166 lines
5.7 KiB
Python
166 lines
5.7 KiB
Python
from __future__ import annotations
|
|
|
|
import json
|
|
import logging
|
|
import os
|
|
import subprocess
|
|
import sys
|
|
import time
|
|
from pathlib import Path
|
|
|
|
from .job_manager import JobManager
|
|
from .log_config import get_logger
|
|
from .schemas import JobPaths
|
|
|
|
log = get_logger("service.runner")
|
|
|
|
|
|
STAGE_RULES: list[tuple[str, str, int]] = [
|
|
("--- 1. 读取数据 ---", "reading_data", 10),
|
|
("--- 2. 智能画幅计算 ---", "sizing_canvas", 20),
|
|
("--- 3. 生成掩膜", "building_mask", 35),
|
|
("--- 4. 计算权重 ---", "computing_weights", 45),
|
|
("--- 5. 启动生成", "placing_words", 65),
|
|
("--- 6. 高清渲染 ---", "rendering", 85),
|
|
("已保存:", "writing_outputs", 92),
|
|
("✅ 完成", "writing_outputs", 99),
|
|
]
|
|
|
|
|
|
class JobRunner:
|
|
def __init__(self, project_root: Path, manager: JobManager) -> None:
|
|
self.project_root = project_root
|
|
self.manager = manager
|
|
self.script_path = self.project_root / "wordcloud_generate_hybrid.py"
|
|
|
|
def _parse_stage(self, line: str, current_stage: str, current_progress: int) -> tuple[str, int]:
|
|
for token, stage, progress in STAGE_RULES:
|
|
if token in line:
|
|
return stage, progress
|
|
|
|
if "尝试 #" in line or "尺度" in line or "二分重试" in line:
|
|
progress = max(current_progress, 70)
|
|
return "placing_words", min(progress + 1, 84)
|
|
|
|
return current_stage, current_progress
|
|
|
|
def run(self, job_id: str, paths: JobPaths, config: dict) -> None:
|
|
t_start = time.time()
|
|
log.info("=" * 50)
|
|
log.info("[Runner] 任务启动 job_id=%s", job_id)
|
|
log.info(" config_path = %s", paths.config_path)
|
|
log.info(" output_dir = %s", paths.output_dir)
|
|
log.info(" 子进程 python = %s", sys.executable)
|
|
self.manager.set_status(job_id, status="running", stage="starting", progress_percent=1, message="任务启动")
|
|
|
|
with paths.config_path.open("w", encoding="utf-8") as f:
|
|
json.dump(config, f, ensure_ascii=False, indent=2)
|
|
log.info(" 配置文件已写入")
|
|
|
|
cmd = [
|
|
sys.executable,
|
|
str(self.script_path),
|
|
"--config",
|
|
str(paths.config_path),
|
|
]
|
|
|
|
env = os.environ.copy()
|
|
process = subprocess.Popen(
|
|
cmd,
|
|
cwd=str(self.project_root),
|
|
stdout=subprocess.PIPE,
|
|
stderr=subprocess.STDOUT,
|
|
text=True,
|
|
bufsize=1,
|
|
env=env,
|
|
)
|
|
|
|
stage = "starting"
|
|
progress = 1
|
|
|
|
assert process.stdout is not None
|
|
for raw in process.stdout:
|
|
line = raw.rstrip("\n")
|
|
log.info("[Pipeline] %s", line)
|
|
stage, progress = self._parse_stage(line, stage, progress)
|
|
self.manager.add_event(
|
|
job_id,
|
|
kind="log",
|
|
stage=stage,
|
|
progress_percent=progress,
|
|
message=line,
|
|
)
|
|
|
|
ret = process.wait()
|
|
elapsed = time.time() - t_start
|
|
log.info("[Runner] 子进程退出 code=%d 耗时=%.2fs", ret, elapsed)
|
|
|
|
png = next(paths.output_dir.glob("*.png"), None)
|
|
svg = next(paths.output_dir.glob("*[!_stroke].svg"), None)
|
|
svg_stroke = next(paths.output_dir.glob("*_stroke.svg"), None)
|
|
db = next(paths.output_dir.glob("*.db"), None)
|
|
metrics = next(paths.output_dir.glob("*metrics*.json"), None)
|
|
|
|
log.info("[Runner] 产物扫描:")
|
|
log.info(" png = %s", png)
|
|
log.info(" svg = %s", svg)
|
|
log.info(" svg_stroke = %s", svg_stroke)
|
|
log.info(" db = %s", db)
|
|
log.info(" metrics = %s", metrics)
|
|
|
|
artifacts = {
|
|
"png": str(png) if png else "",
|
|
"svg": str(svg) if svg else "",
|
|
"svg_stroke": str(svg_stroke) if svg_stroke else "",
|
|
"db": str(db) if db else "",
|
|
"metrics": str(metrics) if metrics else "",
|
|
}
|
|
self.manager.set_artifacts(job_id, artifacts)
|
|
|
|
if ret == 0 and not png:
|
|
log.error("[Runner] 任务失败:退出码=0 但未找到输出图片")
|
|
self.manager.add_event(
|
|
job_id,
|
|
kind="status",
|
|
stage="failed",
|
|
progress_percent=100,
|
|
message="任务失败:未找到输出图片",
|
|
)
|
|
self.manager.set_status(
|
|
job_id,
|
|
status="failed",
|
|
stage="failed",
|
|
progress_percent=100,
|
|
message="任务失败",
|
|
error="missing png artifact",
|
|
)
|
|
return
|
|
|
|
if ret == 0:
|
|
log.info("[Runner] ✅ 任务完成 job_id=%s 总耗时=%.2fs", job_id, elapsed)
|
|
self.manager.add_event(
|
|
job_id,
|
|
kind="status",
|
|
stage="completed",
|
|
progress_percent=100,
|
|
message="任务完成",
|
|
)
|
|
self.manager.set_status(job_id, status="success", stage="completed", progress_percent=100, message="任务完成")
|
|
else:
|
|
log.error("[Runner] ❌ 任务失败 job_id=%s exit_code=%d 耗时=%.2fs", job_id, ret, elapsed)
|
|
self.manager.add_event(
|
|
job_id,
|
|
kind="status",
|
|
stage="failed",
|
|
progress_percent=100,
|
|
message=f"任务失败,退出码: {ret}",
|
|
)
|
|
self.manager.set_status(
|
|
job_id,
|
|
status="failed",
|
|
stage="failed",
|
|
progress_percent=100,
|
|
message="任务失败",
|
|
error=f"script exited with code {ret}",
|
|
)
|