github/NarratoAI
1
0
Fork 0
mirror of https://github.com/linyqh/NarratoAI.git synced 2026-02-03 11:48:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #58 from linyqh/dev

Browse Source 
v0.3.6
This commit is contained in:
linyq 2024-11-18 12:21:50 +08:00 committed by GitHub
commit 782c163bde
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 185 additions and 82 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

209
webui/components/basic_settings.py
View File

@ -66,6 +66,51 @@ def render_proxy_settings(tr):
os.environ["HTTPS_PROXY"] = HTTPS_PROXY os.environ["HTTPS_PROXY"] = HTTPS_PROXY
def test_vision_model_connection(api_key, base_url, model_name, provider, tr):
"""测试视觉模型连接
Args:
api_key: API密钥
base_url: 基础URL
model_name: 模型名称
provider: 提供商名称
Returns:
bool: 连接是否成功
str: 测试结果消息
"""
if provider.lower() == 'gemini':
import google.generativeai as genai
try:
genai.configure(api_key=api_key)
model = genai.GenerativeModel(model_name)
model.generate_content("直接回复我文本'当前网络可用'")
return True, tr("gemini model is available")
except Exception as e:
return False, f"{tr('gemini model is not available')}: {str(e)}"
elif provider.lower() == 'narratoapi':
import requests
try:
# 构建测试请求
headers = {
"Authorization": f"Bearer {api_key}"
}
test_url = f"{base_url.rstrip('/')}/health"
response = requests.get(test_url, headers=headers, timeout=10)
if response.status_code == 200:
return True, tr("NarratoAPI is available")
else:
return False, f"{tr('NarratoAPI is not available')}: HTTP {response.status_code}"
except Exception as e:
return False, f"{tr('NarratoAPI is not available')}: {str(e)}"
else:
return False, f"{tr('Unsupported provider')}: {provider}"
def render_vision_llm_settings(tr): def render_vision_llm_settings(tr):
"""渲染视频分析模型设置""" """渲染视频分析模型设置"""
st.subheader(tr("Vision Model Settings")) st.subheader(tr("Vision Model Settings"))
@ -99,6 +144,22 @@ def render_vision_llm_settings(tr):
st_vision_base_url = st.text_input(tr("Vision Base URL"), value=vision_base_url) st_vision_base_url = st.text_input(tr("Vision Base URL"), value=vision_base_url)
st_vision_model_name = st.text_input(tr("Vision Model Name"), value=vision_model_name) st_vision_model_name = st.text_input(tr("Vision Model Name"), value=vision_model_name)
# 在配置输入框后添加测试按钮
if st.button(tr("Test Connection"), key="test_vision_connection"):
with st.spinner(tr("Testing connection...")):
success, message = test_vision_model_connection(
api_key=st_vision_api_key,
base_url=st_vision_base_url,
model_name=st_vision_model_name,
provider=vision_provider,
tr=tr
)
if success:
st.success(tr(message))
else:
st.error(tr(message))
# 保存视觉模型配置 # 保存视觉模型配置
if st_vision_api_key: if st_vision_api_key:
config.app[f"vision_{vision_provider}_api_key"] = st_vision_api_key config.app[f"vision_{vision_provider}_api_key"] = st_vision_api_key
@ -110,80 +171,80 @@ def render_vision_llm_settings(tr):
config.app[f"vision_{vision_provider}_model_name"] = st_vision_model_name config.app[f"vision_{vision_provider}_model_name"] = st_vision_model_name
st.session_state[f"vision_{vision_provider}_model_name"] = st_vision_model_name st.session_state[f"vision_{vision_provider}_model_name"] = st_vision_model_name
# NarratoAPI 特殊配置 # # NarratoAPI 特殊配置
if vision_provider == 'narratoapi': # if vision_provider == 'narratoapi':
st.subheader(tr("Narrato Additional Settings")) # st.subheader(tr("Narrato Additional Settings"))
#
# Narrato API 基础配置 # # Narrato API 基础配置
narrato_api_key = st.text_input( # narrato_api_key = st.text_input(
tr("Narrato API Key"), # tr("Narrato API Key"),
value=config.app.get("narrato_api_key", ""), # value=config.app.get("narrato_api_key", ""),
type="password", # type="password",
help="用于访问 Narrato API 的密钥" # help="用于访问 Narrato API 的密钥"
) # )
if narrato_api_key: # if narrato_api_key:
config.app["narrato_api_key"] = narrato_api_key # config.app["narrato_api_key"] = narrato_api_key
