rose-ash/sx/content/highlight.py

"""Syntax highlighting using Tailwind classes.

Produces sx source with coloured spans — no external CSS dependencies.
Showcases the on-demand CSS system.
"""
from __future__ import annotations

import re


def highlight_sx(code: str) -> str:
    """Highlight s-expression source code as sx with Tailwind spans."""
    tokens = _tokenize_sx(code)
    parts = []
    for kind, text in tokens:
        escaped = text.replace("\\", "\\\\").replace('"', '\\"')
        if kind == "comment":
            parts.append(f'(span :class "text-stone-400 italic" "{escaped}")')
        elif kind == "string":
            parts.append(f'(span :class "text-emerald-700" "{escaped}")')
        elif kind == "keyword":
            parts.append(f'(span :class "text-violet-600" "{escaped}")')
        elif kind == "component":
            parts.append(f'(span :class "text-rose-600 font-semibold" "{escaped}")')
        elif kind == "special":
            parts.append(f'(span :class "text-sky-700 font-semibold" "{escaped}")')
        elif kind == "paren":
            parts.append(f'(span :class "text-stone-400" "{escaped}")')
        elif kind == "number":
            parts.append(f'(span :class "text-amber-700" "{escaped}")')
        elif kind == "boolean":
            parts.append(f'(span :class "text-orange-600" "{escaped}")')
        else:
            parts.append(f'(span "{escaped}")')
    return "(<> " + " ".join(parts) + ")"


_SX_SPECIALS = {
    "defcomp", "defrelation", "define",
    "if", "when", "cond", "case", "and", "or", "not",
    "let", "let*", "lambda", "fn",
    "do", "begin", "quote",
    "->", "map", "filter", "reduce", "some", "every?",
    "map-indexed", "for-each",
    "&key", "&rest",
}

_SX_TOKEN_RE = re.compile(
    r'(;[^\n]*)'            # comment
    r'|("(?:[^"\\]|\\.)*")' # string
    r'|(:[a-zA-Z][\w?!-]*)' # keyword
    r'|(~[\w-]+)'           # component
    r'|([()[\]{}])'         # parens/brackets
    r'|(\d+\.?\d*)'         # number
    r'|(true|false|nil)\b'  # boolean/nil
    r'|([\w?!+\-*/<>=&.]+)' # symbol
    r'|(\s+)'               # whitespace
    r'|(.)'                 # other
)


def _tokenize_sx(code: str) -> list[tuple[str, str]]:
    tokens = []
    for m in _SX_TOKEN_RE.finditer(code):
        if m.group(1):
            tokens.append(("comment", m.group(1)))
        elif m.group(2):
            tokens.append(("string", m.group(2)))
        elif m.group(3):
            tokens.append(("keyword", m.group(3)))
        elif m.group(4):
            tokens.append(("component", m.group(4)))
        elif m.group(5):
            tokens.append(("paren", m.group(5)))
        elif m.group(6):
            tokens.append(("number", m.group(6)))
        elif m.group(7):
            tokens.append(("boolean", m.group(7)))
        elif m.group(8):
            text = m.group(8)
            if text in _SX_SPECIALS:
                tokens.append(("special", text))
            else:
                tokens.append(("symbol", text))
        elif m.group(9):
            tokens.append(("ws", m.group(9)))
        else:
            tokens.append(("other", m.group(10)))
    return tokens


def highlight_python(code: str) -> str:
    """Highlight Python source code as sx with Tailwind spans."""
    tokens = _tokenize_python(code)
    parts = []
    for kind, text in tokens:
        escaped = text.replace("\\", "\\\\").replace('"', '\\"')
        if kind == "comment":
            parts.append(f'(span :class "text-stone-400 italic" "{escaped}")')
        elif kind == "string":
            parts.append(f'(span :class "text-emerald-700" "{escaped}")')
        elif kind == "keyword":
            parts.append(f'(span :class "text-violet-600 font-semibold" "{escaped}")')
        elif kind == "builtin":
            parts.append(f'(span :class "text-sky-700" "{escaped}")')
        elif kind == "decorator":
            parts.append(f'(span :class "text-amber-600" "{escaped}")')
        elif kind == "number":
            parts.append(f'(span :class "text-amber-700" "{escaped}")')
        else:
            parts.append(f'(span "{escaped}")')
    return "(<> " + " ".join(parts) + ")"


_PY_KEYWORDS = {
    "False", "None", "True", "and", "as", "assert", "async", "await",
    "break", "class", "continue", "def", "del", "elif", "else", "except",
    "finally", "for", "from", "global", "if", "import", "in", "is",
    "lambda", "nonlocal", "not", "or", "pass", "raise", "return",
    "try", "while", "with", "yield",
}

