Kochbuch/recipes/parsing.py

"""Parsing and formatting of ingredient lines, steps and tags.

Recipes are edited as plain text (one ingredient per line, group headers end with ":",
steps separated by blank lines). This module converts between that text and the
structured data stored in the database, and scales amounts for the serving calculator.
"""

import re
from dataclasses import dataclass
from decimal import ROUND_HALF_UP, Decimal, InvalidOperation

MAX_TAGS = 15
MAX_TAG_LENGTH = 40

# Abbreviations that get a canonical spelling.
CANONICAL_UNITS = {
    "g": "g", "gr": "g", "gr.": "g", "gramm": "g",
    "kg": "kg", "kilo": "kg", "kilogramm": "kg",
    "mg": "mg",
    "ml": "ml", "milliliter": "ml",
    "cl": "cl", "dl": "dl",
    "l": "l", "liter": "l", "ltr": "l", "ltr.": "l",
    "el": "EL", "el.": "EL", "essl.": "EL", "esslöffel": "EL", "tbsp": "EL", "tbsp.": "EL",
    "tl": "TL", "tl.": "TL", "teel.": "TL", "teelöffel": "TL", "tsp": "TL", "tsp.": "TL",
    "msp": "Msp.", "msp.": "Msp.", "messerspitze": "Msp.", "messerspitzen": "Msp.",
    "pck": "Pck.", "pck.": "Pck.", "pkg": "Pck.", "pkg.": "Pck.", "päckchen": "Pck.", "packung": "Pck.",
    "stk": "Stück", "stk.": "Stück", "st.": "Stück", "stück": "Stück",
    "prise": "Prise", "prisen": "Prise",
    "cup": "Cup", "cups": "Cup", "oz": "oz", "lb": "lb", "lbs": "lb",
}
# Unit words kept as written (plural forms matter in German).
UNIT_WORDS = {
    "bund", "dose", "dosen", "zehe", "zehen", "tasse", "tassen", "becher", "glas", "gläser",
    "scheibe", "scheiben", "handvoll", "zweig", "zweige", "blatt", "blätter", "schuss",
    "tropfen", "würfel", "stange", "stangen", "knolle", "knollen", "kopf", "köpfe",
    "flasche", "flaschen", "beutel", "tafel", "tafeln", "spritzer", "stiel", "stiele",
}

FRACTION_CHARS = {"½": Decimal("0.5"), "¼": Decimal("0.25"), "¾": Decimal("0.75"), "⅓": Decimal(1) / 3,
                  "⅔": Decimal(2) / 3, "⅛": Decimal("0.125")}
NICE_FRACTIONS = [(Decimal("0.125"), "⅛"), (Decimal("0.25"), "¼"), (Decimal(1) / 3, "⅓"), (Decimal("0.5"), "½"),
                  (Decimal(2) / 3, "⅔"), (Decimal("0.75"), "¾")]

_FRAC = "".join(FRACTION_CHARS)
_NUMBER = rf"(?:\d+\s+\d+\s*/\s*\d+|\d+\s*/\s*\d+|\d+\s*[{_FRAC}]|[{_FRAC}]|\d+(?:[.,]\d+)?)"
_AMOUNT_RE = re.compile(rf"^(?P<a>{_NUMBER})(?:\s*(?:-|–|bis)\s*(?P<b>{_NUMBER}))?(?=[\s\D]|$)", re.IGNORECASE)
_BULLET_RE = re.compile(r"^\s*(?:[-*•·▢□◦]|\d+[.)](?=\s))\s*")
_STEP_NUMBER_RE = re.compile(r"^\s*(?:schritt\s*)?\d+\s*[.):]\s*", re.IGNORECASE)


@dataclass
class ParsedIngredient:
    name: str
    amount: Decimal | None = None
    amount_max: Decimal | None = None
    unit: str = ""
    note: str = ""
    group: str = ""


def parse_number(text):
    text = text.strip().replace(" ", " ")
    if not text:
        return None
    try:
        if "/" in text:
            parts = text.split()
            whole = Decimal(parts[0]) if len(parts) == 2 else Decimal(0)
            num, den = (parts[-1] if len(parts) == 2 else text).replace(" ", "").split("/")
            if Decimal(den) == 0:
                return None
            return whole + Decimal(num) / Decimal(den)
        if text[-1] in FRACTION_CHARS:
            whole = text[:-1].strip()
            return (Decimal(whole) if whole else Decimal(0)) + FRACTION_CHARS[text[-1]]
        return Decimal(text.replace(",", "."))
    except (InvalidOperation, ValueError):
        return None


def _split_unit(rest):
    """Split a leading unit off the remaining text. Returns (unit, remainder)."""
    match = re.match(r"^(\S+)(?:\s+(.*))?$", rest)
    if not match:
        return "", rest
    word, remainder = match.group(1), (match.group(2) or "").strip()
    if not remainder:
        return "", rest  # "2 Eier": the only word is the ingredient, not a unit
    key = word.lower()
    if key in CANONICAL_UNITS:
        return CANONICAL_UNITS[key], remainder
    if key in UNIT_WORDS:
        return word, remainder
    return "", rest


