添加批量混剪功能

.gitignore

@@ -9,3 +9,5 @@ hunjian_main.py
 .DS_Store
 */.DS_Store
 work/*
+final/*
+resource/*

final/README

@@ -0,0 +1 @@
+最后合成的视频文件会放在这里

locales/zh-CN.json

@@ -72,7 +72,23 @@
   "Video normalize...": "视频初始化...",
   "Generate Video...": "生成视频...",
   "Add Subtitles...": "添加字幕...",
-  "Generate Video completed!": "生成视频完成!"
+  "Generate Video completed!": "生成视频完成!",
+
+
+  "Mix Video": "视频混剪区",
+  "Mix Video Scene": "视频片段",
+  "Video Scene Resource": "视频片段资源目录",
+  "Please input video scene resource folder path": "请输入视频片段资源目录路径",
+  "Video Scene Text": "视频片段文案路径",
+  "Please input video scene text path": "请输入视频片段文案路径",
+  "Add More Scene": "添加更多片段",
+  "Delete Extra Scene": "删除多余片段",
+  "One Line Text For One Scene,UTF-8 encoding": "一个场景所需的文字放在TXT文件中的一行,程序会随机从TXT文件中挑选一行文本作为该场景的文字内容,TXT文件必须UTF-8编码",
+  "You Need More Resource": "视频资源不足,请添加资源后重试"
+
+
+
+
 
 
 }

main.py

@@ -2,11 +2,12 @@
 
 import streamlit as st
 
-from config.config import my_config
+from config.config import my_config, audio_voices_azure, audio_voices_ali, audio_voices_tencent
 from services.audio.alitts_service import AliAudioService
 from services.audio.azure_service import AzureAudioService
 from services.audio.tencent_tts_service import TencentAudioService
 from services.captioning.captioning_service import generate_caption, add_subtitles
+from services.hunjian.hunjian_service import concat_audio_list, get_audio_and_video_list
 from services.llm.azure_service import MyAzureService
 from services.llm.baichuan_service import MyBaichuanService
 from services.llm.baidu_qianfan_service import BaiduQianfanService
@@ -16,7 +17,7 @@
 from services.llm.tongyi_service import MyTongyiService
 from services.resource.pexels_service import PexelsService
 from services.resource.pixabay_service import PixabayService
-from services.video.video_service import get_audio_duration, VideoService
+from services.video.video_service import get_audio_duration, VideoService, VideoMixService
 from tools.tr_utils import tr
 from tools.utils import random_with_system_time, get_must_session_option, extent_audio
 
@@ -32,6 +33,16 @@
 audio_output_dir = os.path.abspath(audio_output_dir)
 
 
+def get_audio_voices():
+    selected_audio_provider = my_config['audio']['provider']
+    if selected_audio_provider == 'Azure':
+        return audio_voices_azure
+    if selected_audio_provider == 'Ali':
+        return audio_voices_ali
+    if selected_audio_provider == 'Tencent':
+        return audio_voices_tencent
+
+
 def get_resource_provider():
     resource_provider = my_config['resource']['provider']
     print("resource_provider:", resource_provider)
@@ -71,6 +82,8 @@ def get_audio_service():
 def main_generate_video_content():
     print("main_generate_video_content begin")
     topic = get_must_session_option('video_subject', "请输入要生成的主题")
+    if topic is None:
+        return
     video_language = st.session_state.get('video_language')
     video_length = st.session_state.get('video_length')
 
@@ -97,6 +110,8 @@ def main_try_test_audio():
     else:
         video_content = "你好，我是程序那些事"
     audio_voice = get_must_session_option("audio_voice", "请先设置配音语音")
+    if audio_voice is None:
+        return
     audio_service.read_with_ssml(video_content,
                                  audio_voice,
                                  audio_rate)
@@ -111,7 +126,11 @@ def main_generate_video_dubbing():
     audio_rate = get_audio_rate()
 
     video_content = get_must_session_option("video_content", "请先设置视频主题")
+    if video_content is None:
+        return
     audio_voice = get_must_session_option("audio_voice", "请先设置配音语音")
+    if audio_voice is None:
+        return
     audio_service.save_with_ssml(video_content,
                                  audio_output_file,
                                  audio_voice,
@@ -121,6 +140,16 @@ def main_generate_video_dubbing():
     print("main_generate_video_dubbing end")
 
