physics-IDE is a Linux-first desktop research environment for developing, comparing, and testing scientific theories in a structured workspace. It combines a Tauri-based desktop shell with a lightweight web frontend and Rust-backed analysis utilities.
The immediate direction for physics-IDE is to focus on a Linux-exclusive build and Debian-based deployment workflow. Windows support is currently deferred while the compilation process remains too complex to maintain effectively.
The next major milestone is version 7: turning physics-IDE into a true imported-project research environment where the AI can understand the state of the theory, the available tools, and the history of analysis work without needing the original human operator to re-explain everything each session.
The version 7 vision is to provide a flexible environment for:
- organizing theory material, notes, equations, and manuscripts in one place;
- ingesting theory sources from different paradigms without rejecting them up front;
- generating a structured master manuscript from imported markdown theory content;
- building a project knowledge index so the AI can navigate chapters, sections, and subsections coherently;
- linking theory content to reusable tools, scripts, experiments, and datasets;
- supporting educational and exploratory workflows across mainstream, hybrid, and non-standard theory families;
- connecting theory content to empirical data and transparent evaluation workflows;
- shipping a polished Linux desktop experience with reliable .deb packaging and installation.
The project has moved from a simple desktop shell toward a more practical theory-development workspace.
- added a built-in Markdown Documents viewer with rendered preview, fuzzy search, and editor launch workflow;
- added single-document PDF export from Markdown Documents with output-directory selection;
- connected Markdown Documents and Manuscript Tools with cross-navigation buttons for rapid workflow switching;
- improved manuscript rendering/export behavior so PDF and DOCX are generated through Pandoc conversion;
- added explicit GitHub Username/PAT settings and aligned markdown save behavior with configured GitHub mode vs local-save mode;
- expanded the in-app Help system with:
- GUI Button Glossary,
- Push/Pull Context and Common Errors,
- Startup Initial Setup Workflow and Checklist;
- completed a UI housekeeping pass with terminology alignment and broad tooltip coverage (including keyboard Enter hints on chat/search flows).
- confirmed reliable Gemini model communication from the desktop app;
- fixed markdown file-opening from the project tree view;
- established the version 7 direction around project-aware AI memory and theory indexing;
- prepared the groundwork for a new Tools menu and manuscript-ordering workflow.
- a Tauri desktop app shell with a configurable interface and integrated terminal;
- settings for project roots, theory directories, master-axiom paths, and AI/provider configuration;
- a master-axiom generation flow that scans theory markdown content and produces a structured draft;
- a theory import pipeline that can ingest a source file and split plain-text manuscripts into markdown sections;
- an initial theory-mode classification layer that recognizes mainstream, hybrid, and left-field-style content;
- regression tests for scientific template generation, theory-style classification, and manuscript import.
- formalizing the version 7 implementation plan for project-aware AI memory;
- adding a Tools menu for future project workflows and app functions;
- building a manuscript composition workflow that orders markdown chapters, sections, and subsections into a master document;
- connecting imported theory structure to reusable tools, experiments, and datasets;
- improving support for diverse theory families and educational use cases.
-
Project knowledge index
- Parse the master manuscript and theory markdown directory into a structured topic tree.
- Generate chapter, section, and subsection summaries that can be used by the AI as a compact navigation layer.
-
Compact project digest
- Create a token-efficient digest file that summarizes the theory corpus, assumptions, tools, and experiments.
- Use this digest as a prompt context layer for AI sessions.
-
Manuscript composition workflow
- Allow the user to reorder imported markdown files into a preferred master-document sequence.
- Support logical sorting for numbered sections and appendix-style files.
- Render a combined markdown document from the chosen order.
- Export the result as Markdown, PDF, or DOCX from a new Tools menu workflow.
- Support an optional AI training export that writes a replacement training artifact for project-aware AI context.
-
Tool and experiment registry
- Add a project-level registry of reusable scripts, notebooks, and analysis tools.
- Track prior experiments and link them to the theory topics they support.
-
AI awareness integration
- Inject the project digest, topic index, and tool registry into the AI briefing pipeline.
- Make the AI prefer existing tools and previous analyses before proposing new ones.
- Support in-thread file attachment so users can provide a selected document or image directly to an AI lane.
-
UI polish and workflow consolidation
- Add the new Tools menu and move version-7 functions into that drop-down as they are introduced.
- Keep the Customize menu focused on configuration paths and app settings.
- Expand the terminal area in the left wing to make the workspace tools more usable.
- Node.js and npm
- Rust toolchain
- Tauri prerequisites for your OS
- Pandoc (required for real PDF and DOCX manuscript/markdown export)
npm install
npm run tauri devcargo test --manifest-path src-tauri/Cargo.tomlThis repository is under active development. The current implementation is intentionally modular so new theory parsers, empirical evaluators, and scientific workflows can be added over time.
- Desktop stabilization toward v6.0.0 is tracked in docs/releases/v6-desktop-checklist.md.
- UI-to-function mapping coverage is tracked in docs/releases/v6-control-map.md.
- Provider API keys entered in settings are stored locally on the device in application config as plain text.
- Keys are used only to send requests to the provider selected in the UI.
- For stronger secret handling, prefer environment variables or an OS keychain-backed workflow.
Version 7 marks a major milestone for physics-IDE.
- Project-aware AI behavior is now stable enough to materially outperform a generic side-by-side browser LLM workflow for in-project theory work.
- The desktop workflow now supports a coherent paradigm for theoretical-physics modeling, iteration, and testing, with AI carrying repetitive context-heavy tasks while the human remains the primary director of theory evolution.
Version 8 is focused on refinement, flexibility, and polish.
- model freedom: user-selected OpenAI and Gemini model IDs without code edits;
- workflow consolidation: migrate mature workflows into the top-menu Tools dropdown to free left/right wing real estate;
- layout control: expand View controls so users can toggle pane elements such as file tree and primer-related surfaces;
- primer simplification: evaluate how much primer work can be automated by project-aware context, including an idea-pad-driven pathway that can append daily notes into primer context;
- UX coherence: keep customization centered on path/location setup while reducing repetitive manual context assembly.
Tracking references:
- docs/releases/v7-release-checklist.md
- docs/releases/v7-release-notes.md
- docs/releases/v8-model-flexibility-plan.md
Use this sequence on a local Ubuntu laptop for a clean production build.
- Install system dependencies
sudo apt update
sudo apt install -y \
build-essential \
curl \
wget \
file \
pandoc \
libgtk-3-dev \
libayatana-appindicator3-dev \
librsvg2-dev \
patchelf \
libwebkit2gtk-4.1-dev- Install Node.js 20 LTS (if not already installed)
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
sudo apt install -y nodejs
node -v
npm -v- Install Rust toolchain (if not already installed)
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
source "$HOME/.cargo/env"
rustc -V
cargo -V- Clone and install project dependencies
git clone https://github.com/GPD-Research/physics-ide.git
cd physics-ide
npm install- Run automated checks before packaging
node --test src/ai-config.test.js
cargo test --manifest-path src-tauri/Cargo.toml- Build production desktop artifacts
npm run tauri -- build- Locate artifacts
- Debian package and related artifacts are produced under:
- src-tauri/target/release/bundle/
- Install local Debian package (if generated)
sudo dpkg -i src-tauri/target/release/bundle/deb/*.deb
sudo apt -f install -y- Launch and smoke-check
- open the installed app;
- verify workspace loading, AI provider settings, and AI Testing modal flows;
- run one context probe and confirm report generation/open-report behavior.
If build issues appear, capture full logs:
npm run tauri -- build > build.log 2>&1
tail -n 120 build.log