首次提交

config/config.json

@@ -0,0 +1,113 @@
+{
+  "tencentcloud": {
+    "ocr": {
+      "secret_id": "",
+      "secret_key": "",
+      "max_attempts": 4,
+      "timeout": 30,
+      "dispatch_log_every": 1000,
+      "endpoints": [
+      ]
+    },
+    "cos": {
+      "secret_id": "",
+      "secret_key": "",
+      "region": "",
+      "bucket": ""
+    }
+  },
+  "audio": {
+    "model_size": "small",
+    "device": "cpu",
+    "compute_type": "float16",
+    "device_index": "auto",
+    "cpu_threads": 2
+  },
+  "splitframe": {
+    "interval_seconds": 0.33,
+    "ffmpeg_threads": 3,
+    "use_gpu": true,
+    "gpu_skip_codecs": [
+      "av1"
+    ],
+    "gpu_reserve_last_for_whisper": false,
+    "gpu_skip_indices": [],
+    "gpu_auto_select": true,
+    "gpu_busy_mem_mb": 4096,
+    "crop_bottom_ratio": null,
+    "download": true,
+    "root_path": "./video",
+    "upload_path": "",
+    "delete_video": true,
+    "coscli_path": "./coscli-linux-amd64",
+    "cosconfig_path": ".cos.yaml",
+    "ffmpeg_timeout": 3600
+  },
+  "subtitle": {
+    "min_confidence": 50,
+    "similarity_threshold": 0.7,
+    "ocr_workers": 20,
+    "require_chinese": true,
+    "redundant_bands": 12,
+    "redundant_merge_ratio": 0.97,
+    "no_composite": false,
+    "min_duration_seconds": 0,
+    "merge_workers": 4,
+    "cv_detect": true,
+    "cv_use_region": false,
+    "max_tiles_per_composite": 24,
+    "ocr_max_image_side": 4000,
+    "filter_static": true,
+    "static_frac": 0.25,
+    "filter_lowconf": true,
+    "lowconf_thr": 80,
+    "lowconf_filter_frac": 0.6,
+    "lowconf_min_det": 200,
+    "strip_static_prefix": false,
+    "filter_watermark_pos": true,
+    "wm_dominant_frac": 0.25,
+    "wm_min_count": 20,
+    "split_far_boxes": true,
+    "max_duration_seconds": 0,
+    "long_video_list": "./超长待重跑.txt",
+    "ocr_no_drop": true,
+    "ocr_max_passes": 5,
+    "ocr_retry_sleep": 2,
+    "ocr_fail_list": "./OCR失败待重跑.txt",
+    "max_global_ocr": 0,
+    "no_subtitle_reasons": [
+      "字幕太少",
+      "冗余文字过多",
+      "冗余文字太多",
+      "无字幕区域",
+      "非字幕内容",
+      "置信度太低",
+      "未检测",
+      "特殊字体",
+      "视频过短",
+      "视频过长"
+    ]
+  },
+  "dispatch": {
+    "enabled": true,
+    "source": "cos",
+    "run_long_video": false,
+    "auto_fallback_long_video": false,
+    "global_dedup_on_startup": false,
+    "machine_id": "",
+    "jobs_root": "",
+    "shard_size": 2000,
+    "keep_alive": true,
+    "idle_poll_seconds": 60,
+    "claim_shuffle": true,
+    "claim_match": "",
+    "heartbeat_seconds": 300,
+    "daily_backup_path": "",
+    "monitor_root": "",
+    "long_list_backup_path": "",
+    "long_list_clear_after_upload": true,
+    "long_list_min_batch": 1000,
+    "daily_backup_hour": 10,
+    "stop_flag_key": ""
+  }
+}

