MCP Appium

from __future__ import annotations import re from dataclasses import dataclass from typing import Dict, Iterable, List, Sequence, Tuple from xml.etree import ElementTree as ET BOUNDS_PATTERN = re.compile(r"\[(\d+),(\d+)\]\[(\d+),(\d+)\]") @dataclass class NodeSnapshot: text: str = "" content_desc: str = "" resource_id: str = "" class_name: str = "" package: str = "" bounds: Tuple[int, int, int, int] = (0, 0, 0, 0) def prompt_line(self, index: int) -> str: text = self.text or self.content_desc or "<empty>" return ( f"{index}. text='{text}' desc='{self.content_desc or '-'}' " f"id='{self.resource_id or '-'}' class='{self.class_name}' bounds={self.bounds}" ) def as_dict(self) -> Dict[str, str | Tuple[int, int, int, int]]: return { "text": self.text, "content_desc": self.content_desc, "resource_id": self.resource_id, "class_name": self.class_name, "bounds": self.bounds, } def preferred_locator(self) -> Dict[str, str] | None: if self.resource_id: return {"strategy": "id", "value": self.resource_id} if self.content_desc: return {"strategy": "accessibility_id", "value": self.content_desc} if self.text: return {"strategy": "text", "value": self.text} return None class NodeParser: """Extracts meaningful node summaries from Appium XML or adb dumps.""" @staticmethod def parse_xml(xml_payload: str) -> List[NodeSnapshot]: try: tree = ET.fromstring(xml_payload) except ET.ParseError: return [] snapshots: List[NodeSnapshot] = [] for element in tree.iter(): if not element.attrib: continue snapshots.append( NodeSnapshot( text=element.attrib.get("text", ""), content_desc=element.attrib.get("content-desc", ""), resource_id=element.attrib.get("resource-id", ""), class_name=element.attrib.get("class", ""), package=element.attrib.get("package", ""), bounds=_parse_bounds(element.attrib.get("bounds", "")), ) ) return snapshots @staticmethod def parse_accessibility_dump(payload: str) -> List[NodeSnapshot]: """Very lightweight parser for `adb shell dumpsys accessibility` output.""" snapshots: List[NodeSnapshot] = [] current: NodeSnapshot | None = None for line in payload.splitlines(): stripped = line.strip() if stripped.startswith(("View[", "AccessibilityNodeInfo[")): if current: snapshots.append(current) current = NodeSnapshot() continue if current is None: continue if stripped.startswith("text:"): current.text = stripped.split(":", 1)[1].strip() elif stripped.startswith("contentDescription:"): current.content_desc = stripped.split(":", 1)[1].strip() elif stripped.startswith("resourceName:"): current.resource_id = stripped.split(":", 1)[1].strip() elif stripped.startswith("className:"): current.class_name = stripped.split(":", 1)[1].strip() elif stripped.startswith("packageName:"): current.package = stripped.split(":", 1)[1].strip() elif stripped.startswith("boundsInScreen:"): current.bounds = _parse_bounds_from_dump(stripped) if current: snapshots.append(current) return snapshots def summarize_nodes(nodes: Sequence[NodeSnapshot], limit: int = 40) -> str: excerpt = list(nodes[:limit]) return "\n".join(node.prompt_line(idx + 1) for idx, node in enumerate(excerpt)) def _parse_bounds(raw: str) -> Tuple[int, int, int, int]: match = BOUNDS_PATTERN.search(raw) if not match: return (0, 0, 0, 0) return tuple(int(match.group(i)) for i in range(1, 5)) # type: ignore[return-value] def _parse_bounds_from_dump(line: str) -> Tuple[int, int, int, int]: coords = [int(value) for value in re.findall(r"-?\d+", line)] if len(coords) >= 4: return tuple(coords[:4]) # type: ignore[return-value] return (0, 0, 0, 0)