You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
249 lines
8.2 KiB
249 lines
8.2 KiB
#!/usr/bin/env python3
|
|
"""校验物件设计 design.md 格式是否符合规范。
|
|
|
|
用法：
|
|
# 校验单个物件
|
|
python validate_props_design.py .mv/{song}/{theme}/props/{prop}/design.md
|
|
|
|
# 校验 props 目录下所有物件
|
|
python validate_props_design.py .mv/{song}/{theme}/props/
|
|
|
|
规则：
|
|
- 必须有一级标题 # 物件名
|
|
- 必须包含 5 个二级标题分区：外观描述、叙事作用、生图设定、候选图、选定图
|
|
- 外观描述必须包含至少 3 项关键词（类型、形态、颜色/材质、尺寸/比例、细节特征）
|
|
- 生图设定必须包含：提示词、模型、参考图
|
|
- 候选图必须有至少 1 条列表项
|
|
- 候选图路径对应的文件必须存在
|
|
"""
|
|
|
|
import re
|
|
import sys
|
|
from pathlib import Path
|
|
|
|
REQUIRED_SECTIONS = ["外观描述", "叙事作用", "生图设定", "候选图", "选定图"]
|
|
PROMPT_FIELDS = ["提示词：", "模型：", "参考图："]
|
|
|
|
# 中英 section / field 映射
|
|
SECTION_ALIASES = {
|
|
"外观描述": "Appearance",
|
|
"叙事作用": "Narrative Role",
|
|
"生图设定": "Image Settings",
|
|
"候选图": "Candidates",
|
|
"选定图": "Selected",
|
|
}
|
|
PROMPT_FIELD_ALIASES = {
|
|
"提示词：": ("Prompt：", "Prompt:"),
|
|
"模型：": ("Model：", "Model:"),
|
|
"参考图：": ("Reference：", "Reference:"),
|
|
}
|
|
IMG_EXTS = {".png", ".jpg", ".jpeg", ".webp"}
|
|
|
|
H1_RE = re.compile(r"^#\s+(.+)$")
|
|
H2_RE = re.compile(r"^##\s+(.+)$")
|
|
LIST_RE = re.compile(r"^-\s+(.+)$")
|
|
CANDIDATE_RE = re.compile(r"^-\s+(candidate_\S+)")
|
|
|
|
|
|
def parse_design(text: str) -> dict:
|
|
"""解析 design.md，返回结构化数据。"""
|
|
lines = text.split("\n")
|
|
result = {"title": None, "sections": {}}
|
|
current_section = None
|
|
|
|
for line in lines:
|
|
stripped = line.strip()
|
|
|
|
h1 = H1_RE.match(stripped)
|
|
if h1:
|
|
result["title"] = h1.group(1)
|
|
continue
|
|
|
|
h2 = H2_RE.match(stripped)
|
|
if h2:
|
|
current_section = h2.group(1)
|
|
result["sections"][current_section] = []
|
|
continue
|
|
|
|
if current_section is not None:
|
|
result["sections"][current_section].append(stripped)
|
|
|
|
return result
|
|
|
|
|
|
def get_section(sections: dict, zh_name: str) -> list:
|
|
"""从 sections 字典中获取内容，兼容中英 section header。"""
|
|
en_name = SECTION_ALIASES.get(zh_name, "")
|
|
return sections.get(zh_name, sections.get(en_name, []))
|
|
|
|
|
|
def has_section(sections: dict, zh_name: str) -> bool:
|
|
"""检查 section 是否存在，兼容中英。"""
|
|
en_name = SECTION_ALIASES.get(zh_name, "")
|
|
return zh_name in sections or en_name in sections
|
|
|
|
|
|
def has_prompt_field(content: str, zh_field: str) -> bool:
|
|
"""检查生图设定字段是否存在，兼容中英。"""
|
|
if zh_field in content:
|
|
return True
|
|
for alias in PROMPT_FIELD_ALIASES.get(zh_field, ()):
|
|
if alias in content:
|
|
return True
|
|
return False
|
|
|
|
|
|
def validate_one(design_path: Path) -> list[str]:
|
|
"""校验单个 design.md，返回问题列表。"""
|
|
issues = []
|
|
prop_dir = design_path.parent
|
|
prop_name = prop_dir.name
|
|
|
|
text = design_path.read_text(encoding="utf-8")
|
|
data = parse_design(text)
|
|
|
|
# 物件名
|
|
if not data["title"]:
|
|
issues.append(f"ERROR [{prop_name}] 缺少一级标题 # 物件名")
|
|
|
|
# 必需分区
|
|
for section in REQUIRED_SECTIONS:
|
|
if not has_section(data["sections"], section):
|
|
en = SECTION_ALIASES.get(section, "")
|
|
issues.append(f"ERROR [{prop_name}] 缺少分区 ## {section}/{en}")
|
|
|
|
# 外观描述：需要包含至少 3 项列表项
|
|
appearance = get_section(data["sections"], "外观描述")
|
|
if appearance:
|
|
items = [l for l in appearance if LIST_RE.match(l)]
|
|
if len(items) < 3:
|
|
issues.append(
|
|
f"WARN [{prop_name}] 外观描述/Appearance 列表项不足（需至少 3 项，当前 {len(items)} 项）"
|
|
)
|
|
|
|
# 叙事作用：需要有内容
|
|
narrative = get_section(data["sections"], "叙事作用")
|
|
if narrative:
|
|
content = "\n".join(narrative).strip()
|
|
if not content:
|
|
