update VideoCompress

VideoCompress/config.json

@@ -1,22 +1,22 @@
 {
     "save_to": "single",
+    "crf": 18,
     "bitrate": null,
-    "crf": 26,
-    "codec": "h264_qsv",
-    "hwaccel": "qsv",
+    "codec": "h264",
+    "hwaccel": null,
     "extra": [],
-    "ffmpeg": "C:/tools/ffmpeg/bin/ffmpeg.exe",
-    "ffprobe": "C:/tools/ffmpeg/bin/ffprobe",
+    "ffmpeg": "ffmpeg",
+    "ffprobe": "ffprobe",
     "manual": null,
     "video_ext": [
         ".mp4",
         ".mkv"
     ],
-    "compress_dir_name": "compress_qsv",
+    "compress_dir_name": "compress",
     "resolution": null,
-    "fps": "30",
+    "fps": 30,
     "test_video_resolution": "1920x1080",
-    "test_video_fps": "30",
+    "test_video_fps": 30,
     "test_video_input": "compress_video_test.mp4",
     "test_video_output": "compressed_video_test.mp4",
     "disable_hwaccel_when_fail": true

VideoCompress/get_frame.py

@@ -122,11 +122,15 @@ def get_video_frame_count(
     }
 
     for key in fallback_order:
+        try:
             try:
                 func = methods.get(key)
                 if not func:
                     continue
                 n = func(path, stream_index)
+            except Exception:
+                logging.debug(f"Errored to get frame with {key}.",exc_info=True)
+                continue
             if isinstance(n, int) and n >= 0:
                 return n
             else:

VideoCompress/main.py

@@ -8,33 +8,35 @@ from time import time
 from rich.logging import RichHandler
 from rich.progress import Progress
 from pickle import dumps, loads
-from typing import Optional, Callable,Literal
+from typing import Optional, Callable, Literal, List, Any, TYPE_CHECKING
 import atexit
 import re
 import get_frame
-from pydantic import BaseModel,Field,field_validator,model_validator
+import json
 
-class Config(BaseModel):
-    save_to:Literal["single","multi"] = Field("single",description="保存到单文件夹，或者每个子文件夹创建compress_dir")
-    crf: Optional[int] = Field(None, ge=0, le=51, description="CRF值，范围0-51")
+try:
+    from pydantic import BaseModel, Field, field_validator, model_validator
+    HAS_PYDANTIC = True
+
+    class Config(BaseModel):
+        save_to: Literal["single", "multi"] = Field("single", description="保存到单文件夹，或者每个子文件夹创建compress_dir")
+        crf: Optional[int] = Field(18, ge=0, le=51, description="CRF值，范围0-51")
         bitrate: Optional[str] = Field(None, description="比特率，格式如: 1000k, 2.5M, 1500B")
-    codec: str = Field("h264",description="ffmpeg的codec，如果使用GPU需要对应设置")
-    hwaccel:Optional[Literal["amf","qsv","cuda"]] = Field(None,description="使用GPU加速")
-    extra:Optional[list[str]] = Field(None,description="插入到ffmpeg输出前的自定义参数")
-    ffmpeg:str = "ffmpeg"
-    ffprobe:str = "ffprobe"
-    manual:Optional[list[str]] = Field(None,description=r"手动设置ffmpeg，命令ffmpeg -i {input} {manual} {output}")
-    video_ext:list[str] = Field([".mp4", ".mkv"],description="视频文件后缀，含.")
-    compress_dir_name:str = Field("compress",description="压缩文件夹名称")
-    resolution: Optional[str] = Field(None,description="统一到特定尺寸，None为不使用缩放")
-    fps:int = Field(30,description="fps",ge=0)
-    test_video_resolution:str = "1920x1080"
-    test_video_fps:int = Field(30,ge=0)
-    test_video_input:str = "compress_video_test.mp4"
-    test_video_output:str =  "compressed_video_test.mp4"
-    disable_hwaccel_when_fail:bool  = Field(True,description="当运行失败时，禁用硬件加速")
-    
-    
+        codec: str = Field("h264", description="ffmpeg的codec，如果使用GPU需要对应设置")
+        hwaccel: Optional[Literal["amf", "qsv", "cuda"]] = Field(None, description="使用GPU加速")
+        extra: List[str] = Field([], description="插入到ffmpeg输出前的自定义参数")
+        ffmpeg: str = "ffmpeg"
+        ffprobe: str = "ffprobe"
+        manual: Optional[List[str]] = Field(None, description=r"手动设置ffmpeg，命令ffmpeg -i {input} {manual} {output}")
+        video_ext: List[str] = Field([".mp4", ".mkv"], description="视频文件后缀，含.")
+        compress_dir_name: str = Field("compress", description="压缩文件夹名称")
+        resolution: Optional[str] = Field(None, description="统一到特定尺寸，None为不使用缩放")
+        fps: int = Field(30, description="fps", ge=0)
+        test_video_resolution: str = "1920x1080"
+        test_video_fps: int = Field(30, ge=0)
+        test_video_input: str = "compress_video_test.mp4"
+        test_video_output: str = "compressed_video_test.mp4"
+        disable_hwaccel_when_fail: bool = Field(True, description="当运行失败时，禁用硬件加速")
 