st.session_state['narrato_api_key'] = narrato_api_key # st.session_state['narrato_api_key'] = narrato_api_key
#
narrato_api_url = st.text_input( # narrato_api_url = st.text_input(
tr("Narrato API URL"), # tr("Narrato API URL"),
value=config.app.get("narrato_api_url", "http://127.0.0.1:8000/api/v1/video/analyze") # value=config.app.get("narrato_api_url", "http://127.0.0.1:8000/api/v1/video/analyze")
) # )
if narrato_api_url: # if narrato_api_url:
config.app["narrato_api_url"] = narrato_api_url # config.app["narrato_api_url"] = narrato_api_url
st.session_state['narrato_api_url'] = narrato_api_url # st.session_state['narrato_api_url'] = narrato_api_url
#
# 视频分析模型配置 # # 视频分析模型配置
st.markdown("##### " + tr("Vision Model Settings")) # st.markdown("##### " + tr("Vision Model Settings"))
narrato_vision_model = st.text_input( # narrato_vision_model = st.text_input(
tr("Vision Model Name"), # tr("Vision Model Name"),
value=config.app.get("narrato_vision_model", "gemini-1.5-flash") # value=config.app.get("narrato_vision_model", "gemini-1.5-flash")
) # )
narrato_vision_key = st.text_input( # narrato_vision_key = st.text_input(
tr("Vision Model API Key"), # tr("Vision Model API Key"),
value=config.app.get("narrato_vision_key", ""), # value=config.app.get("narrato_vision_key", ""),
type="password", # type="password",
help="用于视频分析的模型 API Key" # help="用于视频分析的模型 API Key"
) # )
#
if narrato_vision_model: # if narrato_vision_model:
config.app["narrato_vision_model"] = narrato_vision_model # config.app["narrato_vision_model"] = narrato_vision_model
st.session_state['narrato_vision_model'] = narrato_vision_model # st.session_state['narrato_vision_model'] = narrato_vision_model
if narrato_vision_key: # if narrato_vision_key:
config.app["narrato_vision_key"] = narrato_vision_key # config.app["narrato_vision_key"] = narrato_vision_key
st.session_state['narrato_vision_key'] = narrato_vision_key # st.session_state['narrato_vision_key'] = narrato_vision_key
#
# 文案生成模型配置 # # 文案生成模型配置
st.markdown("##### " + tr("Text Generation Model Settings")) # st.markdown("##### " + tr("Text Generation Model Settings"))
narrato_llm_model = st.text_input( # narrato_llm_model = st.text_input(
tr("LLM Model Name"), # tr("LLM Model Name"),
value=config.app.get("narrato_llm_model", "qwen-plus") # value=config.app.get("narrato_llm_model", "qwen-plus")
) # )
narrato_llm_key = st.text_input( # narrato_llm_key = st.text_input(
tr("LLM Model API Key"), # tr("LLM Model API Key"),
value=config.app.get("narrato_llm_key", ""), # value=config.app.get("narrato_llm_key", ""),
type="password", # type="password",
help="用于文案生成的模型 API Key" # help="用于文案生成的模型 API Key"
) # )
#
if narrato_llm_model: # if narrato_llm_model:
config.app["narrato_llm_model"] = narrato_llm_model # config.app["narrato_llm_model"] = narrato_llm_model
st.session_state['narrato_llm_model'] = narrato_llm_model # st.session_state['narrato_llm_model'] = narrato_llm_model
if narrato_llm_key: # if narrato_llm_key:
config.app["narrato_llm_key"] = narrato_llm_key # config.app["narrato_llm_key"] = narrato_llm_key
st.session_state['narrato_llm_key'] = narrato_llm_key # st.session_state['narrato_llm_key'] = narrato_llm_key
#
# 批处理配置 # # 批处理配置
narrato_batch_size = st.number_input( # narrato_batch_size = st.number_input(
tr("Batch Size"), # tr("Batch Size"),
min_value=1, # min_value=1,
max_value=50, # max_value=50,
value=config.app.get("narrato_batch_size", 10), # value=config.app.get("narrato_batch_size", 10),
help="每批处理的图片数量" # help="每批处理的图片数量"
) # )
if narrato_batch_size: # if narrato_batch_size:
config.app["narrato_batch_size"] = narrato_batch_size # config.app["narrato_batch_size"] = narrato_batch_size
st.session_state['narrato_batch_size'] = narrato_batch_size # st.session_state['narrato_batch_size'] = narrato_batch_size
def render_text_llm_settings(tr): def render_text_llm_settings(tr):

