first commit

This commit is contained in:
2026-09-02 11:44:52 +08:00
commit 0c8fa2653e
309 changed files with 57278 additions and 0 deletions
@@ -0,0 +1,401 @@
#!/usr/bin/env python3
"""
宜搭自定义页面 schema 生成/修改(编排:get-schema → 编译 + 构建 → update-schema
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
【调用前必读】references/yida-custom-page-codegen.md
JSX 入口签名 / Hooks 限制 / 行内样式 / 跨表单联动 5 种模式 / SEARCH/REPLACE
增量改写 / 常见坑速查表)。本脚本 --help 仅给出基本用法,**不要**只看
--help 就直接拼 JSX,几乎必踩坑。
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
用法:
python yida_custom_page_update.py --app APP_X --form FORM-XXX --code-file page.jsx --yes
python yida_custom_page_update.py --app APP_X --form FORM-XXX --code 'import ...' --yes
python yida_custom_page_update.py --app APP_X --form FORM-XXX --show-current
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
--show-current 模式:
只拉取现有 schema 并输出当前代码,不做修改。
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
写入模式:
全量替换代码。脚本调用纯 Python 编译管线(JSX→createElement 转换 +
Hooks 兼容层 _customState/didMount),构建标准 Jsx 组件 schema,并保留
page_id 和已有 dataSource。**零第三方依赖**(仅需 Python 3.7+ 标准库,
无需 pip install、无需 Node.js)。空页面和已有代码的页面均可使用。
跨表单联动场景:JSX 内可通过 Yida.api.form.* 直接读写同应用内任意表单,
搭配 `yida_form_inspector.py --action fields-snippet` 取目标表的字段常量片段。
详见 references/yida-custom-page-codegen.md §9。
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
"""
from __future__ import annotations
import argparse
import json
import os
import subprocess
import sys
from pathlib import Path
from typing import Any
_SCRIPT_DIR = Path(__file__).resolve().parent
if str(_SCRIPT_DIR) not in sys.path:
sys.path.insert(0, str(_SCRIPT_DIR))
from yida_page_compiler import compile_jsx_to_schema # noqa: E402
from yida_page_schema import extract_source_code # noqa: E402
from yida_jsx_pipeline import field_check, lint_check # noqa: E402
MAX_CODE_FILE_SIZE = 1 * 1024 * 1024
MAX_INLINE_CODE = 200 * 1024
def _gather_allowed_roots() -> list[Path]:
"""收集所有允许的路径根目录。任一命中即放行,详见 _resolve_safe_path。
优先级(靠前的优先,仅影响报错提示顺序):
1. OPENYIDA_ALLOWED_ROOTS:显式多根,以 os.pathsep / ':' / ',' 分隔。
2. OPENCLAW_WORKSPACE:老环境变量,向后兼容。
3. 当前工作目录 cwd:兼容原行为。
4. 临时目录 tempdir:供脚本中转使用。
"""
roots: list[Path] = []
extra = os.environ.get("OPENYIDA_ALLOWED_ROOTS", "")
if extra:
seps = [os.pathsep, ":", ","]
parts: list[str] = [extra]
for sep in seps:
parts = [seg for chunk in parts for seg in chunk.split(sep)]
for part in parts:
part = part.strip()
if part:
roots.append(Path(part).expanduser().resolve())
legacy = os.environ.get("OPENCLAW_WORKSPACE")
if legacy:
roots.append(Path(legacy).expanduser().resolve())
roots.append(Path.cwd().resolve())
import tempfile as _tempfile
roots.append(Path(_tempfile.gettempdir()).resolve())
roots.append(Path("/tmp").resolve())
roots.append(Path("/private/tmp").resolve())
# 去重保序
seen: set[str] = set()
uniq: list[Path] = []
for r in roots:
s = str(r)
if s not in seen:
seen.add(s)
uniq.append(r)
return uniq
def _resolve_safe_path(path_str: str) -> Path:
target = Path(path_str).expanduser()
target = target.resolve() if target.is_absolute() else (Path.cwd() / target).resolve()