         @field_validator('bitrate')
         @classmethod
@@ -57,15 +59,12 @@ class Config(BaseModel):
             return v
 
         @field_validator("compress_dir_name")
-    # @field_validator("test_video_input")
-    # @field_validator("test_video_output")
         @classmethod
-    def valid_path(cls, v:str) -> str:
-        if re.search(r'[\\/:*?"<>|\x00-\x1F]',v):
+        def valid_path(cls, v: str) -> str:
+            if re.search(r'[\\/:*?"<>|\x00-\x1F]', v):
                 raise ValueError("某配置不符合目录名语法")
             return v
 
-    
         @model_validator(mode='after')
         def validate_mutual_exclusive(self):
             crf_none = self.crf is None
@@ -77,31 +76,98 @@ class Config(BaseModel):
 
             return self
         
+        def dump(self):
+            return self.model_dump()
+
+except ImportError:
+    HAS_PYDANTIC = False
+    from dataclasses import dataclass, asdict
+    import copy
+    @dataclass
+    class Config:
+        save_to: str = "single"
+        crf: Optional[int] = 18
+        bitrate: Optional[str] = None
+        codec: str = "h264"
+        hwaccel: Optional[str] = None
+        extra: List[str] = []
+        ffmpeg: str = "ffmpeg"
+        ffprobe: str = "ffprobe"
+        manual: Optional[List[str]] = None
+        video_ext: List[str] = [".mp4", ".mkv"]
+        compress_dir_name: str = "compress"
+        resolution: Optional[str] = None
+        fps: int = 30
+        test_video_resolution: str = "1920x1080"
+        test_video_fps: int = 30
+        test_video_input: str = "compress_video_test.mp4"
+        test_video_output: str = "compressed_video_test.mp4"
+        disable_hwaccel_when_fail: bool = True
+
+        def update(self, other):
+            if isinstance(other, dict):
+                d = other
+            elif isinstance(other, Config):
+                d = asdict(other)
+            else:
+                return
+            
+            for k, v in d.items():
+                if hasattr(self, k):
+                    setattr(self, k, v)
+                    
+        def copy(self):
+            return copy.deepcopy(self)
+            
+        def dump(self):
+            return asdict(self)
+
 
