"""STYLE.md sections 1 and 4 as code: the banned words and the counts.

Shared by test_messages.py (UX-08) and test_ui_text.py (UX-13 P5), so the two
checks cannot drift apart. Moved here unchanged from test_messages.py.
"""

import re

BANNED = re.compile(
    r"\b(?:(?:SAFETY|TARIFF|TEL|CFG|FCST|OPT|UX|HW|FW|WITNESS|PKG|ARCH)-\d+|"
    r"integrator|anti-?windup|windup|clamp(?:ed|s)?|i_clamped|saturat\w*|"
    r"BUG SIGNAL|latch(?:ed)?|arbiter|allocator|registers?|0x[0-9A-Fa-f]+|"
    r"layer [123]|heartbeat|watchdog|telegrams?|ACKs?|CRC|OTA|Modbus|planner|"
    r"capaciteitstarief|journalctl|ssh|grep|"
    r"\d+\.\d+\.\d+\.\d+|[\w-]+\.(?:py|md|yaml|json|db))\b|§\s?\d", re.I)
ENTITY = re.compile(r"\b(?:sensor|number|binary_sensor|input_\w+|switch|select)"
                    r"\.[a-z0-9_*]+")
SNAKE = re.compile(r"\b[a-z0-9]+(?:_[a-z0-9]+)+\b")
SHOUT = re.compile(r"\b[A-Z]{2,}[A-Z0-9]*(?:[\s-]+[A-Z]{2,}[A-Z0-9]*)+\b")
PLACEHOLDER = re.compile(r"\{[^{}]*\}|%[-+ #0-9.]*[sdrfgx%]")


def words(text):
    return [w for w in re.findall(r"\S+", PLACEHOLDER.sub("X", text))
            if re.search(r"[A-Za-z0-9]", w)]


def sentences(text):
    """Sentence count: a stop followed by a space and a capital, or the end."""
    t = PLACEHOLDER.sub("X", text.strip())
    t = re.sub(r"\b(e\.g|i\.e|etc)\.", "eg", t)
    return len([s for s in re.split(r"[.!?](?:\s+|$)", t) if s.strip()])


def title_problems(title):
    out = []
    if len(words(title)) > 5:
        out.append(f"title over 5 words ({len(words(title))})")
    if title.lower().startswith("goodwe"):
        out.append("title carries a GoodWe prefix")
    if title.rstrip().endswith("."):
        out.append("title ends in a full stop")
    return out + text_problems(title, keys=False, entities=False)


def text_problems(text, *, keys, entities):
    t = PLACEHOLDER.sub("X", text)
    out = []
    hit = sorted({m.group(0) for m in BANNED.finditer(t)})
    if hit:
        out.append("banned: " + ", ".join(hit[:4]))
    if not entities and ENTITY.search(t):
        out.append("entity id: " + ENTITY.search(t).group(0))
    if not keys and SNAKE.search(ENTITY.sub("", t)):
        out.append("option key: " + SNAKE.search(ENTITY.sub("", t)).group(0))
    if SHOUT.search(t):
        out.append("capitals: " + SHOUT.search(t).group(0))
    return out