45
webui/components/script_settings.py
View File

@ -205,6 +205,40 @@ def render_video_details(tr):
def render_script_buttons(tr, params): def render_script_buttons(tr, params):
"""渲染脚本操作按钮""" """渲染脚本操作按钮"""
# 新增三个输入框，放在同一行
input_cols = st.columns(3)
with input_cols[0]:
skip_seconds = st.number_input(
"skip_seconds",
min_value=0,
value=st.session_state.get('skip_seconds', config.frames.get('skip_seconds', 0)),
help=tr("Skip the first few seconds"),
key="skip_seconds_input"
)
st.session_state['skip_seconds'] = skip_seconds
with input_cols[1]:
threshold = st.number_input(
"threshold",
min_value=0,
value=st.session_state.get('threshold', config.frames.get('threshold', 30)),
help=tr("Difference threshold"),
key="threshold_input"
)
st.session_state['threshold'] = threshold
with input_cols[2]:
vision_batch_size = st.number_input(
"vision_batch_size",
min_value=1,
max_value=20,
value=st.session_state.get('vision_batch_size', config.frames.get('vision_batch_size', 5)),
help=tr("Vision processing batch size"),
key="vision_batch_size_input"
)
st.session_state['vision_batch_size'] = vision_batch_size
# 生成/加载按钮 # 生成/加载按钮
script_path = st.session_state.get('video_clip_json_path', '') script_path = st.session_state.get('video_clip_json_path', '')
if script_path == "auto": if script_path == "auto":
@ -287,7 +321,6 @@ def generate_script(tr, params):
with st.spinner("正在生成脚本..."): with st.spinner("正在生成脚本..."):
if not params.video_origin_path: if not params.video_origin_path:
st.error("请先选择视频文件") st.error("请先选择视频文件")
st.stop()
return return
# ===================提取键帧=================== # ===================提取键帧===================
@ -323,8 +356,8 @@ def generate_script(tr, params):
# 处理视频并提取关键帧 # 处理视频并提取关键帧
processor.process_video_pipeline( processor.process_video_pipeline(
output_dir=video_keyframes_dir, output_dir=video_keyframes_dir,
skip_seconds=config.frames.get("skip_seconds", 0), skip_seconds=st.session_state.get('skip_seconds'),
threshold=config.frames.get("threshold", 30) threshold=st.session_state.get('threshold')
) )
else: else:
processor = video_processor.VideoProcessor(params.video_origin_path) processor = video_processor.VideoProcessor(params.video_origin_path)
@ -353,7 +386,7 @@ def generate_script(tr, params):
except Exception as cleanup_err: except Exception as cleanup_err:
logger.error(f"清理失败的关键帧目录时出错: {cleanup_err}") logger.error(f"清理失败的关键帧目录时出错: {cleanup_err}")
raise Exception(f"关键帧提取<EFBFBD><EFBFBD>败: {str(e)}") raise Exception(f"关键帧提取失败: {str(e)}")
# 根据不同的 LLM 提供商处理 # 根据不同的 LLM 提供商处理
vision_llm_provider = st.session_state.get('vision_llm_providers').lower() vision_llm_provider = st.session_state.get('vision_llm_providers').lower()
@ -374,7 +407,7 @@ def generate_script(tr, params):
analyzer = vision_analyzer.VisionAnalyzer( analyzer = vision_analyzer.VisionAnalyzer(
model_name=vision_model, model_name=vision_model,
api_key=vision_api_key api_key=vision_api_key,
) )
update_progress(40, "正在分析关键帧...") update_progress(40, "正在分析关键帧...")
@ -388,7 +421,7 @@ def generate_script(tr, params):
analyzer.analyze_images( analyzer.analyze_images(
images=keyframe_files, images=keyframe_files,
prompt=config.app.get('vision_analysis_prompt'), prompt=config.app.get('vision_analysis_prompt'),
batch_size=config.frames.get("vision_batch_size", 5) batch_size=config.frames.get("vision_batch_size", st.session_state.get('vision_batch_size', 5))
) )
) )
loop.close() loop.close()

13
webui/i18n/zh.json
View File

@ -125,6 +125,15 @@
"Text API Key": "文案生成 API 密钥", "Text API Key": "文案生成 API 密钥",
"Text Base URL": "文案生成接口地址", "Text Base URL": "文案生成接口地址",
"Text Model Name": "文案生成模型名称", "Text Model Name": "文案生成模型名称",
"Account ID": "账户 ID" "Account ID": "账户 ID",
"Skip the first few seconds": "跳过开头多少秒",
"Difference threshold": "差异阈值",
"Vision processing batch size": "视觉处理批次大小",
"Test Connection": "测试连接",
"gemini model is available": "Gemini 模型可用",
"gemini model is not available": "Gemini 模型不可用",
"NarratoAPI is available": "NarratoAPI 可用",
"NarratoAPI is not available": "NarratoAPI 不可用",
"Unsupported provider": "不支持的提供商"
} }
} }