shumengya
50edff80f5
包含 extensions、skills、prompts、settings、auth、models、mcp 等配置。 排除 node_modules、npm 缓存、sessions 等运行时数据。
176 lines
7.7 KiB
Python
176 lines
7.7 KiB
Python
#!/usr/bin/env python3
|
||
"""Create an editable source-file evidence list before code extraction."""
|
||
|
||
from __future__ import annotations
|
||
|
||
import argparse
|
||
from pathlib import Path
|
||
from typing import Any
|
||
|
||
from common import COPYRIGHT_CODE_EXTS, FRONTEND_EXTS, ensure_dir, is_known_config_file, iter_project_files, read_json, rel, write_json
|
||
from extract_code_material import LINES_PER_PAGE, SPLIT_THRESHOLD_PAGES, category_weight, should_skip_file
|
||
|
||
|
||
DEFAULT_MAX_FILES = 0
|
||
|
||
|
||
def evidence_for(path: Path, project: Path) -> str:
|
||
priority, _ = category_weight(path, project)
|
||
if priority == 0:
|
||
return "入口文件证据"
|
||
if priority == 10:
|
||
return "路由文件证据"
|
||
if priority == 20:
|
||
return "页面文件证据"
|
||
if priority == 30:
|
||
return "数据交互文件证据"
|
||
if priority == 40:
|
||
return "状态或数据文件证据"
|
||
if priority == 50:
|
||
return "页面组成文件证据"
|
||
if priority == 60:
|
||
return "通用能力文件证据"
|
||
if priority == 90:
|
||
return "样式文件证据"
|
||
if path.suffix.lower() not in FRONTEND_EXTS:
|
||
return "补充源码证据"
|
||
return "普通源码文件"
|
||
|
||
|
||
def build_candidates(project: Path) -> list[dict[str, Any]]:
|
||
files = [p for p in iter_project_files(project, COPYRIGHT_CODE_EXTS) if not should_skip_file(p) and not is_known_config_file(p)]
|
||
files.sort(key=lambda p: category_weight(p, project))
|
||
candidates: list[dict[str, Any]] = []
|
||
for path in files:
|
||
try:
|
||
line_count = len(path.read_text(encoding="utf-8", errors="replace").splitlines())
|
||
except Exception:
|
||
line_count = 0
|
||
priority, _ = category_weight(path, project)
|
||
candidates.append(
|
||
{
|
||
"path": rel(path, project),
|
||
"selected": False,
|
||
"line_count": line_count,
|
||
"priority": priority,
|
||
"selection_tier": "frontend" if path.suffix.lower() in FRONTEND_EXTS else "supplement",
|
||
"evidence": evidence_for(path, project),
|
||
"model_reason": "",
|
||
}
|
||
)
|
||
return candidates
|
||
|
||
|
||
def selected_line_estimate(item: dict[str, Any]) -> int:
|
||
return int(item.get("line_count") or 0) + 2
|
||
|
||
|
||
def selection_stats(candidates: list[dict[str, Any]]) -> dict[str, int]:
|
||
selected_items = [item for item in candidates if item.get("selected")]
|
||
return {
|
||
"selected_count": len(selected_items),
|
||
"selected_lines": sum(selected_line_estimate(item) for item in selected_items),
|
||
}
|
||
|
||
|
||
def all_candidate_lines(candidates: list[dict[str, Any]]) -> int:
|
||
return sum(selected_line_estimate(item) for item in candidates)
|
||
|
||
|
||
def write_selection_md(path: Path, data: dict[str, Any]) -> None:
|
||
lines = [
|
||
"# 代码文件候选清单",
|
||
"",
|
||
"请先确认要抽取哪些源码文件，再运行代码材料抽取。",
|
||
"",
|
||
"本清单只列出候选源码证据，不默认决定抽取文件。",
|
||
"模型需要先理解项目业务、页面入口和源码职责，再填写 `selected/model_reason`。",
|
||
f"当前已选约 {data['estimated_selected_pages']} 页，全部候选源码约 {data['estimated_all_candidate_pages']} 页。",
|
||
"",
|
||
"```text",
|
||
"STOP_FOR_USER",
|
||
"NEXT_ACTION: 请由模型先填写 草稿/代码文件选择.json 的抽取选择和选择理由，再让用户确认；确认后运行 confirm_stage.py --stage code-selection。",
|
||
"```",
|
||
"",
|
||
"确认方式：",
|
||
"",
|
||
"1. 模型根据项目业务和代码入口选择最能体现软件功能的文件。",
|
||
"2. 把需要抽取的文件设为 `selected: true`，并填写 `model_reason`。",
|
||
"3. 代码材料按完整文件原样复制，不支持只抽取某个文件的中间行段。",
|
||
"4. 用户确认模型选择后，再记录 `code-selection` 门禁。",
|
||
"",
|
||
"## 默认选中文件",
|
||
"",
|
||
"| 文件 | 行数 | 模型选择理由 |",