_PY_BUILTINS = {
    "print", "len", "range", "str", "int", "float", "list", "dict",
    "set", "tuple", "type", "isinstance", "getattr", "setattr", "hasattr",
    "super", "property", "staticmethod", "classmethod", "enumerate", "zip",
    "map", "filter", "sorted", "reversed", "any", "all", "min", "max",
    "abs", "sum", "open", "input", "format", "repr", "id", "hash",
}

_PY_TOKEN_RE = re.compile(
    r'(#[^\n]*)'                         # comment
    r'|("""[\s\S]*?"""|\'\'\'[\s\S]*?\'\'\')' # triple-quoted string
    r'|("(?:[^"\\]|\\.)*")'             # double-quoted string
    r"|('(?:[^'\\]|\\.)*')"             # single-quoted string
    r'|(@\w+)'                           # decorator
    r'|(\d+\.?\d*(?:e[+-]?\d+)?)'       # number
    r'|([a-zA-Z_]\w*)'                  # identifier
    r'|(\s+)'                            # whitespace
    r'|(.)'                              # other
)


def _tokenize_python(code: str) -> list[tuple[str, str]]:
    tokens = []
    for m in _PY_TOKEN_RE.finditer(code):
        if m.group(1):
            tokens.append(("comment", m.group(1)))
        elif m.group(2):
            tokens.append(("string", m.group(2)))
        elif m.group(3):
            tokens.append(("string", m.group(3)))
        elif m.group(4):
            tokens.append(("string", m.group(4)))
        elif m.group(5):
            tokens.append(("decorator", m.group(5)))
        elif m.group(6):
            tokens.append(("number", m.group(6)))
        elif m.group(7):
            text = m.group(7)
            if text in _PY_KEYWORDS:
                tokens.append(("keyword", text))
            elif text in _PY_BUILTINS:
                tokens.append(("builtin", text))
            else:
                tokens.append(("ident", text))
        elif m.group(8):
            tokens.append(("ws", m.group(8)))
        else:
            tokens.append(("other", m.group(9)))
    return tokens


def highlight_bash(code: str) -> str:
    """Highlight bash source code as sx with Tailwind spans."""
    tokens = _tokenize_bash(code)
    parts = []
    for kind, text in tokens:
        escaped = text.replace("\\", "\\\\").replace('"', '\\"')
        if kind == "comment":
            parts.append(f'(span :class "text-stone-400 italic" "{escaped}")')
        elif kind == "string":
            parts.append(f'(span :class "text-emerald-700" "{escaped}")')
        elif kind == "variable":
            parts.append(f'(span :class "text-violet-600" "{escaped}")')
        elif kind == "keyword":
            parts.append(f'(span :class "text-sky-700 font-semibold" "{escaped}")')
        elif kind == "flag":
            parts.append(f'(span :class "text-amber-700" "{escaped}")')
        else:
            parts.append(f'(span "{escaped}")')
    return "(<> " + " ".join(parts) + ")"


_BASH_KEYWORDS = {
    "if", "then", "else", "elif", "fi", "for", "while", "do", "done",
    "case", "esac", "in", "function", "return", "exit", "export",
    "source", "local", "readonly", "declare", "set", "unset",
}

_BASH_TOKEN_RE = re.compile(
    r'(#[^\n]*)'                          # comment
    r'|("(?:[^"\\]|\\.)*")'              # double-quoted string
    r"|('(?:[^'\\]|\\.)*')"              # single-quoted string
    r'|(\$\{?\w+\}?|\$\()'              # variable
    r'|(--?[\w-]+)'                       # flag
    r'|([a-zA-Z_][\w-]*)'               # word
    r'|(\s+)'                             # whitespace
    r'|(.)'                               # other
)


def _tokenize_bash(code: str) -> list[tuple[str, str]]:
    tokens = []
    for m in _BASH_TOKEN_RE.finditer(code):
        if m.group(1):
            tokens.append(("comment", m.group(1)))
        elif m.group(2):
            tokens.append(("string", m.group(2)))
        elif m.group(3):
            tokens.append(("string", m.group(3)))
        elif m.group(4):
            tokens.append(("variable", m.group(4)))
        elif m.group(5):
            tokens.append(("flag", m.group(5)))
        elif m.group(6):
            text = m.group(6)
            if text in _BASH_KEYWORDS:
                tokens.append(("keyword", text))
            else:
                tokens.append(("word", text))
        elif m.group(7):
            tokens.append(("ws", m.group(7)))
        else:
            tokens.append(("other", m.group(8)))
    return tokens


def highlight(code: str, language: str = "lisp"):
    """Highlight code in the given language. Returns SxExpr for wire format."""
    from shared.sx.parser import SxExpr, serialize
    if language in ("lisp", "sx", "sexp"):
        return SxExpr(highlight_sx(code))
    elif language in ("python", "py"):
        return SxExpr(highlight_python(code))
    elif language in ("bash", "sh", "shell"):
        return SxExpr(highlight_bash(code))
    # Fallback: no highlighting, just escaped text
    return SxExpr("(span " + serialize(code) + ")")