Skip to content

devlog.md

Latest commit

 

History

History
112 lines (91 loc) · 5.68 KB

devlog.md

File metadata and controls

112 lines (91 loc) · 5.68 KB
Raw

Google Sites Clone — Development Log

Chronological summary of 11 development sessions (#86–#94) + 2 Dev.to articles.

📊 Overview

Metric Value
Chat sessions 9 (#86–#94)
Dev.to articles 2 (#13, #14)
Total lines 26 448
Total size ~1.04 MB
Period March 2026

📁 File Index

# File Lines Size Topic
86 Schematization 2 063 93 KB Initial clone — Node.js generator, 33 HTML pages, images, YouTube previews
87 Schematization 3 663 175 KB README/ARCH, 81 YouTube on member pages, parser comparison (10 options), SingleFile CLI
88 Fixing Merge-Video Deployment 5 664 258 KB Full SingleFile clone (34 pages), Puppeteer vs SingleFile, Vercel deploy, repo creation, credential leak fix
89 Fixing Leaked Credentials 2 205 79 KB Google/GitHub OAuth stubs, SVG icons (moon/sun/github), hero centering, remote fix
90 Implementing Video Pipeline 2 185 86 KB README (readme-guidelines), Vercel deploy gsclone.osovsky.com, favicon, Alternatives section
91 Video Pipeline and Report Integration 567 29 KB lib/video.js (YT/Vimeo scan), build.js iframe→thumbnail, report.js → site/, Vimeo regex
92 Publishing CLI Tool 2 740 107 KB Video Grid (YT/Vimeo/GDrive), root page filter, clickable report chips, npm publish
93 что надо сделать 4 341 135 KB Phase 5 SaaS: Real OAuth, Resend, R2 storage, Lifecycle Rules, Usage Tiers
94 Google Analytics Integration 2 662 111 KB Rate limit (Upstash Redis), GitHub Star gate, Usage Tiers ($99 Unlimited), GA4
13 autoposting-pipeline 187 10 KB Dev.to article: Content Distribution Pipeline — 9 types, 20+ platforms
14 Google sites devto-article 171 7 KB Dev.to article: AI Clone Failure Story — 5.4/10 performance review

🔄 Development Timeline

Phase 1 — Schematization Clone (#86–#87)

  • Scraped Google Sites «Схематизация» via read_url_content (text-only)
  • Built Node.js generator → 33 HTML pages + style.css + nav.js
  • Downloaded 84 images from Google CDN, deduplicated to ~48
  • Extracted 12+ YouTube videos, downloaded thumbnails
  • Created Markdown versions with frontmatter
  • Wrote README.md and ARCHITECTURE.md for smd.bar repo
  • Added 81 YouTube thumbnails to 7 member pages
  • Compared 10 parsing tools → chose SingleFile CLI (#8) for best fidelity

Phase 2 — Full SingleFile Rebuild (#88)

  • Installed single-file-cli, saved all 34 pages (7–12 MB each, base64 inline)
  • Post-processing: internal link rewriting, Google redirect unwrapping, YouTube iframe→thumbnail
  • Fixed hidden sidebar (display:none → display:block)
  • Medium link migration (@osowski → osovsky.medium.com)
  • Served locally via npx serve

Phase 3 — google-sites-clone Project (#88–#89)

  • Created maximosovsky/google-sites-clone repo on GitHub
  • Fixed credential leak (remote pointed to agent-workflows)
  • Added Google/GitHub OAuth stubs (localStorage-based)
  • Replaced emoji theme toggle with SVG moon/sun from site-design/icons
  • Centered hero layout

Phase 4 — Pipeline & CLI (#90–#92)

  • README formatted per readme-guidelines
  • Deployed to Vercel → gsclone.osovsky.com
  • Added favicon from Google Sites logo SVG
  • Compared competitors (HTTrack, google-sites-backup) → added Alternatives section
  • Implemented lib/video.js — scans _content/ for YouTube (90), Vimeo (6), GDrive (3)
  • Built Video Grid: CSS Grid with thumbnails injected into SF pages after <body>
  • Root page filtering: index.html reserved for navigation shell
  • Report improvements: clickable colored chips, Total site MB
  • Published to npm: npx google-sites-clone

Phase 5 — SaaS Features (#93–#94)

  • Implemented Real Google OAuth + Real GitHub OAuth (Vercel API routes)
  • Integrated Resend for email delivery
  • Created Cloudflare R2 bucket gsclone + 3 Lifecycle Rules (zips 7d, previews 7d, reports 360d)
  • Designed Usage Tiers:
    • Free (Google Auth) — 1 clone, 250 MB ZIP
    • Starred (GitHub + ⭐) — 5/day, 20/month, 250 MB ZIP
    • Unlimited ($99/mo) — offered via email when Starred exhausted
  • Rate limit via Upstash Redis (clones:daily:{email}:{date})
  • GitHub Star gate: GET /user/starred/maximosovsky/google-sites-clone
  • Google Analytics (GA4) integration

✍️ Dev.to Articles (unpublished)

#13 — Autoposting Pipeline

Title: How I Designed a Content Distribution Pipeline — 9 Content Types, 20+ Platforms, One Interactive Diagram

  • 187 lines, covers the routing logic for articles, videos, podcasts, events
  • Open source: github.com/maximosovsky/avtoposting

#14 — Google Sites Clone Failure Story

Title: I Asked an AI to Clone a Website. It Wrote 15 Scripts, Made 12 Commits, and Still Failed

  • 171 lines, documents 6 AI failures, performance review 5.4/10
  • Compares 10 scraping tools, highlights Google Sites vendor lock-in
  • Live clone: maximosovsky.github.io/smd.bar/schematization/

🏗️ Key Technical Decisions

Decision Rationale
Two-Pass Pipeline (SingleFile + Puppeteer) SF captures CSS/images, PP captures text/links/YouTube
Iframe-based navigation Sidebar + iframe isolates SF CSS from nav shell
Video Grid with inline styles all:initial resets SF styles, no external CSS conflicts
Root page skip in build index.html reserved for generated nav, content page keeps transliterated name
Cookie-less rate limit (Upstash Redis) Reliable per-email tracking, can't be bypassed by clearing cookies
R2 Lifecycle Rules Zero-code auto-cleanup: ZIPs 7 days, reports 360 days