 root = None
 if os.environ.get("INSTALL", "0") == "1":
-    CFG_FILE= Path(os.getenv("APPDATA", "C:/")) / "VideoCompress" / "config.json"
+    CFG_FILE = Path(os.getenv("APPDATA", "C:/")) / "VideoCompress" / "config.json"
 else:
-    CFG_FILE= Path(sys.path[0]) / "config.json"
-CFG = {
-    "save_to": "single",
-    "crf": "18",
-    "bitrate": None,
-    "codec": "h264",
-    "hwaccel": None,
-    "extra": [],
-    "ffmpeg": "ffmpeg",
-    "manual": None,
-    "video_ext": [".mp4", ".mkv"],
-    "compress_dir_name": "compress",
-    "resolution": None,
-    "fps": "30",
-    "test_video_resolution": "1920x1080",
-    "test_video_fps": "30",
-    "test_video_input": "compress_video_test.mp4",
-    "test_video_output": "compressed_video_test.mp4",
-    "disable_hwaccel_when_fail": True,
-}
+    CFG_FILE = Path(sys.path[0]) / "config.json"
+
+if CFG_FILE.exists():
+    try:
+        import json
+
+        if HAS_PYDANTIC:
+            assert BaseModel # type: ignore
+            assert issubclass(Config, BaseModel)
+            CFG = Config.model_validate_json(CFG_FILE.read_text())
+        else:
+            assert Config
+            cfg:dict[str, Any] = json.loads(CFG_FILE.read_text())
+            CFG = Config(**cfg)
+        
+        get_frame.ffprobe = CFG.ffprobe
+        logging.debug(CFG)
+    except KeyboardInterrupt as e:
+        raise e
+    except Exception as e:
+        logging.warning("Invalid config file, ignored.")
+        logging.debug(e)
+else:
+    try:
+        if HAS_PYDANTIC:
+            if TYPE_CHECKING:
+                assert BaseModel # type: ignore
+                assert issubclass(Config, BaseModel)
+            CFG = Config() # type: ignore
+            CFG_FILE.write_text(CFG.model_dump_json(indent=4))
+        else:
+            import json
+            if TYPE_CHECKING:
+                assert Config
+                assert asdict # type: ignore
+            CFG = Config() # type: ignore
+            CFG_FILE.write_text(json.dumps(asdict(CFG), indent=4))
+        
+        logging.info("Config file created.")
+    except Exception as e:
+        logging.warning("Failed to create config file.", exc_info=e)
 
 
 def get_cmd(video_path: str | Path, output_file: str | Path) -> list[str]:
@@ -110,23 +176,23 @@ def get_cmd(video_path: str | Path, output_file: str | Path) -> list[str]:
     if isinstance(output_file, Path):
         output_file = str(output_file.resolve())
 
-    if CFG["manual"] is not None:
-        command = [CFG["ffmpeg"], "-hide_banner", "-i", video_path]
-        command.extend(CFG["manual"])
+    if CFG.manual is not None:
+        command = [CFG.ffmpeg, "-hide_banner", "-i", video_path]
+        command.extend(CFG.manual)
         command.append(output_file)
         return command
 
     command = [
-        CFG["ffmpeg"],
+        CFG.ffmpeg,
         "-hide_banner",
     ]
-    if CFG["hwaccel"] is not None:
+    if CFG.hwaccel is not None:
         command.extend(
             [
                 "-hwaccel",
-                CFG["hwaccel"],
+                CFG.hwaccel,
                 "-hwaccel_output_format",
-                CFG["hwaccel"],
+                CFG.hwaccel,
                 
             ]
         )
@@ -137,47 +203,47 @@ def get_cmd(video_path: str | Path, output_file: str | Path) -> list[str]:
         ]
     )
 
-    if CFG["bitrate"] is not None:
+    if CFG.bitrate is not None:
 
-        if CFG["resolution"] is not None:
+        if CFG.resolution is not None:
             command.extend(
                 [
                     "-vf",
-                    f"scale={CFG['resolution']}",
+                    f"scale={CFG.resolution}",
                 ]
             )
         command.extend(
             [
                 "-c:v",
-                CFG["codec"],
+                CFG.codec,
                 "-b:v",
-                CFG["bitrate"],
+                CFG.bitrate,
                 "-r",
-                str(CFG["fps"]),
+                str(CFG.fps),
                 "-y",
             ]
         )
     else:
-        if CFG["resolution"] is not None:
+        if CFG.resolution is not None:
             command.extend(
                 [
                     "-vf",
-                    f"scale={CFG['resolution']}",
+                    f"scale={CFG.resolution}",
                 ]
             )
         command.extend(
             [
                 "-c:v",
-                CFG["codec"],
+                CFG.codec,
                 "-global_quality",
-                str(CFG["crf"]),
+                str(CFG.crf),
                 "-r",
-                str(CFG["fps"]),
+                str(CFG.fps),
                 "-y",
             ]
         )
 