roots = _gather_allowed_roots()
for root in roots:
try:
target.relative_to(root)
return target
except ValueError:
continue
listing = "\n - ".join(str(r) for r in roots)
raise ValueError(
f"路径超出允许范围:{path_str}\n"
f"已尝试的允许根目录:\n - {listing}\n"
f"提示:设置 OPENYIDA_ALLOWED_ROOTS(允许多根,冒号/逗号分隔)或 OPENCLAW_WORKSPACE 扩展允许范围。"
)
def _run_dws(args: list[str], dry_run: bool = False) -> Any | None:
cmd = ["dws"] + args
if dry_run:
print(f" [dry-run] {' '.join(cmd)}")
return {"dry_run": True}
try:
result = subprocess.run(cmd, capture_output=True, text=True, timeout=120)
except FileNotFoundError:
print(" [FAIL] 找不到 'dws' 命令", file=sys.stderr)
return None
except subprocess.TimeoutExpired:
print(" [FAIL] dws 超时", file=sys.stderr)
return None
if result.returncode != 0:
err = result.stderr.strip() or result.stdout.strip()
print(f" [FAIL] dws 失败 (exit {result.returncode}): {err}", file=sys.stderr)
return None
try:
return json.loads(result.stdout)
except json.JSONDecodeError as e:
print(f" [FAIL] 非 JSON: {e}\n 输出: {result.stdout[:300]}", file=sys.stderr)
return None
def _unwrap_content(data: Any) -> Any:
"""兼容 dws JSON 输出的常见包裹层。"""
current = data
for _ in range(4):
if not isinstance(current, dict):
return current
if "content" in current:
current = current["content"]
continue
if "data" in current and isinstance(current["data"], dict):
current = current["data"]
continue
return current
return current
def _extract_form_type(info: Any) -> str:
"""从 get-info 的不同返回形态中提取 formType/type。"""
candidates: list[Any] = []
current = info
for _ in range(4):
if not isinstance(current, dict):
break
candidates.append(current)
next_obj = None
for key in ("content", "data", "result"):
value = current.get(key)
if isinstance(value, dict):
next_obj = value
break
if next_obj is None:
break
current = next_obj
for item in candidates:
value = item.get("formType") or item.get("type") or item.get("pageType")
if isinstance(value, str) and value.strip():
return value.strip().lower()
return ""
def _check_display_target(app: str, form: str, force: bool = False) -> bool:
"""发布前确认目标是自定义展示页,避免覆盖普通表单/流程表单。"""
print("Step 0: 校验发布目标")
info = _run_dws(["yida", "design", "form", "get-info", "--app", app,
"--form", form, "--format", "json"])
form_type = _extract_form_type(info)
if form_type == "display":
print(" [OK] 目标类型 display")
return True
if force:
reason = form_type or "unknown"
print(f" [WARN] 目标类型为 {reason},已按 --force 跳过保护")
return True
if not info:
print(" [FAIL] 无法获取目标页面类型,已拒绝写入", file=sys.stderr)
elif form_type:
print(f" [FAIL] 目标 formType={form_type},不是 display 自定义页面,已拒绝写入",
file=sys.stderr)
else:
print(" [FAIL] get-info 返回中未找到 formType,已拒绝写入", file=sys.stderr)
print(" [HINT] 请确认 --form 是 display 页面;确认无误时可加 --force 显式绕过",
file=sys.stderr)
return False
def _load_code(args: argparse.Namespace) -> str:
if args.code_file:
safe = _resolve_safe_path(args.code_file)
if not safe.exists():
raise ValueError(f"文件不存在: {safe}")
if safe.stat().st_size > MAX_CODE_FILE_SIZE:
raise ValueError(f"文件过大 (限制 {MAX_CODE_FILE_SIZE:,} 字节)")
return safe.read_text(encoding="utf-8")
elif args.code:
if len(args.code.encode("utf-8")) > MAX_INLINE_CODE:
raise ValueError(f"--code 过长 (限制 {MAX_INLINE_CODE:,} 字节)")
return args.code
else:
raise ValueError("必须提供 --code-file 或 --code")
def _extract_existing_data_source(schema: dict) -> dict | None:
"""从已有 schema 中提取 Page 组件的 dataSource,用于 merge 保留用户自定义数据源。"""
try:
return schema["pages"][0]["componentsTree"][0].get("dataSource")
except (KeyError, IndexError, TypeError):
return None
def main() -> int:
ap = argparse.ArgumentParser(
description=(
"宜搭自定义页面 schema 生成/修改 "
"【调用前必读】references/yida-custom-page-codegen.md"
"JSX 写法 / Hooks 限制 / 跨表单联动 / 常见坑),不要只看 --help 就拼 JSX"
),
formatter_class=argparse.RawDescriptionHelpFormatter, epilog=__doc__)
ap.add_argument("--app", required=True, help="应用编码 appType")
ap.add_argument("--form", required=True, help="页面 formUuid")
ap.add_argument("--code-file", help="新代码文件路径")
ap.add_argument("--code", help="新代码内联字符串")
ap.add_argument("--show-current", action="store_true", help="只输出当前代码不修改")
ap.add_argument("--yes", action="store_true", help="确认写入")
ap.add_argument("--dry-run", action="store_true", help="只编译不写入")
ap.add_argument("--skip-field-check", action="store_true",
help="跳过字段 ID 对账预检(不推荐)")
ap.add_argument("--skip-lint", action="store_true",
help="跳过 JSX 静态检查(30 条宜搭专属陷阱,不推荐)")
ap.add_argument("--force", action="store_true",
help="跳过发布目标 formType=display 保护(仅确认目标无误时使用)")
args = ap.parse_args()
if not args.show_current and not args.code_file and not args.code:
print("错误: 必须提供 --code-file / --code 或 --show-current", file=sys.stderr)
return 1
if not args.show_current and not args.dry_run:
if not _check_display_target(args.app, args.form, force=args.force):
return 1
# Step 1: 拉取现有 schema
print("Step 1: 获取现有 schema")
resp = _run_dws(["yida", "design", "form", "get-schema", "--app", args.app,
"--form", args.form, "--format", "json"], dry_run=args.dry_run)
if args.dry_run and not args.show_current:
try:
new_code = _load_code(args)
except ValueError as e:
print(f"错误: {e}", file=sys.stderr)
return 1
result = compile_jsx_to_schema(new_code, form_uuid=args.form)
if not result.get("ok"):
errors = result.get("errors", [])
err_msgs = "; ".join(e.get("message", "") for e in errors)
print(f" [FAIL] 编译失败: {err_msgs}", file=sys.stderr)
lint = result.get("lint", {})
if lint.get("warnings"):
for w in lint["warnings"]:
print(f" [WARN] {w.get('message', w)}", file=sys.stderr)
return 1
schema_json = result["schema"]
print(json.dumps({"ok": True, "dry_run": True, "formUuid": args.form,
"codeSize": len(new_code),
"schemaSize": len(schema_json)}, ensure_ascii=False, indent=2))
return 0
if not resp:
return 1
schema = resp
print(" [OK] 拿到 schema")
# --show-current 模式
if args.show_current:
try:
current_code = extract_source_code(schema)
except (ValueError, TypeError):
current_code = None
if current_code is None:
print(" [WARN] schema 中没有可提取的自定义页面代码")
print(json.dumps({"ok": False, "error": "not_a_custom_page"}, ensure_ascii=False))
return 1
print(json.dumps({"ok": True, "formUuid": args.form,
"codeSize": len(current_code),
"currentCode": current_code}, ensure_ascii=False, indent=2))
