Assembling Records

The assemble subcommand scans a project directory for supported quantum chemistry output files and calculation folders, parses the available calculation data, and stores the results as records in a CHEMSMART database.

Each record is uniquely identified by its record_id. If a newly parsed calculation generates a record_id that already exists in the database, the existing record is updated with the newly parsed data rather than duplicated. This allows assemble to be safely re-run on the same project directory after additional calculations have finished or existing outputs have been updated.

Usage

chemsmart run database assemble [-d path/to/directory] [-p gaussian|orca|xtb]
                                [-i index] [-o outfile.db] [--include-failed]

Options

Option	Type	Description
`-d, --directory`	string	Root directory to scan recursively for supported output files and folders. Defaults to the current working directory.
`-p, --program`	string	Restrict parsing to output files/folders from a specific program. Supported values are `gaussian`, `orca`, and `xtb`. If omitted, outputs from all supported programs are scanned.
`-i, --index`	string	1-based index or slice used to select structures from files containing multiple structures. Defaults to `:` (all structures).
`-o, --output`	string	Output database filename. The `.db` extension is appended automatically if it is not provided. Defaults to `database.db`.
`--include-failed`	flag	Include parseable partial data from failed calculations, which are skipped by default and marked by termination status in the database.

Examples

Assemble all supported calculations under the current directory:

chemsmart run database assemble

Assemble only Gaussian outputs from a project directory:

chemsmart run database assemble -d results/ -p gaussian -o gaussian_results.db

Assemble only xTB calculation outputs from a project directory:

For xTB calculations, assemble recursively scans the root directory to arbitrary depth. Each recognized xTB calculation directory is treated as one record.

Example directory layout:

xtb_results/
├── molecule1_ohess/
├── project_a/
│   ├── molecule2_hess/
│   └── nested/
│       └── molecule3_opt/
└── project_b/
    └── molecule4_sp/

chemsmart run database assemble -d xtb_results/ -p xtb -o xtb_results.db

Every xTB calculation directory must contain exactly one xTB main output together with the required auxiliary files. Directories containing multiple xTB main outputs are ambiguous: CHEMSMART logs an error, skips that directory, and continues assembling other valid calculations.

Assemble only the final structure from each calculation source:

chemsmart run database assemble -d results/ -i -1 -o final_geoms.db