def parse_ingredient_line(line, group=""):
    line = _BULLET_RE.sub("", line.strip()).strip()
    if not line:
        return None
    amount = amount_max = None
    rest = line
    match = _AMOUNT_RE.match(line)
    if match:
        amount = parse_number(match.group("a"))
        amount_max = parse_number(match.group("b")) if match.group("b") else None
        if amount is not None:
            rest = line[match.end():].strip()
            # "200g Mehl": unit glued to the number
            glued = re.match(r"^([a-zA-ZäöüÄÖÜ.]+)\s+(.*)$", rest)
            if glued and glued.group(1).lower() in CANONICAL_UNITS and not line[match.end():match.end() + 1].isspace():
                unit, rest = CANONICAL_UNITS[glued.group(1).lower()], glued.group(2)
            else:
                unit, rest = _split_unit(rest)
        else:
            unit = ""
    else:
        unit = ""
    name, note = rest, ""
    if "," in rest:
        name, note = (part.strip() for part in rest.split(",", 1))
    if not name:
        name, note = note, ""
    return ParsedIngredient(name=name, amount=amount, amount_max=amount_max, unit=unit, note=note, group=group)


def is_group_header(line):
    stripped = line.strip()
    return stripped.endswith(":") and len(stripped) > 1 and not _AMOUNT_RE.match(stripped)


def parse_ingredients(text):
    """Parse the ingredient textarea into a list of ParsedIngredient."""
    items, group = [], ""
    for raw in (text or "").splitlines():
        if not raw.strip():
            continue
        if is_group_header(raw):
            group = _BULLET_RE.sub("", raw.strip())[:-1].strip()[:100]
            continue
        item = parse_ingredient_line(raw, group)
        if item:
            items.append(item)
    return items


def parse_steps(text):
    """Blank lines separate steps; without blank lines every line is a step."""
    text = (text or "").replace("\r\n", "\n").strip()
    if not text:
        return []
    chunks = re.split(r"\n\s*\n", text) if re.search(r"\n\s*\n", text) else text.split("\n")
    steps = []
    for chunk in chunks:
        chunk = " ".join(part.strip() for part in chunk.split("\n") if part.strip())
        chunk = _STEP_NUMBER_RE.sub("", chunk).strip()
        if chunk:
            steps.append(chunk)
    return steps


def normalize_tag(tag):
    tag = re.sub(r"\s+", " ", (tag or "").strip().lstrip("#")).lower()
    return tag[:MAX_TAG_LENGTH]


def parse_tags(text):
    seen, tags = set(), []
    for raw in re.split(r"[,;\n]", text or ""):
        tag = normalize_tag(raw)
        if tag and tag not in seen:
            seen.add(tag)
            tags.append(tag)
    return tags[:MAX_TAGS]


# --- Formatting & scaling ------------------------------------------------------------


def format_amount(value):
    if value is None:
        return ""
    value = Decimal(value)
    whole = int(value)
    frac = value - whole
    for target, glyph in NICE_FRACTIONS:
        if abs(frac - target) < Decimal("0.02"):
            return f"{whole} {glyph}" if whole else glyph
    if frac < Decimal("0.02"):
        return str(whole)
    if frac > Decimal("0.98"):
        return str(whole + 1)
    rounded = value.quantize(Decimal("0.01"), rounding=ROUND_HALF_UP).normalize()
    return f"{rounded:f}".replace(".", ",")


def _round_metric(value):
    """Gram/millilitre amounts: no fractions; coarser rounding for big numbers."""
    if value >= 100:
        step = Decimal(5)
    elif value >= 10:
        step = Decimal(1)
    else:
        return value
    return (value / step).quantize(Decimal(1), rounding=ROUND_HALF_UP) * step


def scale_amount(amount, amount_max, unit, factor):
    """Scale an amount (and optional upper bound of a range) and normalise metric units
    (1250 g -> 1,25 kg). Returns (amount, amount_max, unit)."""
    if amount is None:
        return None, None, unit
    values = [Decimal(v) * Decimal(factor) if v is not None else None for v in (amount, amount_max)]
    if unit in ("g", "ml"):
        values = [_round_metric(v) if v is not None else None for v in values]
        if values[0] >= 1000:
            values = [v / 1000 if v is not None else None for v in values]
            unit = "kg" if unit == "g" else "l"
    return values[0], values[1], unit


def format_ingredient(amount, amount_max, unit, name, note=""):
    parts = []
    if amount is not None:
        parts.append(format_amount(amount) + (f"-{format_amount(amount_max)}" if amount_max is not None else ""))
    if unit:
        parts.append(unit)
    parts.append(name)
    text = " ".join(parts)
    return f"{text}, {note}" if note else text


def ingredients_to_text(ingredients):
    """Inverse of parse_ingredients for the edit form. Accepts objects with the same attributes."""
    lines, group = [], ""
    for ing in ingredients:
        if ing.group != group:
            if lines:
                lines.append("")
            if ing.group:
                lines.append(f"{ing.group}:")
            group = ing.group
        lines.append(format_ingredient(ing.amount, ing.amount_max, ing.unit, ing.name, ing.note))
    return "\n".join(lines)


def steps_to_text(steps):
    return "\n\n".join(steps)