How do their APIs compare?

Neither of these repositories exposes a runtime API in the traditional sense — they're both content/education repos, not services or libraries with a public interface — but there are a couple of "API-like" surfaces worth distinguishing:

practical-tutorials/project-based-learning This repo is essentially a curated README (a list of tutorial links organized by language) [README excerpt]. Its only "programmatic" surface is a set of internal automation scripts and GitHub Actions used to maintain that README's integrity — not an API meant for external consumers:

• scripts/check_readme.py contains functions like escape_md_cell for safely rendering untrusted fetched data (e.g., a linked page's <title>) into markdown table cells posted to GitHub issues [scripts/check_readme.py:L887-L903], normalize_url for canonicalizing URLs for comparison [scripts/check_readme.py:L177-L186], load_url_items_from_file for reading a JSON list of URLs or the output of a check-diff --json command [scripts/check_readme.py:L772-L789], and _default_state_entry for building tracking-state records used in link-liveness checks [scripts/check_readme.py:L822-L825].
• These functions are invoked from GitHub Actions workflows like .github/workflows/validate-pr.yml, which runs check_readme.py check-links against added URLs to detect broken/blocked links in PRs [.github/workflows/validate-pr.yml:L36-L48].
• So the "interface" here is a CLI/workflow-oriented internal tool, not a reusable public API — there's no package, no importable module surface intended for consumers.

thealgorithms/python This repo is a collection of independent, educational algorithm implementations organized by topic (backtracking, boolean_algebra, audio_filters, etc.), each exposing plain Python functions with type hints and doctests rather than a unified API:

• Example: generate_parenthesis(n: int) -> list[str] in [backtracking/generate_parentheses.py:L48-L82], word_break(input_string: str, word_dict: set[str]) -> bool in [backtracking/word_break.py:L48-L75], print_solution(grid: Matrix) -> None in [backtracking/sudoku.py:L112-L134], and simplify_kmap/main() in [boolean_algebra/karnaugh_map_simplification.py:L32-L56].
• Each file is self-contained and runnable standalone (if __name__ == "__main__" blocks with doctest.testmod()), and the project's CONTRIBUTING.md explicitly requires type hints and mypy-testable signatures for all submissions [CONTRIBUTING.md:L182-L194].
• There's no single unifying "API" — it's a library of many small, independently callable functions, meant to be read and learned from (per the README: "Implementations are for learning purposes only") rather than imported as a cohesive package.

Comparison summary

Both repos are non-service, non-library-distribution projects, so "API comparison" mostly comes down to: one has automation-script internals for repo upkeep, the other has hundreds of small, independently documented educational functions.

FOLLOW_UPS:

• What does check_readme.py's link-checking pipeline do?
• How are algorithms tested in thealgorithms/python?
• What triggers the stale-PR workflow in project-based-learning?