our-claude-skills/custom-skills/32-notion-writer/code/scripts/test_parser.py

#!/usr/bin/env python3
"""Parser tests for notion_writer.py — run with `python test_parser.py`."""

import json
import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).parent))

from notion_writer import markdown_to_notion_blocks, parse_rich_text


FIXTURE = """# Top Heading

## Section With **Bold** And `code`

Plain paragraph with *italic*, **bold**, `inline`, a [link](https://example.com), and ~~strike~~.

- First bullet
- Second bullet with **bold**
* Alt bullet marker

1. First numbered
2. Second numbered

- [ ] Unchecked task
- [x] Checked task

> A quote with *emphasis*

```python
print("hello")
```

---

| Task | Priority | Status |
|------|----------|--------|
| Fix HTML reporter | P0 | Pending |
| Korean encoding | **P1** | In Progress |
| Type hints | P2 | [link](https://example.com/types) |
"""


def _assert(cond, msg):
    if not cond:
        print(f"  ✗ FAIL: {msg}")
        raise SystemExit(1)
    print(f"  ✓ {msg}")


def test_rich_text_plain():
    spans = parse_rich_text("hello world")
    _assert(len(spans) == 1, "plain text emits one span")
    _assert(spans[0]["text"]["content"] == "hello world", "plain content preserved")
    _assert("annotations" not in spans[0], "plain text has no annotations")


def test_rich_text_bold():
    spans = parse_rich_text("a **bold** b")
    _assert(len(spans) == 3, "bold splits into 3 spans (pre / bold / post)")
    _assert(spans[0]["text"]["content"] == "a ", "pre-text preserved")
    _assert(spans[1]["text"]["content"] == "bold", "bold content stripped of markers")
    _assert(spans[1].get("annotations", {}).get("bold") is True, "bold annotation set")
    _assert(spans[2]["text"]["content"] == " b", "post-text preserved")


def test_rich_text_italic():
    spans = parse_rich_text("x *em* y")
    _assert(any(s.get("annotations", {}).get("italic") for s in spans), "italic annotation set")


def test_rich_text_code():
    spans = parse_rich_text("run `npm test` now")
    code_spans = [s for s in spans if s.get("annotations", {}).get("code")]
    _assert(len(code_spans) == 1, "code span emitted")
    _assert(code_spans[0]["text"]["content"] == "npm test", "code content stripped of backticks")


def test_rich_text_link():
    spans = parse_rich_text("see [the docs](https://example.com/docs) please")
    link_spans = [s for s in spans if s["text"].get("link")]
    _assert(len(link_spans) == 1, "link span emitted")
    _assert(link_spans[0]["text"]["link"]["url"] == "https://example.com/docs", "link URL preserved")
    _assert(link_spans[0]["text"]["content"] == "the docs", "link text preserved")


def test_rich_text_strike():
    spans = parse_rich_text("~~gone~~ now")
    strike = [s for s in spans if s.get("annotations", {}).get("strikethrough")]
    _assert(len(strike) == 1, "strikethrough annotation set")


def test_rich_text_mixed():
    spans = parse_rich_text("**bold** and *italic* and `code` and [link](https://x.com)")
    kinds = set()
    for s in spans:
        ann = s.get("annotations") or {}
        if ann.get("bold"): kinds.add("bold")
        if ann.get("italic"): kinds.add("italic")
        if ann.get("code"): kinds.add("code")
        if s["text"].get("link"): kinds.add("link")
    _assert(kinds == {"bold", "italic", "code", "link"}, "all four inline types present")


def test_blocks_headings():
    blocks = markdown_to_notion_blocks(FIXTURE)
    types = [b["type"] for b in blocks]
    _assert("heading_1" in types, "H1 emitted")
    _assert("heading_2" in types, "H2 emitted")