-    command.extend(CFG["extra"])
+    command.extend(CFG.extra)
     command.append(output_file)
     logging.debug(f"Create CMD: {command}")
     return command
@@ -219,7 +285,7 @@ def process_video(
 
     if compress_dir is None:
         # 在视频文件所在目录下创建 compress 子目录（如果不存在）
-        compress_dir = video_path.parent / CFG["compress_dir_name"]
+        compress_dir = video_path.parent / CFG.compress_dir_name
     else:
         assert root
         compress_dir /= video_path.parent.relative_to(root)
@@ -277,7 +343,7 @@ def process_video(
             assert result.stdout is not None
             logging.debug(result.stdout.read())
             logging.debug(total)
-            if CFG["hwaccel"] == "mediacodec" and CFG["codec"] in [
+            if CFG.hwaccel == "mediacodec" and CFG.codec in [
                 "h264_mediacodec",
                 "hevc_mediacodec",
             ]:
@@ -285,32 +351,36 @@ def process_video(
                     "mediacodec硬件加速器已知在较短片段上存在异常，将禁用加速重试。"
                 )
                 output_file.unlink(missing_ok=True)
-                bak = CFG.copy()
-                CFG["hwaccel"] = None
-                CFG["codec"] = "h264" if CFG["codec"] == "h264_mediacodec" else "hevc"
+                bak = CFG.codec, CFG.hwaccel
+                CFG.hwaccel = None
+                CFG.codec = "h264" if CFG.codec == "h264_mediacodec" else "hevc"
                 assert not output_file.exists()
                 ret = process_video(video_path, compress_dir, update_func)
-                CFG.update(bak)
+                CFG.codec, CFG.hwaccel = bak
                 if not ret:
                     logging.error("重试仍然失败。")
                     return False
                 else:
                     return True
-            elif CFG["disable_hwaccel_when_fail"] and CFG["hwaccel"] is not None:
+            elif CFG.disable_hwaccel_when_fail and CFG.hwaccel is not None:
                 logging.info("正在禁用硬件加速器重试，进度条可能发生混乱。")
                 output_file.unlink(missing_ok=True)
-                bak = CFG.copy()
-                CFG["hwaccel"] = None
+                if TYPE_CHECKING:
+                    assert BaseModel # type: ignore
+                    assert isinstance(CFG, BaseModel)
+
+                bak = CFG.codec, CFG.hwaccel
+                CFG.hwaccel = None
                 if (
-                    CFG["codec"].endswith("_mediacodec")
-                    or CFG["codec"].endswith("_qsv")
-                    or CFG["codec"].endswith("_nvenc")
-                    or CFG["codec"].endswith("_amf")
+                    CFG.codec.endswith("_mediacodec")
+                    or CFG.codec.endswith("_qsv")
+                    or CFG.codec.endswith("_nvenc")
+                    or CFG.codec.endswith("_amf")
                 ):
-                    CFG["codec"] = CFG["codec"].split("_")[0]
+                    CFG.codec = CFG.codec.split("_")[0]
                 assert not output_file.exists()
                 ret = process_video(video_path, compress_dir, update_func)
-                CFG.update(bak)
+                CFG.codec, CFG.hwaccel = bak
                 if not ret:
                     logging.error("重试仍然失败。")
                     return False
@@ -329,7 +399,7 @@ def process_video(
 
 
 def traverse_directory(root_dir: Path):
-    video_extensions = set(CFG["video_ext"])
+    video_extensions = set(CFG.video_ext)
     sm = None
     # 获取视频文件列表和帧数信息
     video_files:list[Path] = []
@@ -338,8 +408,8 @@ def traverse_directory(root_dir: Path):
         d = que.pop()
         for file in d.glob("*") if d.is_dir() else [d]:
             if (
-                file.parent.name == CFG["compress_dir_name"]
-                or file.name == CFG["compress_dir_name"]
+                file.parent.name == CFG.compress_dir_name
+                or file.name == CFG.compress_dir_name
             ):
                 continue
             if file.is_file() and file.suffix.lower() in video_extensions:
@@ -426,9 +496,9 @@ def traverse_directory(root_dir: Path):
                     )
                     prog.update(main_task, completed=completed_start + x)
 
