https://github.com/sanzgiri/thinking_tools

Tools for Thinking app, based on Daniel Dennett's book "Intuition Pumps & Other Tools for Thinking"
https://github.com/sanzgiri/thinking_tools

critical-thinking daniel-dennett intuition-pumps mental-models nextjs ollama philosophy react thinking-tools typescript

Last synced: 11 days ago
JSON representation

Tools for Thinking app, based on Daniel Dennett's book "Intuition Pumps & Other Tools for Thinking"

• Host: GitHub
• URL: https://github.com/sanzgiri/thinking_tools
• Owner: sanzgiri
• License: mit
• Created: 2025-12-04T19:22:21.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2026-06-20T07:13:47.000Z (about 1 month ago)
• Last Synced: 2026-06-20T07:22:45.807Z (about 1 month ago)
• Topics: critical-thinking, daniel-dennett, intuition-pumps, mental-models, nextjs, ollama, philosophy, react, thinking-tools, typescript
• Language: Python
• Homepage: https://tools-for-thinking.netlify.app/
• Size: 266 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Dennett's Thinking Tools

An interactive web app presenting all 77 "intuition pumps" and tools for thinking from Daniel C. Dennett's *Intuition Pumps and Other Tools for Thinking* — each with a detailed guide and a practical exercise.

**Live site:** https://tools-for-thinking.netlify.app/

![Screenshot of Dennett's Thinking Tools](docs/screenshot.png)

**Tech stack:** Next.js (App Router) · TypeScript · CSS Modules · Ollama (local LLM for content) · Python (content pipeline) · Netlify (hosting).

## Project Structure

- `dennet.txt`: Source text.

- `slugify.py`: Single source of truth for turning a tool name into a URL slug (kept in sync with `site/src/lib/tools.ts`).

- `process_dennet.py`: Parses `dennet.txt` into `tools.json`, computing the canonical `slug` for each tool.

- `generate_content.py`: Generates detailed markdown content using Ollama (with retries, validation, and atomic writes).

- `add_references.py`: Appends a `## References` section to each tool's markdown, sourced from the "Works cited" list in `dennet.txt`.

- `remove_references.py`: Strips `## References` sections back out of the markdown.

- `main.py`: Pipeline runner that ties the above scripts together (see below).

- `sync_content.sh`: Copies `tools.json` and `content/` into `site/` so the app can read them.

- `content/`: Generated markdown files (one per tool, named `.md`).

- `site/`: Next.js web application.

## Setup

1. **Install Python dependencies**:

   ```bash

   uv sync

   ```

   (Or prefix commands with `uv run ...`.)

2. **Run the content pipeline**:

   `main.py` orchestrates the Python steps. The default (no args) runs `parse` then `generate`.

   ```bash

   uv run python main.py                 # parse + generate (default)

   uv run python main.py parse           # rebuild tools.json only (fast, no Ollama)

   uv run python main.py generate        # generate markdown via Ollama

   uv run python main.py references      # append References sections

   uv run python main.py clean           # strip References sections

   uv run python main.py parse generate references   # chain steps

   ```

   *Notes:*

   - Content generation uses a local Ollama instance (`mistral:latest` model), so Ollama must be running for the `generate` step.

   - Generation takes time. You can stop and resume it; it skips tools whose `content/.md` already exists. Failed/invalid responses are retried and are not written, so they will be retried on the next run.

3. **Build the site**:

   The site needs `tools.json` and `content/` copied into `site/`. Use the sync script:

   ```bash

   ./sync_content.sh

   cd site

   npm install

   npm run dev     # local preview at http://localhost:3000

   # or

   npm run build   # production build (static export of all 77 tool pages)

   ```

## Deployment

The project is deployed to **Netlify** and configured via `netlify.toml` (these settings come from the file, not the dashboard):

- **Base directory**: `.` (repo root, so the build can run `sync_content.sh`)

- **Build command**: `chmod +x sync_content.sh && ./sync_content.sh && cd site && npm install && npm run build`

- **Publish directory**: `site/.next`

- **Plugin**: `@netlify/plugin-nextjs`

Pushing to the `main` branch on GitHub triggers an automatic build and deploy.

## Notes

- The canonical tool slug is computed once in `process_dennet.py` and stored in `tools.json` (the `slug` field). All downstream code (content generation, references, the site) reads that field rather than recomputing it. `slugify.py` and `slugify()` in `site/src/lib/tools.ts` exist only as fallbacks and must be kept in sync.

- The app uses **Vanilla CSS** (CSS Modules) for styling.

- It features a "Rich Aesthetic" with dark mode, gradients, and glassmorphism.

- The homepage supports **live search** and **category filtering** (client-side, in `site/src/app/ToolBrowser.tsx`).

- Each tool page generates its own SEO metadata (`generateMetadata`) from the tool's name and short description.

- If content for a tool is missing, the tool page displays a "Content Generating..." placeholder with the base description. (With all 77 files generated, this is a fallback that is not normally shown.)

## License

[MIT](LICENSE) © Ashutosh Sanzgiri.