[BUG] Dpdata failed to process `abacus/md` (the `develop` branch of `abacus`)

[ZhouXY-PKU]

Bug summary

The energy keyword in running_md.log have changed from final_etot_is to #TOTAL ENERGY# in the develop branch of abacus, so dpdata/abacus/md.py cannot process the log files successfully.

dpdata Version

1.0.1

Input Files, Running Commands, Error Log, etc.

sys = dpdata.LabeledSystem("../",fmt="abacus/pw/md")

File "/home/dpdata/abacus/md.py", line 188, in get_frame
energy = get_energy(outlines, ndump, dump_freq)
File "/home/dpdata/abacus/md.py", line 150, in get_energy
assert ndump == len(energy), (
AssertionError: Number of total energies in running_md.log = 0. Number of frames in MD_dump = 20. Please check.

Steps to Reproduce

Just use a running_md.log of develop branch of abacus to reproduce.

Further Information, Files, and Links

No response

[dosubot]

Hi @ZhouXY-PKU! I'm Dosu and I’m helping the dpdata team.

Thanks for reporting this! You've identified a confirmed issue - the MD parser in dpdata only looks for "final etot is" in dpdata/abacus/md.py, while the ABACUS develop branch now uses "TOTAL ENERGY".

Interestingly, PR #909 already added support for both keywords in the SCF parser (dpdata/abacus/scf.py), but the MD parser was not updated at the same time.

The fix would be to modify the get_energy function in dpdata/abacus/md.py to check for both keywords:

if "final etot is" in line or "TOTAL ENERGY" in line:

This would match how the SCF parser handles both ABACUS LTS and develop versions.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

  

[ZhouXY-PKU]

@dosu Thanks for your advice, PR #950 is ready for check now.

[dosubot]

You're welcome! Great work on submitting PR #950 so quickly. The maintainers should be able to review it from here. Thanks for contributing the fix! 🎉

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

  

[ZhouXY-PKU]

@dosu Actually, abacus/relax.py has the same issue of energy keyword as the format of running_cell-relax.log of the develop branch of abacus has also been modified.

[dosubot]

Good catch! You're right - dpdata/abacus/relax.py also only checks for "final etot is" and doesn't support the "TOTAL ENERGY" keyword.

Currently only the SCF parser supports both keywords. Both the MD and relax parsers need the same fix. Would you like to add the "TOTAL ENERGY" support to relax.py as part of PR #950 as well, or open a separate PR for that?

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}