-            if CFG["save_to"] == "single":
+            if CFG.save_to == "single":
                 process_video(
-                    file, root_dir / CFG["compress_dir_name"], update_progress
+                    file, root_dir / CFG.compress_dir_name, update_progress
                 )
             else:
                 process_video(file, None, update_progress)
@@ -450,7 +520,7 @@ def test():
 
     try:
         subprocess.run(
-            [CFG["ffmpeg"], "-version"], stdout=-3, stderr=-3
+            [CFG.ffmpeg, "-version"], stdout=-3, stderr=-3
         ).check_returncode()
     except Exception as e:
         print(__file__)
@@ -458,7 +528,7 @@ def test():
         exit(-1)
     try:
         ret = subprocess.run(
-            f"{CFG['ffmpeg']} -hide_banner -f lavfi -i testsrc=duration=1:size={CFG['test_video_resolution']}:rate={CFG['test_video_fps']} -c:v libx264 -y -pix_fmt yuv420p {CFG['test_video_input']}".split(),
+            f"{CFG.ffmpeg} -hide_banner -f lavfi -i testsrc=duration=1:size={CFG.test_video_resolution}:rate={CFG.test_video_fps} -c:v libx264 -y -pix_fmt yuv420p {CFG.test_video_input}".split(),
             stdout=subprocess.PIPE,
             stderr=subprocess.PIPE,
             text=True,
@@ -469,8 +539,8 @@ def test():
             logging.debug(ret.stderr)
             ret.check_returncode()
         cmd = get_cmd(
-            CFG["test_video_input"],
-            CFG["test_video_output"],
+            CFG.test_video_input,
+            CFG.test_video_output,
         )
         ret = subprocess.run(
             cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True
@@ -512,30 +582,6 @@ def main(_root=None):
     logging.info("-------------------------------")
     logging.info(datetime.now().strftime("Video Compress started at %Y/%m/%d %H:%M"))
 
-    if CFG_FILE.exists():
-        try:
-            import json
-
-            cfg: dict = json.loads(CFG_FILE.read_text())
-            cfg_model = Config(**cfg)
-            CFG.update(cfg_model.model_dump())
-            get_frame.ffprobe = CFG["ffprobe"]
-            logging.debug(cfg_model)
-            logging.debug(CFG)
-        except KeyboardInterrupt as e:
-            raise e
-        except Exception as e:
-            logging.warning("Invalid config file, ignored.")
-            logging.debug(e)
-    else:
-        try:
-            import json
-
-            CFG_FILE.write_text(json.dumps(CFG, indent=4))
-            logging.info("Config file created.")
-        except Exception as e:
-            logging.warning("Failed to create config file.", exc_info=e)
-
     if _root is not None:
         root = Path(_root)
     else:
@@ -546,7 +592,7 @@ def main(_root=None):
             sys.exit(1)
         root = Path(sys.argv[1])
 
-    if root.name.lower() == CFG["compress_dir_name"].lower():
+    if root.name.lower() == CFG.compress_dir_name.lower():
         logging.critical("请修改目标目录名为非compress。")
         logging.error("Error termination via invalid input.")
         sys.exit(1)
@@ -576,5 +622,5 @@ def main(_root=None):
 
 
 if __name__ == "__main__":
-    sys.argv.append(r'C:\Users\flt\Documents\WeChat Files\wxid_m8h0igh8p52p22\FileStorage\Video')
+    # sys.argv.append(r'C:\Users\flt\Documents\WeChat Files\wxid_m8h0igh8p52p22\FileStorage\Video')
     main()

