#!/usr/bin/env python3
"""Every section a WAI specification text cites in another WAI text exists.

Reads wai/SPEC.md, wai/extensions/*.md and wai/SECURITY.md. In each, a reference to a
section of SPEC.md or of an extension is checked against the headings of the file it
names. Three spellings are recognised, including where the reference breaks across
lines:

  [text](X.md) §N           a link, then the section
  [text §N](X.md)           the section inside the link text
  X §N, X.md §N, `X` §N     an extension's file name (or SPEC) in prose, then the section

where §N is a section number of one or more parts (§7, §2.4.3), and a range written
§A–§B checks both ends. A heading is numbered when it begins `#`…, then the number, as
in `## 7.`, `### 6.4 ` or `## 7.1.`. Files under wai/research/ are dated records, kept as
written, and are not read.

Exit status 1, with one line per reference that names a missing section; 0 otherwise.

    python3 scripts/check-wai-section-refs.py [repo-root]
"""
import os, re, sys

ROOT = os.path.abspath(sys.argv[1] if len(sys.argv) > 1 else os.path.join(os.path.dirname(os.path.abspath(__file__)), ".."))
WAI = os.path.join(ROOT, "wai")
EXT = os.path.join(WAI, "extensions")

HEADING = re.compile(r"^#{1,6}[ \t]+([0-9]+(?:\.[0-9]+)*)\.?(?=[ \t])", re.M)
SECTION = r"§[ \t]*([0-9]+(?:\.[0-9]+)*)"
RANGE = re.compile(r"[ \t]*[–-][ \t]*§[ \t]*([0-9]+(?:\.[0-9]+)*)")


def sections(path):
    with open(path, encoding="utf-8") as f:
        return set(HEADING.findall(f.read()))


def texts():
    out = [os.path.join(WAI, "SPEC.md")]
    out += sorted(os.path.join(EXT, n) for n in os.listdir(EXT) if n.endswith(".md"))
    out.append(os.path.join(WAI, "SECURITY.md"))  # it cites the texts it points readers to
    return out


def main():
    files = texts()
    known = {os.path.normpath(p): sections(p) for p in files}
    stems = sorted((os.path.splitext(os.path.basename(p))[0] for p in files if os.path.dirname(p) == EXT),
                   key=len, reverse=True)
    by_stem = {os.path.splitext(os.path.basename(p))[0]: os.path.normpath(p) for p in files if os.path.dirname(p) == EXT}
    by_stem["SPEC"] = os.path.normpath(os.path.join(WAI, "SPEC.md"))
    # A file name in prose, not inside a link target or a longer word or path.
    bare = re.compile(r"(?<![\w/.(-])`?(" + "|".join(map(re.escape, stems + ["SPEC"])) + r")(?:\.md)?`?[ \t\n]+" + SECTION)
    link_then = re.compile(r"\[[^\]]*\]\(([^)\s]+\.md)\)[ \t\n]*" + SECTION)
    link_with = re.compile(r"\[([^\]]*?)" + SECTION + r"[^\]]*\]\(([^)\s]+\.md)\)")

    findings = []
    for path in files:
        with open(path, encoding="utf-8") as f:
            text = f.read()
        here = os.path.dirname(path)

        def check(target, num, at):
            target = os.path.normpath(target)
            if target not in known:
                return  # not a WAI specification text (a README, a research note)
            if num not in known[target]:
                line = text.count("\n", 0, at) + 1
                rel = os.path.relpath(path, ROOT)
                findings.append(f"{rel}:{line}: §{num} of {os.path.relpath(target, ROOT)} does not exist")

        def check_range(target, m, end):
            r = RANGE.match(text, end)
            if r:
                check(target, r.group(1), m.start())

        for m in link_then.finditer(text):
            target = os.path.join(here, m.group(1).split("#")[0])
            check(target, m.group(2), m.start())
            check_range(target, m, m.end())
        for m in link_with.finditer(text):
            target = os.path.join(here, m.group(3).split("#")[0])
            check(target, m.group(2), m.start())
        for m in bare.finditer(text):
            check(by_stem[m.group(1)], m.group(2), m.start())
            check_range(by_stem[m.group(1)], m, m.end())

    for f in sorted(set(findings)):
        print(f)
    return 1 if findings else 0


if __name__ == "__main__":
    sys.exit(main())