github/NarratoAI
1
0
Fork 0
mirror of https://github.com/linyqh/NarratoAI.git synced 2026-05-01 14:18:19 +00:00
Code Issues Packages Projects Releases Wiki Activity
346 Commits 2 Branches 27 Tags
Commit Graph

346 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
viccy
be653c5748
Merge pull request #236 from linyqh/codex/refactor-documentary-frame-analysis-pipeline 
refactor(documentary): centralize frame analysis pipeline
v0.7.8 		2026-04-03 13:16:02 +08:00
linyq
d5c63cf4b4 chore: bump version to 0.7.8 2026-04-03 13:09:26 +08:00
linyq
e53156f4f2 fix(documentary): normalize streamlit progress values 2026-04-03 12:57:24 +08:00
linyq
abc9db22e5 Fix documentary narration parsing and explicit vision overrides 2026-04-03 12:04:09 +08:00
linyq
4e2560651f fix(documentary): restore narration repair and explicit vision overrides 2026-04-03 11:29:27 +08:00
linyq
a8b6a5bb6b fix(documentary): fail on malformed narration payload 2026-04-03 02:45:33 +08:00
linyq
d678bf62b1 fix(documentary): centralize final script generation in shared service 2026-04-03 02:38:54 +08:00
linyq
ac63fea953 refactor(documentary): route adapters through shared analysis service 2026-04-03 02:24:30 +08:00
linyq
df034d104b fix(documentary): keep frames when batch summary is missing 2026-04-03 02:09:02 +08:00
linyq
ad02059e5d fix(documentary): validate batch response contract before success 2026-04-03 02:04:21 +08:00
linyq
4d21c43b89 feat(documentary): preserve failed batches and add vision concurrency 2026-04-03 01:54:47 +08:00
linyq
8201911b82 fix(documentary): harden fast-path fallback and cache key prefix 2026-04-03 01:42:43 +08:00
linyq
3d76bff442 perf(documentary): add fast frame extraction and cache keys 2026-04-03 01:30:51 +08:00
linyq
40a48cc9ff fix(documentary): align batch result fields with prompt contract 2026-04-03 01:23:05 +08:00
linyq
c83841a2e0 chore(gitignore): restore minimal tests ignore exception 2026-04-03 01:18:39 +08:00
linyq
f9539eac8c fix(documentary): tighten prompt contract and config guards 2026-04-03 01:14:41 +08:00
linyq
1d148370c5 feat(documentary): add shared frame analysis contract 2026-04-03 00:55:19 +08:00
linyq
093c8aa329 test: ignore manual llm smoke scripts in pytest 2026-04-03 00:47:30 +08:00
linyq
1057bd215c chore: ignore local worktrees 2026-04-03 00:36:46 +08:00
linyq
8a68f44f54 chore: 出于安全考虑，已移除 LiteLLM 依赖，统一使用 OpenAI 兼容请求链路 v0.7.7 2026-03-28 00:35:40 +08:00
linyq
16dbbf3461 refactor(config): 重构配置系统以支持默认值和模型名称规范化 
- 新增 defaults.py 提供共享默认配置和模型名称处理工具
- 重构 config.py 使用默认值填充缺失配置
- 修改 openai_compatible_provider.py 简化模型名称处理逻辑
- 更新 WebUI 组件使用新的默认值系统
- 添加测试用例验证配置引导和模型名称处理
