svg-backend/backend/app/services/svg_normalizer.py

import json
import re
from typing import Any
from xml.etree import ElementTree as StdET


SHAPE_TAGS = {"rect", "circle", "ellipse", "path", "polygon", "polyline", "line"}
CONTAINER_TAGS = {"g"}
TEXT_TAGS = {"text", "tspan"}


def _local_name(tag: str) -> str:
    if "}" in tag:
        return tag.split("}", 1)[1]
    return tag


def _parse_classes(value: str | None) -> list[str]:
    if not value:
        return []
    return [item for item in value.strip().split() if item]


def _to_float(value: str | None) -> float | None:
    if value is None or value == "":
        return None
    try:
        return float(value)
    except ValueError:
        return None


def _infer_kind(element_id: str | None, classes: list[str], tag: str) -> str:
    haystack = " ".join([element_id or ""] + classes).lower()

    if "seat" in haystack or "place" in haystack:
        return "seat"
    if "sector" in haystack or "zone" in haystack:
        return "sector"
    if "group" in haystack:
        return "group"
    if tag in SHAPE_TAGS:
        return "shape"
    if tag in CONTAINER_TAGS:
        return "container"
    if tag in TEXT_TAGS:
        return "text"
    return "other"


def _extract_prefixed_id(value: str | None, prefix: str) -> str | None:
    if not value:
        return None
    low = value.lower()
    pref = f"{prefix}-"
    if low.startswith(pref):
        return value[len(pref):]
    return None


def _extract_seat_parts_from_id(value: str | None) -> tuple[str | None, str | None]:
    if not value:
        return None, None

    patterns = [
        r"^seat[-_]?([a-zA-Z]+)[-_]?(\d+)$",
        r"^place[-_]?([a-zA-Z]+)[-_]?(\d+)$",
        r"^([a-zA-Z]+)[-_]?(\d+)$",
    ]

    for pattern in patterns:
        match = re.match(pattern, value)
        if match:
            return match.group(1), match.group(2)

    return None, None


def _build_parent_map(root) -> dict[int, dict[str, str | None]]:
    parent_map: dict[int, dict[str, str | None]] = {}

    def walk(node, current_sector_id: str | None, current_group_id: str | None):
        node_id = node.attrib.get("id")
        classes = _parse_classes(node.attrib.get("class"))
        kind = _infer_kind(node_id, classes, _local_name(node.tag))

        sector_id = current_sector_id
        group_id = current_group_id

        explicit_sector = node.attrib.get("data-sector-id") or _extract_prefixed_id(node_id, "sector")
        explicit_group = node.attrib.get("data-group-id") or _extract_prefixed_id(node_id, "group")

        if kind == "sector":
            sector_id = explicit_sector or node_id or sector_id

        if kind == "group":
            group_id = explicit_group or node_id or group_id

        parent_map[id(node)] = {
            "sector_id": sector_id,
            "group_id": group_id,
        }

        for child in list(node):
            walk(child, sector_id, group_id)

    walk(root, None, None)
    return parent_map


def normalize_svg_bytes(content: bytes) -> dict[str, Any]:
    root = StdET.fromstring(content)
    parent_map = _build_parent_map(root)

    elements: list[dict[str, Any]] = []
    seats: list[dict[str, Any]] = []
    groups: list[dict[str, Any]] = []
    sectors: list[dict[str, Any]] = []

    for node in root.iter():
        tag = _local_name(node.tag)

        if tag == "svg":
            continue

        element_id = node.attrib.get("id")
        classes = _parse_classes(node.attrib.get("class"))
        kind = _infer_kind(element_id=element_id, classes=classes, tag=tag)

        inherited = parent_map.get(id(node), {})
        inherited_sector_id = inherited.get("sector_id")
        inherited_group_id = inherited.get("group_id")

        explicit_sector_id = node.attrib.get("data-sector-id")
        explicit_group_id = node.attrib.get("data-group-id")
        explicit_seat_id = node.attrib.get("data-seat-id")
        explicit_row = node.attrib.get("data-row")
        explicit_seat_number = node.attrib.get("data-seat-number")

        row_from_id, seat_number_from_id = _extract_seat_parts_from_id(element_id)

        seat_id = explicit_seat_id or (element_id if kind == "seat" else None)
        sector_id = explicit_sector_id or inherited_sector_id
        group_id = explicit_group_id or inherited_group_id
        row = explicit_row or row_from_id
        seat_number = explicit_seat_number or seat_number_from_id

        item = {
            "id": element_id,
            "tag": tag,
            "kind": kind,
            "classes": classes,
            "x": _to_float(node.attrib.get("x")),
            "y": _to_float(node.attrib.get("y")),
            "cx": _to_float(node.attrib.get("cx")),
            "cy": _to_float(node.attrib.get("cy")),
            "width": _to_float(node.attrib.get("width")),
            "height": _to_float(node.attrib.get("height")),
            "href": node.attrib.get("href") or node.attrib.get("{http://www.w3.org/1999/xlink}href"),
            "seat_id": seat_id,
            "sector_id": sector_id,
            "group_id": group_id,
            "row": row,
            "seat_number": seat_number,
        }

        elements.append(item)

        if kind == "seat":
            seats.append(item)
        elif kind == "group":
            groups.append(item)
        elif kind == "sector":
            sectors.append(item)

    return {
        "summary": {
            "elements_count": len(elements),
            "seats_count": len(seats),
            "groups_count": len(groups),
            "sectors_count": len(sectors),
        },
        "contract": {
            "seat_fields": ["seat_id", "sector_id", "group_id", "row", "seat_number"],
            "priority": [
                "data-* attributes",
                "inherited parent sector/group",
                "fallback to element id",
            ],
        },
        "elements": elements,
        "seats": seats,
        "groups": groups,
        "sectors": sectors,
    }


def normalize_svg_bytes_to_json(content: bytes) -> tuple[str, dict[str, Any]]:
    payload = normalize_svg_bytes(content)
    return json.dumps(payload, ensure_ascii=False, indent=2), payload