issues.append(f"WARN [{prop_name}] 叙事作用/Narrative Role 内容为空")
|
|
|
|
# 生图设定：需要提示词、模型、参考图
|
|
img_settings = get_section(data["sections"], "生图设定")
|
|
if img_settings:
|
|
content = "\n".join(img_settings)
|
|
for field in PROMPT_FIELDS:
|
|
if not has_prompt_field(content, field):
|
|
issues.append(f"ERROR [{prop_name}] 生图设定/Image Settings 缺少字段 {field}")
|
|
|
|
# 候选图：需要至少 1 条，且文件存在
|
|
candidates_lines = get_section(data["sections"], "候选图")
|
|
if candidates_lines:
|
|
candidates = []
|
|
for line in candidates_lines:
|
|
m = CANDIDATE_RE.match(line)
|
|
if m:
|
|
candidates.append(m.group(1))
|
|
if not candidates:
|
|
issues.append(f"ERROR [{prop_name}] 候选图/Candidates 为空（需要至少 1 张）")
|
|
for fname in candidates:
|
|
actual_name = fname.split()[0] if " " in fname else fname
|
|
fpath = prop_dir / actual_name
|
|
if not fpath.exists():
|
|
issues.append(f"WARN [{prop_name}] 候选图文件不存在: {actual_name}")
|
|
|
|
# 选定图：检查是否已选定
|
|
selected_lines = get_section(data["sections"], "选定图")
|
|
if selected_lines:
|
|
content = "\n".join(selected_lines).strip()
|
|
if not content:
|
|
issues.append(f"INFO [{prop_name}] 选定图/Selected 尚未填写（待用户挑选）")
|
|
elif "selected" in content:
|
|
sel_path = prop_dir / "selected.jpg"
|
|
sel_png = prop_dir / "selected.png"
|
|
if not sel_path.exists() and not sel_png.exists():
|
|
issues.append(f"WARN [{prop_name}] selected.jpg/png 文件不存在")
|
|
|
|
return issues
|
|
|
|
|
|
def find_designs(path: Path) -> list[Path]:
|
|
"""查找所有 design.md 文件。"""
|
|
if path.is_file() and path.name == "design.md":
|
|
return [path]
|
|
if path.is_dir():
|
|
designs = []
|
|
for d in sorted(path.iterdir()):
|
|
if d.is_dir():
|
|
dm = d / "design.md"
|
|
if dm.exists():
|
|
designs.append(dm)
|
|
return designs
|
|
return []
|
|
|
|
|
|
def print_summary(all_issues: dict[str, list[str]]) -> None:
|
|
print("=== 物件设计校验报告 ===\n")
|
|
print(f"物件数量: {len(all_issues)}")
|
|
print()
|
|
|
|
total_errors = 0
|
|
total_warns = 0
|
|
|
|
for prop_name, issues in all_issues.items():
|
|
errors = [i for i in issues if "ERROR" in i]
|
|
warns = [i for i in issues if "WARN" in i]
|
|
infos = [i for i in issues if "INFO" in i]
|
|
total_errors += len(errors)
|
|
total_warns += len(warns)
|
|
|
|
status = "PASS" if not errors else "FAIL"
|
|
print(f" [{status}] {prop_name} — {len(errors)} 错误, {len(warns)} 警告")
|
|
|
|
print()
|
|
if total_errors == 0 and total_warns == 0:
|
|
no_info = all(not any("INFO" in i for i in iss) for iss in all_issues.values())
|
|
if no_info:
|
|
print("ALL PASS")
|
|
else:
|
|
print("ALL PASS (部分物件待用户选图)")
|
|
else:
|
|
all_issues_flat = [i for iss in all_issues.values() for i in iss]
|
|
print(f"发现 {len(all_issues_flat)} 个问题:\n")
|
|
for issue in all_issues_flat:
|
|
print(f" {issue}")
|
|
|
|
|
|
def main():
|
|
if len(sys.argv) < 2:
|
|
print("用法:")
|
|
print(" python validate_props_design.py <design.md 路径>")
|
|
print(" python validate_props_design.py <props 目录路径>")
|
|
sys.exit(1)
|
|
|
|
target = Path(sys.argv[1]).resolve()
|
|
if not target.exists():
|
|
print(f"ERROR: 路径不存在: {target}")
|
|
sys.exit(1)
|
|
|
|
designs = find_designs(target)
|
|
if not designs:
|
|
print(f"ERROR: 未找到 design.md 文件: {target}")
|
|
sys.exit(1)
|
|
|
|
print(f"\n--- 物件设计校验 ---\n")
|
|
|
|
all_issues = {}
|
|
for dp in designs:
|
|
prop_name = dp.parent.name
|
|
issues = validate_one(dp)
|
|
all_issues[prop_name] = issues
|
|
|
|
print_summary(all_issues)
|
|
|
|
has_error = any("ERROR" in i for iss in all_issues.values() for i in iss)
|
|
sys.exit(1 if has_error else 0)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|
|
|