 2026-03-28 00:34:01 +08:00
linyq
3396644593 feat: 移除 LiteLLM 依赖并迁移至 OpenAI 兼容接口 
- 移除 LiteLLM 相关代码和依赖，改用原生 OpenAI 兼容接口
- 重构 LLM 服务提供商注册逻辑，仅支持 OpenAI 兼容接口
- 更新配置文件和文档，移除 LiteLLM 相关说明
- 添加新的测试用例验证 OpenAI 兼容接口集成
- 更新 WebUI 组件以适配新的 OpenAI 兼容接口
 2026-03-27 23:49:58 +08:00
linyq
a6f2e0d815 docs: add Speclip recommendation to README files. 2026-03-16 21:51:34 +08:00
linyq
7ca594a788 feat(voice): 添加代理支持和改进 edge-tts 字幕处理 
添加代理配置支持，包括代理开关和地址设置
重构 edge-tts 字幕处理逻辑，兼容新旧 API 版本
改进错误处理和日志记录，增加连接超时设置
更新 edge-tts 依赖至最新版本
 2026-03-10 21:46:03 +08:00
viccy
ca4bfd1b44
Update README.md 2026-03-02 16:29:46 +08:00
linyq
0706b00577 feat: 修复 windows 环境下 短剧混剪 报错 bug，添加字幕文本处理模块，优化字幕读取和规范化逻辑 2026-01-10 02:28:01 +08:00
linyq
08f682bb50 fix: 优化短剧混剪字幕上传逻辑，与短剧解说保持一致 2025-12-25 10:43:28 +08:00
linyq
26f0dfeab5 fix: 修复短剧混剪字幕在 windows 环境下加载失败 2025-12-25 01:16:00 +08:00
linyq
5e46ea2746 fix: 优化短剧混剪提示词 v0.7.6 2025-12-24 13:49:54 +08:00
linyq
d12dfbaf0f fix: 优化注释 2025-12-24 13:26:20 +08:00
linyq
5f73c57313 fix: 优化短剧混剪使用体验 2025-12-24 13:18:52 +08:00
linyq
5fdf0b6a32 fix: 修复短剧混剪 AttributeError: 'NoneType' object has no attribute 'lower' 报错 2025-12-24 11:00:52 +08:00
linyq
4f964ad98d fix: 修复开发调试代码残留。已在当前版本中修复。服务端日志已清空。已建议所有用户重置 Key 2025-12-12 12:20:32 +08:00
linyq
dfb96e9b0f 更新了示例配置文件，并移除了日文README (坚决拥护中国🇨🇳领土主权🔥) 2025-12-12 11:42:50 +08:00
linyq
97bb59220f fix: 移除未使用的 logger 导入 2025-12-12 11:42:12 +08:00
linyq
169daac94d fix: 移除未使用的 tkinter 导入 2025-12-12 11:42:12 +08:00
linyq
c0e3ff045a fix: 更新版本号至 0.7.5 2025-12-12 11:42:12 +08:00
linyq
7b9ef2f244 feat: 新增 IndexTTS2 零样本语音克隆引擎支持 
添加 IndexTTS2 TTS 引擎配置和实现,支持零样本语音克隆功能。包括配置保存加载、API 调用、参考音频上传、高级参数设置(温度、top_p、top_k、束搜索、重复惩罚等),并在 WebUI 中提供完整的配置界面和使用说明。
 2025-12-12 11:42:12 +08:00
linyq
854cfab460 feat: 显示字幕引擎支持提示 2025-12-12 11:42:12 +08:00
linyq
474ebe46e2 feat: 新增基础设置项并提供中文翻译 2025-12-12 11:42:12 +08:00
linyq
46042d17d6 fix: 优化标题样式 2025-12-12 11:42:12 +08:00
linyq
eb57d2a0fe feat: 更新 webui 界面以支持新功能 2025-12-12 11:42:12 +08:00
linyq
d5f089c9a7 feat: 优化LLM服务配置与迁移适配，并更新相关UI设置及中文翻译 2025-12-12 11:42:12 +08:00
linyq
77c0aa47f2 feat: 增强 LiteLLM 提供商配置并更新基本设置界面 2025-12-12 11:42:12 +08:00
linyq
efa02d83ca fix: 更新版本号 2025-12-12 11:42:12 +08:00
linyq
eca1fcbe67 fix: 修改荒野建造提示词为更加通用的提示词 2025-12-12 11:42:12 +08:00
linyq
d7b1b51a36 fix: 使用 litellm 管理模型供应商 2025-12-12 11:40:44 +08:00
harry
4423195313 Fix the problem that Tencent cloud tts sound setting is invalid in the generation video function 2025-12-12 11:38:06 +08:00
harry
4b0f7c3bb9 新增qwen3 tts服务 2025-12-12 11:38:06 +08:00
Joran Joran
bad4a95ced fix: 修复docker-compose.yml中对于resourse的只读权限。 2025-12-12 11:38:06 +08:00