calc_utils/README.md

@@ -0,0 +1,73 @@
+# calc-utils
+
+个人使用的计算化学工具集，主要基于 [ASE (Atomic Simulation Environment)](https://wiki.fysik.dtu.dk/ase/) 和 [RDKit](https://www.rdkit.org/)。
+
+包含了一些方便的转换工具，以及针对特定服务器环境（PBS/Slurm/Custom）定制的 `ase.calculators.gaussian` 补丁。
+
+专用软件，`futils.gaussian`在不同服务器环境中无法直接运行，必须予以修改。
+
+## 安装
+
+需要 Python 3.12+。
+
+```bash
+git clone https://github.com/your-repo/calc-utils.git
+cd calc-utils
+pip install .
+```
+
+## 功能模块
+
+### 1. `futils.gaussian` (**Breaking Change**)
+
+这是一个对 `ase.calculators.gaussian` 的深度定制和 Monkey Patch。
+
+**注意：导入此模块会直接修改 `ase.calculators.gaussian` 中的类定义。**
+
+主要修改内容：
+- **强制任务提交脚本**：计算器的 `command` 默认被设置为调用 `gsub_wait` 脚本。
+  - 默认路径硬编码为 `/home/fanhj/calcs/lele/tools/gsub_wait`（需要在 `futils/gaussian.py` 中按需修改 `GSUB` 变量）。
+- **文件后缀变更**：输入文件使用 `.gin` 而非 `.gjf`，输出文件默认读取 `.out`。
+- **参数增强**：`__init__` 方法提供了更详细的 Type Hinting 和默认参数（如 `mem="30GB"`, `proc=32`）。
+- **辅助方法**：增加了 `mod()` 方法用于快速复制并修改计算器参数。
+
+```python
+from futils.gaussian import Gaussian
+from ase import Atoms
+
+# 使用定制后的 Gaussian 计算器
+# 注意：这会尝试调用 gsub_wait 提交任务
+calc = Gaussian(label='test_calc', method='B3LYP', basis='6-31G(d)')
+```
+
+### 2. `futils.rdkit2ase`
+
+提供 RDKit 分子对象 (`rdkit.Chem.Mol`) 与 ASE 原子对象 (`ase.Atoms`) 之间的无缝转换，**保留 3D 坐标**。
+
+```python
+from futils.rdkit2ase import MolToAtoms, AtomsToMol
+from rdkit import Chem
+
+# RDKit -> ASE
+mol = Chem.MolFromMolFile("molecule.mol")
+atoms = MolToAtoms(mol) 
+
+# ASE -> RDKit
+new_mol = AtomsToMol(atoms)
+```
+
+### 3. `futils.rdkit_utils`
+
+一些 RDKit 绘图辅助函数。
+- `draw2D(mol)`: 生成 SVG 格式的 2D 分子图。
+- `draw3D(mol)`: 使用 IPythonConsole 绘制 3D 分子图。
+
+## 脚本工具 (`bin/`)
+
+本项目包含了一些用于任务提交管理的 Shell 脚本，适用于特定的集群环境。
+
+- **`gsub`**: 任务提交脚本。支持本地或通过 SSH 远程提交到名为 `cluster` 的主机。
+- **`gsub_wait`**: 提交任务并阻塞等待完成，用于 ASE Calculator 的 `command` 调用，以便实现 Python 脚本的同步执行。
+
+**配置说明**：
+使用前请检查 `bin/` 下的脚本以及 `futils/gaussian.py` 中的 `GSUB` 路径，根据您的服务器环境进行调整。