 
+def main_generate_video_dubbing_for_mix():
+    print("main_generate_video_dubbing_for_mix begin")
+    audio_service = get_audio_service()
+    audio_rate = get_audio_rate()
+    audio_output_file_list, video_dir_list = get_audio_and_video_list(audio_service, audio_rate)
+    st.session_state["audio_output_file_list"] = audio_output_file_list
+    st.session_state["video_dir_list"] = video_dir_list
+    print("main_generate_video_dubbing_for_mix end")
+
+
 def get_audio_rate():
     audio_provider = my_config['audio']['provider']
     if audio_provider == "Azure":
@@ -180,7 +209,11 @@ def main_get_video_resource():
     print("main_get_video_resource begin")
     resource_service = get_resource_provider()
     query = get_must_session_option("video_keyword", "请先设置视频关键字")
+    if query is None:
+        return
     audio_file = get_must_session_option("audio_output_file", "请先生成配音文件")
+    if audio_file is None:
+        return
     audio_length = get_audio_duration(audio_file)
     print("audio_length:", audio_length)
     return_videos, total_length = resource_service.handle_video_resource(query, audio_length, 50, False)
@@ -213,7 +246,11 @@ def main_generate_ai_video(video_generator):
             main_get_video_resource()
             st.write(tr("Video normalize..."))
             audio_file = get_must_session_option("audio_output_file", "请先生成配音文件")
+            if audio_file is None:
+                return
             video_list = get_must_session_option("return_videos", "请先生成视频资源文件")
+            if video_list is None:
+                return
 
             video_service = VideoService(video_list, audio_file)
             print("normalize video")
@@ -226,6 +263,72 @@ def main_generate_ai_video(video_generator):
             if enable_subtitles:
                 st.write(tr("Add Subtitles..."))
                 subtitle_file = get_must_session_option('captioning_output', "请先生成字幕文件")
+                if subtitle_file is None:
+                    return
+
+                font_name = st.session_state.get('subtitle_font')
+                font_size = st.session_state.get('subtitle_font_size')
+                primary_colour = st.session_state.get('subtitle_color')
+                outline_colour = st.session_state.get('subtitle_border_color')
+                outline = st.session_state.get('subtitle_border_width')
+                alignment = st.session_state.get('subtitle_position')
+                add_subtitles(video_file, subtitle_file,
+                              font_name=font_name,
+                              font_size=font_size,
+                              primary_colour=primary_colour,
+                              outline_colour=outline_colour,
+                              outline=outline,
+                              alignment=alignment)
+                print("final file with subtitle:", video_file)
+            st.session_state["result_video_file"] = video_file
+            status.update(label=tr("Generate Video completed!"), state="complete", expanded=False)
+
+
+def main_generate_ai_video_for_mix(video_generator):
+    print("main_generate_ai_video_for_fix begin:")
+    with video_generator:
+        st_area = st.status(tr("Generate Video in process..."), expanded=True)
+        with st_area as status:
+            st.write(tr("Generate Video Dubbing..."))
+            main_generate_video_dubbing_for_mix()
+            st.write(tr("Video normalize..."))
+            video_dir_list = get_must_session_option("video_dir_list", "请选择视频目录路径")
+            audio_file_list = get_must_session_option("audio_output_file_list", "请先生成配音文件列表")
+
+            video_mix_servie = VideoMixService()
+            # 使用 zip() 函数遍历两个列表并获得配对
+            i = 0
+            audio_output_file_list = []
+            final_video_file_list = []
+            for video_dir, audio_file in zip(video_dir_list, audio_file_list):
+                print(f"Video Directory: {video_dir}, Audio File: {audio_file}")
+                if i == 0:
+                    matching_videos, total_length = video_mix_servie.match_videos_from_dir(video_dir,
+                                                                                           audio_file, True)
+                else:
+                    matching_videos, total_length = video_mix_servie.match_videos_from_dir(video_dir,
+                                                                                           audio_file, False)
+                i = i + 1
+                audio_output_file_list.append(audio_file)
+                final_video_file_list.extend(matching_videos)
+
+            final_audio_output_file = concat_audio_list(audio_output_file_list)
+            st.session_state['audio_output_file'] = final_audio_output_file
+            st.write(tr("Generate Video subtitles..."))
+            main_generate_subtitle()
+            video_service = VideoService(final_video_file_list, final_audio_output_file)
+            print("normalize video")
+            video_service.normalize_video()
+            st.write(tr("Generate Video..."))
+            video_file = video_service.generate_video_with_audio()
+            print("final file without subtitle:", video_file)
+
+            enable_subtitles = st.session_state.get("enable_subtitles")
+            if enable_subtitles:
+                st.write(tr("Add Subtitles..."))
+                subtitle_file = get_must_session_option('captioning_output', "请先生成字幕文件")
+                if subtitle_file is None:
+                    return
 
