Summarize Gaussian, ORCA, and xTB¶
Put outputs from different programs in one batch and export final status and energy through common fields.
Input¶
The .out extension can match both ORCA and xTB. Automatic mode probes the file contents.
Python¶
from molop import AutoParser
batch = AutoParser("results/*", n_jobs=3)
summary = batch.to_summary_df(
frame=-1,
brief=False,
flatten_columns=True,
)
columns = [
"Calc Parameter.Software",
"Calc Parameter.Method",
"Status.IsNormal",
"Energy.total_energy.hartree",
]
result = summary.reindex(columns=columns)
print(result.to_string(index=False))
result.to_csv("qm_summary.csv", index=False)
Output¶
qm_summary.csv contains one row per successfully parsed file. The stable output below uses the
bundled single-file ORCA sample:
Output
For a mixed directory, the result has one row per successfully parsed input and unavailable fields
remain empty. Environment-dependent columns such as absolute file paths should be handled separately;
do not treat them as fixed output. Do not compare total_energy values across unrelated methods or
Hamiltonians without scientific normalization.
To reproduce the displayed row without a mixed directory, use:
sample = AutoParser("water_mp2.out", n_jobs=1)
sample_summary = sample.to_summary_df(
frame=-1,
brief=False,
flatten_columns=True,
)
sample_columns = [
"Calc Parameter.Software",
"Calc Parameter.Method",
"Status.IsNormal",
"Energy.total_energy.hartree",
]
print(sample_summary.reindex(columns=sample_columns).to_string(index=False))
Output
Inspect detected programs¶
Typical output:
See Gaussian log, ORCA output, and xTB output for exact capabilities.