cpp-flashcards/org/study_deck_02/gen-flashcards.py

#!/usr/bin/env python3
"""Auto-generate Anki flashcards from problem .org files.

Generates one recognition card per problem:
  Front: Problem number + name
  Back:  Problem statement (first paragraph)
"""

import re
import sys
from pathlib import Path

BASE = Path(__file__).parent
DSA = BASE / "dsa"
OUTPUT = BASE / "toolkit" / "gen-problem-cards.org"

TOPIC_DISPLAY = {
    "1-d-dynamic-programming": "1D DP",
    "2-d-dynamic-programming": "2D DP",
    "advanced-graphs": "Advanced Graphs",
    "arrays-hashing": "Arrays & Hashing",
    "backtracking": "Backtracking",
    "binary-search": "Binary Search",
    "bit-manipulation": "Bit Manipulation",
    "graphs": "Graphs",
    "greedy": "Greedy",
    "heap-priority-queue": "Heap / Priority Queue",
    "intervals": "Intervals",
    "linked-list": "Linked List",
    "math-geometry": "Math & Geometry",
    "sliding-window": "Sliding Window",
    "stack": "Stack",
    "trees": "Trees",
    "tries": "Tries",
    "two-pointers": "Two Pointers",
}


def parse_file(path: Path) -> dict | None:
    """Extract problem metadata and statement from an .org file."""
    text = path.read_text()

    # Match heading: * TODO 0217. Contains Duplicate :easy:
    m = re.search(
        r"^\* (?:TODO|DONE) (\d{4})\. (.+?) :(easy|medium|hard):",
        text,
        re.MULTILINE,
    )
    if not m:
        return None

    num = m.group(1)
    name = m.group(2).strip()
    difficulty = m.group(3)

    # Topic = parent directory name
    topic_slug = path.parent.name
    topic = TOPIC_DISPLAY.get(topic_slug, topic_slug.replace("-", " ").title())

    # Problem statement: first non-empty, non-heading, non-property line after :END:
    after_props = text[m.end():]
    lines = after_props.split("\n")
    statement_lines = []
    in_block = False
    for line in lines:
        stripped = line.strip()
        if stripped == ":END:":
            continue
        if not stripped:
            if statement_lines:
                break
            continue
        if stripped.startswith("#+") or stripped.startswith("*") or stripped.startswith(":"):
            if statement_lines:
                break
            continue
        statement_lines.append(stripped)

    statement = " ".join(statement_lines).strip()
    # Clean up org markup for plain text
    statement = re.sub(r"[~=*/_]", "", statement)

    return {
        "num": num,
        "name": name,
        "difficulty": difficulty,
        "topic": topic,
        "statement": statement,
        "path": path,
    }


def generate_cards(problems: list[dict]) -> str:
    """Generate org-mode flashcard content."""
    parts = []
    parts.append("#+TITLE: Auto-Generated Problem Cards")
    parts.append("#+ANKI_DECK: study_deck_02")
    parts.append("")

    for p in sorted(problems, key=lambda x: int(x["num"])):
        num = p["num"]
        name = p["name"]
        diff = p["difficulty"]
        topic = p["topic"]
        stmt = p["statement"]

        # Card: given problem name → what does it ask?
        title = f"LC {num}. {name} — what does it ask?"
        tags = f":leetcode:{diff}:{topic.replace(' ', '-').lower()}:retrieval::recognition:"

        parts.append(f"* {title} {tags}")
        parts.append(":PROPERTIES:")
        parts.append(":ANKI_NOTE_TYPE: Basic")
        parts.append(":END:")
        parts.append("** Front")
        parts.append(f"What does LeetCode {num} *{name}* ask you to do?")
        parts.append("** Back")
        parts.append(stmt)
        parts.append("")

    return "\n".join(parts)


def main():
    dry_run = "--dry-run" in sys.argv

    org_files = sorted(DSA.rglob("*.org"))
    org_files = [f for f in org_files if f.name != "udfs.org"]
    print(f"Found {len(org_files)} problem files")

    problems = []
    skipped = 0
    for f in org_files:
        p = parse_file(f)
        if p:
            problems.append(p)
        else:
            skipped += 1
            print(f"  skipped: {f.name}")

    print(f"Parsed: {len(problems)} problems, {skipped} skipped")

    content = generate_cards(problems)

    if dry_run:
        print(f"\nWould write {len(problems)} cards to {OUTPUT}")
        # Show first 3 cards as preview
        preview = content.split("\n\n\n")[:3]
        print("\n--- Preview ---")
        print("\n\n".join(preview))
    else:
        OUTPUT.parent.mkdir(parents=True, exist_ok=True)
        OUTPUT.write_text(content)
        print(f"\nWrote {len(problems)} cards to {OUTPUT}")


if __name__ == "__main__":
    main()