config/config_utils.py

@@ -0,0 +1,131 @@
+# config_utils.py
+import json
+import os
+
+class ConfigManager:
+    _instance = None
+    _config = None
+    
+    def __new__(cls):
+        if cls._instance is None:
+            cls._instance = super(ConfigManager, cls).__new__(cls)
+            cls._instance.load_config()
+        return cls._instance
+    
+    def load_config(self, config_path="./config/config.json"):
+        """加载配置文件"""
+        try:
+            if not os.path.exists(config_path):
+                # 如果配置文件不存在，创建一个默认配置
+                default_config = {
+                    "tencentcloud": {
+                        "ocr": {
+                            "secret_id": "please_set_your_secret_id",
+                            "secret_key": "please_set_your_secret_key",
+                            "region": "ap-beijing",
+                            "endpoint": "ocr.tencentcloudapi.com"
+                        }
+                    }
+                }
+                with open(config_path, 'w', encoding='utf-8') as f:
+                    json.dump(default_config, f, indent=4, ensure_ascii=False)
+                print(f"已创建默认配置文件: {config_path}")
+                print("请编辑此文件并填写您的腾讯云OCR密钥")
+            
+            with open(config_path, 'r', encoding='utf-8') as f:
+                self._config = json.load(f)
+                
+        except Exception as e:
+            print(f"加载配置文件失败: {e}")
+            self._config = {}
+    
+    def get(self, key, default=None):
+        """获取配置值"""
+        try:
+            keys = key.split('.')
+            value = self._config
+            for k in keys:
+                value = value[k]
+            return value
+        except (KeyError, TypeError):
+            return default
+    
+    def get_tencent_ocr_config(self):
+        """获取腾讯云OCR配置"""
+        return self.get('tencentcloud.ocr', {})
+    
+    def get_splitframe_config(self):
+        """获取Whisper配置"""
+        return self.get('splitframe', {})
+    
+    def get_audio_config(self):
+        """获取音频配置"""
+        return self.get('audio', {})
+    
+    def get_subtitle_config(self):
+        """获取字幕配置"""
+        return self.get('subtitle', {})
+
+    def get_cos_config(self):
+        """获取COS配置"""
+        return self.get('tencentcloud.cos', {})
+
+    def get_dispatch_config(self):
+        """获取 COS 任务分发配置（常驻 worker / 认领锁 / 每日备份）。"""
+        return self.get('dispatch', {})
+
+def _prepend_local_ffmpeg_to_path():
+    """若项目根的 bin/ 下有自带 ffmpeg，则把 bin/ 提到 PATH 最前，
+    让所有 ffmpeg/ffprobe 调用(抽帧、音频、pydub 等)都优先用项目自带的，
+    不再依赖系统 PATH。"""
+    root = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+    bin_dir = os.path.join(root, "bin")
+    ff = os.path.join(bin_dir, "ffmpeg")
+    if os.path.isfile(ff):
+        try:
+            os.chmod(ff, 0o755)
+            ffprobe = os.path.join(bin_dir, "ffprobe")
+            if os.path.isfile(ffprobe):
+                os.chmod(ffprobe, 0o755)
+        except OSError:
+            pass
+        os.environ["PATH"] = bin_dir + os.pathsep + os.environ.get("PATH", "")
+
+
+def _ensure_coscli_executable():
+    """给项目根的 coscli-linux-amd64 自动加执行权限（解压后常丢失 +x，导致无法调用）。"""
+    if os.name == "nt":
+        return
+    root = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+    for name in ("coscli-linux-amd64", "coscli"):
+        p = os.path.join(root, name)
+        if os.path.isfile(p) and not os.access(p, os.X_OK):
+            try:
+                os.chmod(p, 0o755)
+            except OSError:
+                pass
+
+
+_prepend_local_ffmpeg_to_path()
+_ensure_coscli_executable()
+
+
+def _resolve_bin(name):
+    """返回项目 bin/ 下的可执行文件绝对路径(若存在)，否则返回名字本身(回退到 PATH)。"""
+    root = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+    p = os.path.join(root, "bin", name)
+    return p if os.path.isfile(p) else name
+
+
+def get_ffmpeg():
+    """ffmpeg 可执行路径：优先项目自带 bin/ffmpeg，否则 PATH 里的 ffmpeg。"""
+    return _resolve_bin("ffmpeg")
+
+
+def get_ffprobe():
+    """ffprobe 可执行路径：优先项目自带 bin/ffprobe，否则 PATH 里的 ffprobe。"""
+    return _resolve_bin("ffprobe")
+
+
+# 全局配置实例
+config = ConfigManager()