77 lines
2.2 KiB
Python
Executable File
77 lines
2.2 KiB
Python
Executable File
"""
|
|
带可复现性支持的 pipeline 包装器
|
|
|
|
使用方法:
|
|
from gfa_core.pipeline_with_repro import run_stage_with_repro
|
|
|
|
# 替换原来的 run_stage 调用
|
|
result = run_stage_with_repro(stage, job, prefix, cpu=cpu, ...)
|
|
"""
|
|
|
|
import sys
|
|
from pathlib import Path
|
|
from typing import Tuple, Optional
|
|
|
|
# 添加项目路径
|
|
sys.path.insert(0, str(Path(__file__).parent.parent))
|
|
|
|
from apps.web.pipeline_backend import run_stage as _original_run_stage, JobInfo, WORKSPACE_ROOT
|
|
from gfa_core.reproducibility import ReproducibilityBundle
|
|
|
|
|
|
def run_stage_with_repro(
|
|
stage: str,
|
|
job: JobInfo,
|
|
prefix: str,
|
|
cpu: int = 8,
|
|
required_pfams: Optional[str] = None,
|
|
ref_species: Optional[list] = None,
|
|
retention_days: int = 7,
|
|
**kwargs
|
|
) -> Tuple[str, Optional[Path]]:
|
|
"""
|
|
带可复现性支持的 run_stage 包装器
|
|
|
|
在原始 run_stage 基础上，自动记录命令和生成可复现性文件
|
|
"""
|
|
|
|
# 初始化可复现性 bundle
|
|
reproducibility_bundle = None
|
|
try:
|
|
reproducibility_bundle = ReproducibilityBundle(
|
|
job_id=job.job_id,
|
|
prefix=prefix,
|
|
workspace_root=str(WORKSPACE_ROOT)
|
|
)
|
|
except Exception as e:
|
|
print(f"[WARN] ReproducibilityBundle init failed: {e}")
|
|
|
|
# 调用原始 run_stage
|
|
result_msg, manifest = _original_run_stage(
|
|
stage=stage,
|
|
job=job,
|
|
prefix=prefix,
|
|
cpu=cpu,
|
|
required_pfams=required_pfams,
|
|
ref_species=ref_species,
|
|
retention_days=retention_days,
|
|
**kwargs
|
|
)
|
|
|
|
# 如果成功，生成可复现性文件
|
|
if reproducibility_bundle and "✅" in result_msg:
|
|
try:
|
|
reproducibility_bundle.generate_commands_sh()
|
|
reproducibility_bundle.generate_environment_yml()
|
|
reproducibility_bundle.generate_checksums()
|
|
reproducibility_bundle.copy_to_project_reproducibility()
|
|
print(f"[INFO] Reproducibility files generated for {stage}")
|
|
except Exception as e:
|
|
print(f"[WARN] Failed to generate reproducibility files: {e}")
|
|
|
|
return result_msg, manifest
|
|
|
|
|
|
# 保持向后兼容
|
|
run_stage = run_stage_with_repro
|