calc_utils/bin/gsub

@@ -0,0 +1,176 @@
+#!/bin/bash
+set -u
+
+# Usage: gsub <jobname>
+
+job=${1:-}
+if [[ -z "$job" ]]; then
+  echo "Usage: $0 <jobname-without-extension>"
+  exit 1
+fi
+
+# ==========================================
+# 0. 安全检测函数 (Safety Check)
+# ==========================================
+check_dangerous_path() {
+  local path="${1:-}"
+  
+  # 1. Empty check
+  if [[ -z "$path" ]]; then
+    echo "Error: Empty path is dangerous for deletion." >&2
+    return 1
+  fi
+  
+  # 2. Root check
+  if [[ "$path" == "/" ]]; then
+     echo "Error: Root path '/' is dangerous for deletion." >&2
+     return 1
+  fi
+
+  # 3. Space check (optional, but good for safety)
+  if [[ "$path" =~ ^[[:space:]]+$ ]]; then
+     echo "Error: Whitespace path is dangerous." >&2
+     return 1
+  fi
+  
+  return 0
+}
+
+# ==========================================
+# 1. 检查运行环境 (Check Host)
+# ==========================================
+# 如果不是 cluster，尝试通过 SSH 远程调用
+host_short=$(hostname -s 2>/dev/null || hostname)
+if [[ "$host_short" != "cluster" ]]; then
+  # 假设本地挂载路径 /mnt/home 对应远程 /home (根据原脚本逻辑调整)
+  cur_dir=$(pwd)
+  remote_dir="${cur_dir//\/mnt\/home/\/home}"
+  
+  # 定位当前脚本并转换为远程路径
+  # 获取脚本所在目录的绝对路径
+  script_dir=$(cd "$(dirname "$0")" && pwd)
+  script_name=$(basename "$0")
+  local_script="$script_dir/$script_name"
+  
+  # 同样对脚本路径进行替换
+  remote_script="${local_script//\/mnt\/home/\/home}"
+  
+  # 尝试在远程执行自己
+  echo "Running remotely on cluster: $remote_script" >&2
+  ssh cluster "cd '$remote_dir' && '$remote_script' '$job'"
+  exit $?
+fi
+
+# ==========================================
+# 2. 准备作业 (Prepare Job)
+# ==========================================
+
+gin_file="$job.gin"
+if [[ ! -f "$gin_file" ]]; then
+  echo "Error: $gin_file not found in $(pwd)"
+  exit 2
+fi
+
+# 解析配置确定资源 (Parse Proc)
+# 查找 %NProcShared=XX
+proc=$(sed -n 's/^%NProcShared=\([0-9]\+\).*$/\1/pI' "$gin_file" | head -n 1)
+
+queue=""
+ppn=""
+
+if [[ "$proc" == "32" ]]; then
+  queue="n32"
+  ppn="32"
+elif [[ "$proc" == "20" ]]; then
+  queue="n20"
+  ppn="20"
+else
+  echo "Error: Unsupported NProcShared=$proc in $gin_file. Only 20 or 32 allowed."
+  exit 1
+fi
+
+# 清理旧文件 (Clean up old output)
+if [[ -f "$job.out" ]]; then
+  # 原脚本逻辑：休眠并删除
+  # echo "Warning: $job.out exists. Deleting..." >&2
+  # 使用安全检查
+  if check_dangerous_path "$job.out"; then
+    rm "$job.out"
+  else
+    echo "Skipping deletion of unsafe path: $job.out" >&2
+    exit 1
+  fi
+fi
+
+# ==========================================
+# 3. 生成作业脚本 (.job)
+# ==========================================
+job_file="$job.job"
+
+# 使用 heredoc 动态生成 PBS 脚本
+# 整合了原 g16_32.pbs 的内容和 gsub32 的追加内容
+cat > "$job_file" <<EOF
+#!/bin/sh
+#PBS -l nodes=1:ppn=$ppn
+#PBS -q $queue
+#PBS -j oe
+#PBS -N $job
+
+cd \$PBS_O_WORKDIR
+
+# Define Safety Check Function in Job Script
+check_rm_path() {
+  p="\$1"
+  # Empty check
+  if [ -z "\$p" ]; then
+    echo "Refusing to delete empty path"
+    return 1
+  fi
+  # Root check
+  if [ "\$p" = "/" ]; then
+    echo "Refusing to delete root path"
+    return 1
+  fi
+  return 0
+}
+
+export g16root=/share/apps/soft
+source \$g16root/g16/bsd/g16.profile
+
+# Create Scratch Directory
+if [ -n "\$USER" ] && [ -n "\$PBS_JOBID" ]; then
+  mkdir -p /scr/\$USER/\$PBS_JOBID
+  export GAUSS_SCRDIR=/scr/\$USER/\$PBS_JOBID
+else
+  echo "Error: USER or PBS_JOBID not set. Cannot setup scratch."
+  exit 1
+fi
+
+NODES=\`cat \$PBS_NODEFILE | uniq\`
+echo "--------------------------------------------------------"
+echo "  JOBID: \$PBS_JOBID"
+echo "  The job was started at \`date\`"
+echo "  The job was running at \$NODES."
+echo "--------------------------------------------------------"
+
+# Run G16 Job
+echo "Executing: g16 < $gin_file > $job.out"
+g16 < $gin_file > $job.out
+
+echo "--------------------------------------------------------"
+echo "  The job was finished at \`date\`"
+echo "--------------------------------------------------------"
+
+# Delete the tmp File (Cleanup Scratch)
+echo "Cleaning up \$GAUSS_SCRDIR"
+if check_rm_path "\$GAUSS_SCRDIR"; then
+  rm -rf "\$GAUSS_SCRDIR"
+fi
+
+EOF
+
+# ==========================================
+# 4. 提交作业 (Submit)
+# ==========================================
+# qsub 会输出 Job ID，例如 12345.cluster
+qsub "$job_file"

