main
md 1.33 KB

Bender: Stdlib-Only Crawler Phase 1 Log

Issue: #5 (Stdlib-Only Python Crawler) Period: Phase 1 Timestamp: 2026-05-18T10:26:47Z

Objectives

  • Implement stdlib-only Python crawler writing to data/raw/YYYY-WNN.json
  • Support trending rank computation from prior snapshots
  • Support pagination up to GitHub's 1,000-result limit

Accomplishments

  1. Crawler Implementation

    • Pure Python stdlib (urllib) with no external dependencies
    • Default: top 250 search results per query
    • Pagination support up to 1,000 results (GitHub's limit)
  2. Search Queries

    • Created: created:>{date} stars:>50
    • Pushed: pushed:>{date} stars:>50
  3. Significance Filtering

    • Excludes forks
    • Excludes repos without descriptions
    • Excludes repos without READMEs
    • Excludes obvious tutorial/homework/template repos
    • Handles SAML-blocked README endpoints gracefully (skip vs fail)
  4. Data Output

    • Writes data/raw/YYYY-WNN.json for weekly snapshots
    • Computes trending rank from latest prior snapshot when available

Benefits

  • Minimal CI setup (no external dependencies)
  • Deterministic crawls (repeatable results)
  • Cheap execution (suitable for local and Actions runs)
  • Path to deeper crawls without breaking data contract

Status

✅ COMPLETE (Issue #5 closed, implementation in PR #25-#26)