Skip to content

AGENTS.md

Latest commit

 

History

History
89 lines (65 loc) · 3.7 KB

AGENTS.md

File metadata and controls

89 lines (65 loc) · 3.7 KB

AGENTS.md - AI Agent Instructions for dsprep.com

Project Overview

This is dsprep.com, a MkDocs Material static site for data science interview preparation. It is owned by SourceStrongAI (sourcestrongai.com), maintained by Kuldeep Singh Sidhu.

Local Development

  • Conda environment: Use conda run -n mk for all mkdocs commands
  • Dev server: conda run -n mk mkdocs serve (runs on http://127.0.0.1:8000)
  • Deploy: conda run -n mk mkdocs gh-deploy (builds and pushes to gh-pages branch)
  • No CI/CD workflow — deployment is manual via mkdocs gh-deploy

Project Structure

mkdocs.yml                  # Site config, nav, plugins, theme
docs/                       # All site content (markdown, notebooks, assets)
  index.md                  # Homepage
  flashcards.md             # Flashcards app mount point
  contact.md                # Contact page
  privacy.md                # Privacy policy
  Contribute.md             # Contribution guide
  Interview-Questions/      # 15 topic files (markdown with Q&A format)
  Cheat-Sheets/             # Reference sheets (markdown + Jupyter notebooks)
  Machine-Learning/         # ML topic explainers
  Online-Material/          # External resource links
  javascripts/              # JS assets (flashcards.js, mathjax.js, xfile.js)
  stylesheets/extra.css     # Custom CSS overrides
overrides/main.html         # Template overrides (announcement bar, analytics, ads)
hooks/generate_questions_json.py  # Build hook: extracts Q&A to questions.json
site/                       # Built output (do not edit directly)

Deployment

  • Source branch: master
  • Deploy branch: gh-pages (GitHub Pages serves from here)
  • CNAME: dsprep.com
  • After pushing to master, you must run conda run -n mk mkdocs gh-deploy to publish changes

Interview Questions Format

Questions in docs/Interview-Questions/*.md follow this structure:

### Question Title

**Difficulty:** <emoji> Level | **Tags:** `Tag1`, `Tag2` | **Asked by:** Company1, Company2

??? success "View Answer"

    Answer content (4-space indented)
    Supports markdown, code blocks, LaTeX math

The build hook (hooks/generate_questions_json.py) parses these on on_post_build and generates site/assets/questions.json (currently 915 questions) for the flashcards feature.

Flashcards System

  • Page: docs/flashcards.md — minimal HTML mount point
  • Logic: docs/javascripts/flashcards.js — 800+ line client-side app
  • Data: site/assets/questions.json — generated at build time by the hook
  • Features: topic filtering (localStorage), card flip (CSS 3D), shuffle, MathJax support

Announcement Bar

Located in overrides/main.html. Rotating slideshow cycling through:

  • Flashcards feature promotion
  • SourceStrongAI products (ProofMD.ai, ReplyRunner, SEO Audit, OnlineToolsVault, TrendingDraft)
  • SourceStrongAI branding

Analytics & Ads

  • Google Analytics: GA4 property G-CMC251LQT6 (configured in both mkdocs.yml and overrides/main.html gtag snippet)
  • Google AdSense: ca-pub-4988388949365963 (loaded in overrides/main.html)

Key Conventions

  • All contact references should use hi@sourcestrongai.com
  • External product links should open in new tabs (target="_blank" rel="noopener")
  • The PDF export plugin (with-pdf) is disabled by default; enable with ENABLE_PDF_EXPORT=1
  • Example/dummy emails in code samples (SQL, Python, etc.) should NOT be changed
  • Copyright references Kuldeep Singh Sidhu with AGPL-3.0 license