|
||
"| --- | ---: | --- |",
|
||
]
|
||
for item in data["files"]:
|
||
if item.get("selected"):
|
||
lines.append(f"| `{item['path']}` | {item['line_count']} | {item.get('model_reason') or '待模型填写'} |")
|
||
|
||
lines.extend(["", "## 未选候选文件", "", "| 文件 | 行数 | 证据类型 |", "| --- | ---: | --- |"])
|
||
for item in data["files"]:
|
||
if not item.get("selected"):
|
||
lines.append(f"| `{item['path']}` | {item['line_count']} | {item['evidence']} |")
|
||
path.write_text("\n".join(lines) + "\n", encoding="utf-8")
|
||
|
||
|
||
def main() -> None:
|
||
parser = argparse.ArgumentParser()
|
||
parser.add_argument("--project", required=True)
|
||
parser.add_argument("--analysis", help="Optional project analysis JSON; retained for workflow traceability")
|
||
parser.add_argument("--out-dir", default="软件著作权申请资料/草稿")
|
||
parser.add_argument("--max-files", type=int, default=DEFAULT_MAX_FILES, help="Only limits candidate inventory size; does not auto-select files")
|
||
parser.add_argument("--target-pages", type=int, default=SPLIT_THRESHOLD_PAGES)
|
||
parser.add_argument("--lines-per-page", type=int, default=LINES_PER_PAGE)
|
||
args = parser.parse_args()
|
||
|
||
project = Path(args.project)
|
||
if not project.exists():
|
||
raise SystemExit(f"Project not found: {project}")
|
||
if args.analysis and not Path(args.analysis).exists():
|
||
raise SystemExit(f"Analysis JSON not found: {args.analysis}")
|
||
|
||
out_dir = ensure_dir(Path(args.out_dir))
|
||
candidates = build_candidates(project)
|
||
target_lines = max(1, args.target_pages) * max(1, args.lines_per_page)
|
||
if args.max_files:
|
||
candidates = candidates[: args.max_files]
|
||
stats = selection_stats(candidates)
|
||
candidate_lines = all_candidate_lines(candidates)
|
||
selected_pages = (stats["selected_lines"] + args.lines_per_page - 1) // args.lines_per_page if stats["selected_lines"] else 0
|
||
all_pages = (candidate_lines + args.lines_per_page - 1) // args.lines_per_page if candidate_lines else 0
|
||
data = {
|
||
"project_root": str(project.resolve()),
|
||
"selection_required": True,
|
||
"model_selection_required": True,
|
||
"confirmation_required": True,
|
||
"user_confirmed": False,
|
||
"target_pages": args.target_pages,
|
||
"lines_per_page": args.lines_per_page,
|
||
"target_lines": target_lines,
|
||
"estimated_selected_lines": stats["selected_lines"],
|
||
"estimated_selected_pages": selected_pages,
|
||
"estimated_all_candidate_lines": candidate_lines,
|
||
"estimated_all_candidate_pages": all_pages,
|
||
"supplement_rule": "模型优先选择能体现软件核心功能和真实运行逻辑的源码；不足60页时再从其他相关源码中补充；候选源码仍不足时才生成全部代码材料。",
|
||
"confirmation_stage": "code-selection",
|
||
"next_action": "请由模型填写 草稿/代码文件选择.json 的抽取选择和选择理由，再让用户确认；确认后运行 confirm_stage.py --stage code-selection。",
|
||
"instructions": "The script only inventories source files. The model must choose selected/model_reason before user confirmation. Selected files are copied in full.",
|
||
"files": candidates,
|
||
}
|
||
write_json(out_dir / "代码文件选择.json", data)
|
||
write_selection_md(out_dir / "代码文件候选清单.md", data)
|
||
selected_count = sum(1 for item in candidates if item.get("selected"))
|
||
print(f"OK code selection draft: {out_dir}")
|
||
print(f"Candidates: {len(candidates)}")
|
||
print(f"Model selected: {selected_count}")
|
||
print(f"Estimated selected pages: {selected_pages}")
|
||
print(f"Estimated all candidate pages: {all_pages}")
|
||
print("STOP_FOR_USER")
|
||
print(f"NEXT_ACTION: {data['next_action']}")
|
||
|
||
|
||
if __name__ == "__main__":
|
||
main()
|