return 0
# 获取 page_id
page_id = args.form
pages = schema.get("pages", [])
if pages:
page_id = pages[0].get("id", args.form) or args.form
# 提取已有 dataSource(用于 merge
existing_ds = _extract_existing_data_source(schema)
# Step 2: 加载新代码、编译并构建 schema
try:
new_code = _load_code(args)
except ValueError as e:
print(f"错误: {e}", file=sys.stderr)
return 1
if not new_code.strip():
print("错误: 代码不能为空", file=sys.stderr)
return 1
try:
current_code = extract_source_code(schema)
except (ValueError, TypeError):
current_code = None
previous_size = len(current_code) if current_code else 0
print(f"Step 2: 编译 + 构建 schema (新代码 {len(new_code):,} 字节)")
result = compile_jsx_to_schema(new_code, form_uuid=page_id, existing_data_source=existing_ds)
if not result.get("ok"):
errors = result.get("errors", [])
err_msgs = "; ".join(e.get("message", "") for e in errors)
print(f" [FAIL] 编译失败: {err_msgs}", file=sys.stderr)
lint = result.get("lint", {})
if lint.get("warnings"):
for w in lint["warnings"]:
print(f" [WARN] {w.get('message', w)}", file=sys.stderr)
return 1
schema_json = result["schema"]
lint = result.get("lint", {})
if lint.get("warnings"):
for w in lint["warnings"]:
print(f" [WARN] lint: {w.get('message', w)}")
print(f" [OK] 编译成功, schema 大小: {len(schema_json):,} 字节")
# Step 2.5: 字段 ID 对账预检(避免发布后运行时才报 fieldId 不存在)
if not args.skip_field_check:
print("Step 2.5: 字段 ID 对账")
chk = field_check(new_code, args.app)
for w in chk.get("warnings", []):
print(f" [WARN] {w.get('message', w)}")
if not chk.get("ok"):
print(" [FAIL] 字段对账未通过,为避免发布后页面报错,拒绝写入:", file=sys.stderr)
for e in chk.get("errors", []):
print(f" - {e.get('message', e)}", file=sys.stderr)
print(" [HINT] 修复后重试;确认需要忽略可加 --skip-field-check(不推荐)", file=sys.stderr)
return 1
info = chk.get("info", {})
if info.get("skipped"):
print(f" [OK] 跳过({info['skipped']}")
else:
print(f" [OK] 已校验 {info.get('referencedFieldCount', 0)} 个字段引用,"
f"覆盖 {len(info.get('checkedForms', []))} 张表单")
# Step 2.7: JSX 静态检查(避免发布后运行时才报错)
if not args.skip_lint:
print("Step 2.7: JSX 静态检查")
lr = lint_check(new_code, filename=args.code_file or "page.jsx")
for w in lr.get("warnings", []):
print(f" [WARN] L{w['line']} [{w['rule']}] {w['message']}")
if not lr.get("ok"):
print(" [FAIL] JSX 静态检查未通过,为避免发布后页面报错,拒绝写入:", file=sys.stderr)
for e in lr.get("errors", []):
print(f" L{e['line']} [{e['rule']}] {e['message']}", file=sys.stderr)
print(" [HINT] 修复后重试;确认需要忽略可加 --skip-lint(不推荐)", file=sys.stderr)
print(" [HINT] 或在 JSX 中加 // dws-lint-disable-line [rule] 关闭单行检查", file=sys.stderr)
return 1
info = lr.get("info", {})
print(f" [OK] 检查通过(错误 {info.get('errorCount', 0)} / 警告 {info.get('warningCount', 0)}")
# Step 3: 写回
print("Step 3: 写入 schema")
resp = _run_dws(["yida", "design", "form", "update-schema", "--app", args.app,
"--form", args.form, "--form-type", "display",
"--content", schema_json, "--yes", "--format", "json"])
if not resp:
return 1
print(" [OK] 写入成功")
print(json.dumps({"ok": True, "formUuid": args.form, "codeSize": len(new_code),
"previousCodeSize": previous_size}, ensure_ascii=False))
return 0
if __name__ == "__main__":
sys.exit(main())