gaussian-dsh-plugin: A Gaussian 16 Compute Assistant for DeepSeek Harness
yangdw2024/gaussian-dsh-plugin
A DSH plugin that drives Gaussian 16 from natural language: builds inputs from SMILES, submits serial jobs, parses energies and orbitals, and renders figures via Multiwfn/VMD.
gaussian-dsh-plugin is a Gaussian 16 compute assistant that runs inside DeepSeek Harness, linking modeling, submission, parsing and plotting into one natural-language-driven flow. It targets the pain of scattered quantum chemistry steps and constant software switching, letting users describe a need and go from a SMILES string or ChemDraw file all the way to result analysis. Core abilities include RDKit-based 3D structure and Gaussian input generation, serial job management, parsing of energies and HOMO-LUMO data, and asynchronous Multiwfn and VMD rendering of orbitals, ESP surfaces and weak-interaction maps, across 19 tools.
How to Install
dsh plugin --profile <your-profile> add @yangdw2024/gpReplace --profile with the profile name you actually run (see ls ~/.dsh/profiles/): usually web in the browser UI, desktop or desktop-NNN in the desktop app.
- Category
- Tools & Capabilities
- Platform
- DSH Plugin
- Author
- yangdw2024
- Distribution
- Plugin
gaussian-dsh-plugin Key Features
gaussian-dsh-plugin Repo Summary
What Does It Do?
Gaussian computational assistant is a DSH plugin for DeepSeek Harness that chains Gaussian 16, RDKit, Multiwfn and VMD into an AI-callable toolchain. It removes the manual grind of quantum chemistry workflows: starting from a SMILES string or ChemDraw file, it builds the 3D geometry, generates Gaussian input, submits jobs serially, parses the output, and drives post-processing. Maintained by yangdw2024 under the MIT license, it ships RDKit and OpenBabel inside its backend, so target machines need no separate Python setup.
Core Features
- Input generation: turns SMILES or ChemDraw files into 3D structures via RDKit and produces validated Gaussian input files.
- Job management: submits, queries, lists and stops Gaussian jobs serially to avoid crashing multi-core machines.
- Output parsing: extracts energy, convergence, HOMO-LUMO gaps, dipole, frequencies and thermodynamic quantities from .out/.log files.
- Post-processing: asynchronously calls Multiwfn and VMD to render HOMO/LUMO orbital plots, ESP surfaces and RDG/NCI weak-interaction maps.
- Helper queries: online basis-set library search, Gaussian keyword lookup and common error diagnosis, across 19 tools in total.
How to Use This Plugin?
After enabling it in DSH, just describe what you need in natural language, such as optimizing a molecule, and the plugin runs modeling, input generation, submission and parsing on its own. On first use, ask DSH to check backend status and probe post-processing capabilities to confirm Multiwfn and VMD are ready. Because default paths point to the author's own machine, override the Gaussian, formchk, Multiwfn and VMD executable paths plus the working directory through environment variables, then fully quit and restart DSH. An 8-core CPU with 16 GB RAM is recommended, and jobs should stay serial.
How to Troubleshoot This Plugin?
If the backend fails to start, check whether the install path is too long: DSH derives a snapshot directory from the package name, and deep paths can hit the Windows 260-character limit, so deploy under a short path. When the status check returns an error, the environment variables likely still point at the author's paths; follow the official instructions to override each executable path and fully restart DSH. If post-processing plots are unavailable, use the capability probe to confirm Multiwfn and VMD are installed and ready.
This page is an independent rewrite of the plugin's official README — for authoritative documentation and the latest changes, refer to the source: yangdw2024/gaussian-dsh-plugin. The plugin is third-party code that runs on your machine once installed; inclusion does not imply endorsement — please review the source before installing.
DSH Plugin FAQ