def test_blocks_bullets_and_numbers():
    blocks = markdown_to_notion_blocks(FIXTURE)
    types = [b["type"] for b in blocks]
    _assert(types.count("bulleted_list_item") == 3, "3 bulleted items (incl. * marker)")
    _assert(types.count("numbered_list_item") == 2, "2 numbered items")


def test_blocks_todo():
    blocks = markdown_to_notion_blocks(FIXTURE)
    todos = [b for b in blocks if b["type"] == "to_do"]
    _assert(len(todos) == 2, "2 to_do blocks")
    _assert(todos[0]["to_do"]["checked"] is False, "first todo unchecked")
    _assert(todos[1]["to_do"]["checked"] is True, "second todo checked")
    _assert(todos[0]["to_do"]["rich_text"][0]["text"]["content"] == "Unchecked task",
            "todo text stripped of '- [ ]' marker")


def test_blocks_quote_code_divider():
    blocks = markdown_to_notion_blocks(FIXTURE)
    types = [b["type"] for b in blocks]
    _assert("quote" in types, "quote block emitted")
    _assert("code" in types, "code block emitted")
    _assert("divider" in types, "divider block emitted")

    code_block = next(b for b in blocks if b["type"] == "code")
    _assert(code_block["code"]["language"] == "python", "code language preserved")


def test_blocks_table():
    blocks = markdown_to_notion_blocks(FIXTURE)
    tables = [b for b in blocks if b["type"] == "table"]
    _assert(len(tables) == 1, "exactly one table block emitted")

    t = tables[0]["table"]
    _assert(t["table_width"] == 3, "table width = 3 columns")
    _assert(t["has_column_header"] is True, "column header flag set")
    _assert(len(t["children"]) == 4, "4 rows (1 header + 3 body)")

    # Header cell content
    header_row = t["children"][0]["table_row"]["cells"]
    _assert(header_row[0][0]["text"]["content"] == "Task", "header cell 0 = 'Task'")
    _assert(header_row[1][0]["text"]["content"] == "Priority", "header cell 1 = 'Priority'")

    # Body row inline formatting
    row_bold = t["children"][2]["table_row"]["cells"]  # "| Korean encoding | **P1** | In Progress |"
    bold_cell = row_bold[1]
    _assert(bold_cell[0].get("annotations", {}).get("bold") is True,
            "bold inline formatting preserved inside table cell")

    row_link = t["children"][3]["table_row"]["cells"]  # "| Type hints | P2 | [link](...) |"
    link_cell = row_link[2]
    _assert(link_cell[0]["text"].get("link", {}).get("url") == "https://example.com/types",
            "link inline formatting preserved inside table cell")


def test_no_literal_markers_leak():
    """Regression: before the rich-text rewrite, ** and [ ] leaked as literal text."""
    blocks = markdown_to_notion_blocks("A **bold** word and a [link](https://x.com).")
    para = blocks[0]["paragraph"]["rich_text"]
    joined = "".join(s["text"]["content"] for s in para)
    _assert("**" not in joined, "no literal ** in paragraph")
    _assert("[link]" not in joined, "no literal [link] in paragraph")
    _assert("bold" in joined and "link" in joined, "visible words preserved")


def run_all():
    tests = [
        test_rich_text_plain,
        test_rich_text_bold,
        test_rich_text_italic,
        test_rich_text_code,
        test_rich_text_link,
        test_rich_text_strike,
        test_rich_text_mixed,
        test_blocks_headings,
        test_blocks_bullets_and_numbers,
        test_blocks_todo,
        test_blocks_quote_code_divider,
        test_blocks_table,
        test_no_literal_markers_leak,
    ]
    for t in tests:
        print(f"\n{t.__name__}")
        t()

    print("\n" + "=" * 50)
    print(f"✅ All {len(tests)} tests passed")
    print("=" * 50)

    # Dump the fixture's block types for visual sanity check
    blocks = markdown_to_notion_blocks(FIXTURE)
    print("\nFixture produces block types:")
    for b in blocks:
        print(f"  - {b['type']}")


if __name__ == "__main__":
    run_all()