calc_utils/bin/gsub_wait

@@ -0,0 +1,116 @@
+#!/bin/bash
+set -u
+
+# Usage: gsub_wait <jobname>
+
+job=${1:-}
+if [[ -z "$job" ]]; then
+  echo "Usage: $0 <jobname-without-extension>"
+  exit 1
+fi
+
+# ==========================================
+# 1. 提交任务 (Submit Job)
+# ==========================================
+
+# 确定 gsub 命令位置
+# 优先查找当前目录下的 gsub，否则查找 PATH
+if [[ -x "./gsub" ]]; then
+    GSUB_CMD="./gsub"
+else
+    GSUB_CMD="gsub"
+fi
+
+# 调用 gsub 并捕获输出
+# 注意：gsub 内部可能通过 SSH 在远程执行，最终返回 qsub 的输出
+output=$($GSUB_CMD "$job")
+echo "$output"
+
+# ==========================================
+# 2. 检查是否需要等待 (Check Silent Mode)
+# ==========================================
+# 如果 GSUB_SILENT 为 1，则不进行监控，直接退出
+if [[ "${GSUB_SILENT:-0}" == "1" ]]; then
+  exit 0
+fi
+
+# ==========================================
+# 3. 监控任务进度 (Monitor Progress)
+# ==========================================
+
+# 尝试提取 Job ID (例如: 67147.cluster -> 67147)
+jobid_full=$(echo "$output" | grep -oE '[0-9]+\.cluster|[0-9]+' | head -n 1 || true)
+
+if [[ -n "$jobid_full" ]]; then
+  jobid=${jobid_full%%.*}
+  
+  # 准备参数
+  out_file="$job.out"
+  gin_file="$job.gin"
+  end_file="$job.job.o$jobid"
+  
+  if [[ ! -f "$gin_file" ]]; then
+      # 如果 gin 文件找不到（可能是远程路径问题？），跳过监控
+      echo "Warning: $gin_file not found nearby. Skipping monitor."
+      exit 0
+  fi
+  
+  # 计算 Total Steps: (--link1-- 数量) + 1
+  link_count=$(grep -c -- "--link1--" "$gin_file" || true)
+  total=$((link_count + 1))
+  cntDone=0
+  cntSCF=0
+  
+  last_lines=0
+  
+  echo "Monitoring Job $jobid..."
+
+  while true; do
+    # A. 检查 PBS 结束文件 (Job 完成标志)
+    if [[ -f "$end_file" ]]; then
+      echo "Job finished (found $end_file)."
+      break
+    fi
+    
+    # B. 检查并读取 .out 输出文件
+    if [[ -f "$out_file" ]]; then
+      curr_lines=$(wc -l < "$out_file" 2>/dev/null || echo 0)
+      
+      # 如果文件变小（被截断或重新生成），重置读取位置
+      if (( curr_lines < last_lines )); then last_lines=0; fi
+      
+      if (( curr_lines > last_lines )); then
+        # 逐行处理新增内容
+        # 使用进程替换 < <(...) 避免管道导致的子shell变量丢失问题
+        while IFS= read -r line; do
+          
+          # 检查 SCF Done
+          # 正则匹配: SCF Done: ... E ... = (数值) A.U.
+          if [[ "$line" =~ SCF[[:space:]]Done:.*E.*=[[:space:]]*([-0-9.]+)[[:space:]]*A\.U\. ]]; then
+             energy="${BASH_REMATCH[1]}"
+             cntSCF=$((cntSCF + 1))
+             echo "$job: SCF Done: $energy [$cntSCF] ($cntDone/$total)"
+          fi
+          
+          # 检查 Termination
+          if [[ "$line" == *"termination of Gaussian"* ]]; then
+             cntDone=$((cntDone + 1))
+             echo "$job: task done ($cntDone/$total)"
+          fi
+          
+        done < <(tail -n "+$((last_lines + 1))" "$out_file")
+        
+        last_lines=$curr_lines
+      fi
+    fi
+    
+    sleep 2
+  done
+  
+  # C. 最终校验
+  if (( cntDone != total )); then
+    echo "Warning: cntDone ($cntDone) != total ($total)"
+  fi
+else
+  echo "Could not parse Job ID from output. Monitor skipped."
+fi