gedent

gedent is an input generator for computational chemistry workflows. It combines a cascading configuration system with Tera templates to generate input files for quantum chemistry software such as ORCA, Gaussian, XTB, ADF, NWChem, and others.

gedent stands for gerador de entradas — Portuguese for "input generator".

Installation

Requirements

• Rust 1.70 or later

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

From crates.io

cargo install gedent

From source

git clone https://github.com/caprilesport/gedent.git
cd gedent
cargo build --release
# binary at target/release/gedent

Configuration

gedent uses a cascading config system. When you run gedent gen, it walks up from the current directory looking for gedent.toml files, then merges them with the global ~/.config/gedent/gedent.toml. Deeper files win key-by-key — the global config sets defaults, project configs override them.

Config file structure

[gedent]
default_extension = "inp"   # output file extension
software = "orca"           # default software (used for template disambiguation)

[model]
method = "pbe0"
basis_set = "def2-tzvp"
charge = 0
mult = 1
dispersion = "d3bj"
solvent = "water"
solvation_model = "smd"     # smd | cpcm | alpb | cosmo | ...

[resources]
nprocs = 8
mem = 3000                  # MB per core in case of orca

[parameters]
# Arbitrary key-value pairs available in templates as Tera variables.
# Useful for software-specific or job-specific settings.
maxiter = 500
frozen_atoms = [1, 2, 3]

Useful config commands

gedent config print                # show merged config
gedent config print --location     # show per-file contributions + merged result
gedent config edit                 # open nearest local gedent.toml in $EDITOR
gedent config edit --global        # open ~/.config/gedent/gedent.toml
gedent init                        # create a gedent.toml in the current directory

One-off overrides with --var

Any context variable can be overridden for a single run without editing config:

gedent gen scan --var frozen_atoms="[20, 28]" --var nsteps=20 mol.xyz
gedent gen sp --method b3lyp --basis-set def2-svp mol.xyz

Values after --var are parsed as TOML literals, so integers, booleans, and arrays work naturally.

Templates

Templates live in ~/.config/gedent/templates/<software>/<jobtype> and are rendered with Tera, a Jinja2-like engine.

Frontmatter

Each template starts with a Tera comment block that declares its metadata:

{#
software = "orca"
jobtype = "sp"
requires = ["method", "basis_set", "charge", "mult", "nprocs", "mem", "Molecule"]
description = "Single point energy"
#}

• requires — variables that must be present in context before rendering. gedent reports a clear error listing what is missing.
• software and jobtype — used by the template picker and the workflow layer.

Available context variables

All keys from [model], [resources], and [parameters] are injected into the Tera context. Key names match the TOML keys exactly:

Variables are only present if they were set — use {% if x is defined %} before referencing optional ones.

The Molecule object

When an xyz file is provided, a Molecule is injected into context with:

• Molecule.description — comment line from the xyz file
• Molecule.atoms — list of { element, x, y, z } atom objects

Built-in Tera functions

gedent registers these functions in addition to Tera's built-ins:

All index arguments are 1-based.

Template example

{#
software = "orca"
jobtype = "sp"
requires = ["method", "basis_set", "charge", "mult", "nprocs", "mem", "Molecule"]
description = "Single point energy"
#}
! {{ method }} {{ basis_set }}{% if dispersion is defined %} {{ dispersion }}{% endif %}{% if solvation %} SMD({{ solvent }}){% endif %}

%pal
 nprocs {{ nprocs }}
end

%maxcore {{ mem }}

*xyz {{ charge }} {{ mult }}
{{ print_coords(molecule = Molecule) }}
*

Template commands

gedent template list               # list all available templates
gedent template print sp           # print template source
gedent template edit orca/opt      # open template in $EDITOR
gedent template new mytemplate     # create a new template from a preset

Generating inputs

gedent gen sp mol.xyz              # generate sp.inp (or sp.<default_extension>)
gedent gen orca/sp *.xyz           # generate one file per xyz
gedent gen sp mol.xyz --print      # print to stdout instead of writing a file
gedent gen sp mol.xyz --dry-run    # validate and show what would be written, no output
gedent gen sp mol.xyz --show-context  # dump the full Tera context as JSON
gedent gen sp mol.xyz --ext gjf    # override output extension
gedent gen sp --software gaussian mol.xyz  # override software for template lookup

Validation

Before rendering, gedent runs a validation pipeline and reports all issues at once. Errors abort generation; warnings proceed with output.

Checks performed:

• Charge and multiplicity — electron count parity, physically impossible combinations
• Superposed atoms — error if any two atoms are closer than 0.5 Å; warning if closer than half the sum of their covalent radii
• Missing template variables — clear list of what requires but is absent from context
• Solvation compatibility — e.g. XTB in ORCA requires ALPB solvation
• Composite method variables — warning when basis_set or dispersion are set but the method (e.g. pbeh-3c) carries its own

Shell completion

gedent --generate fish   # or bash, zsh

Pipe the output to the appropriate completion file for your shell. Template names are completable in gedent gen and gedent template subcommands.

Contributing

Contributions are welcome — bug reports, feature requests, and pull requests at github.com/caprilesport/gedent.

Acknowledgments

Built on top of:

• Tera — template engine
• clap — CLI parsing
• color-eyre — error reporting
• serde + toml — config parsing

License

MIT