skip and wait for data manually prepared.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.