                 font_name = st.session_state.get('subtitle_font')
                 font_size = st.session_state.get('subtitle_font_size')

pages/01_auto_video.py

@@ -3,7 +3,7 @@
 from config.config import my_config, save_config, languages, audio_languages, audio_voices_azure, transition_types, \
     fade_list, audio_voices_ali, audio_voices_tencent
 from main import main_generate_video_content, main_generate_ai_video, main_generate_video_dubbing, \
-    main_get_video_resource, main_generate_subtitle, main_try_test_audio
+    main_get_video_resource, main_generate_subtitle, main_try_test_audio, get_audio_voices
 from pages.common import common_ui
 from tools.tr_utils import tr
 
@@ -70,17 +70,6 @@ def generate_video(video_generator):
     main_generate_ai_video(video_generator)
 
 
-def get_audio_voices():
-    selected_audio_provider = my_config['audio']['provider']
-    if selected_audio_provider == 'Azure':
-        return audio_voices_azure
-    if selected_audio_provider == 'Ali':
-        return audio_voices_ali
-    if selected_audio_provider == 'Tencent':
-        return audio_voices_tencent
-
-test_mode = my_config["test_mode"]
-
 st.markdown("<h1 style='text-align: center; font-weight:bold; font-family:comic sans ms; padding-top: 0rem;'> \
             AI搞钱工具</h1>", unsafe_allow_html=True)
 st.markdown("<h2 style='text-align: center;padding-top: 0rem;'>自动短视频生成器</h2>", unsafe_allow_html=True)
@@ -215,8 +204,6 @@ def get_audio_voices():
     with llm_columns[3]:
         st.selectbox(label=tr("video Transition effect duration"), key="video_transition_effect_duration",
                      options=["1", "2"])
-    # if test_mode:
-    #     st.button(label=tr("Get Video Resource"), on_click=get_video_resource)
 
 # 字幕
 subtitle_container = st.container(border=True)
@@ -265,13 +252,10 @@ def get_audio_voices():
     with llm_columns[3]:
         st.slider(label=tr("subtitle border width"), min_value=0.0, value=0.0, max_value=4.0, step=1.0,
                   key="subtitle_border_width")
-    # if test_mode:
-    #     st.button(label=tr("Generate subtitle"), on_click=generate_subtitle)
 
 # 生成视频
 video_generator = st.container(border=True)
 with video_generator:
-    # st_status = st.status(tr("Generate Video in process..."), expanded=True)
     st.button(label=tr("Generate Video Button"), type="primary", on_click=generate_video, args=(video_generator,))
 result_video_file = st.session_state.get("result_video_file")
 if result_video_file:

resource/README

@@ -0,0 +1 @@
+资源文件放在这里

services/audio/tencent_recognition_service.py

@@ -61,7 +61,7 @@ def process(self, audioFile, language) -> List[TencentRecognitionResult]:
             if code != 0:
                 print("recognize failed! request_id: ", request_id, " code: ", code, ", message: ", resp["message"])
                 st.toast("腾讯云语音识别失败", icon="⚠️")
-                st.stop()
+                return None
 
             print("request_id: ", request_id)
             # 一个channl_result对应一个声道的识别结果

services/captioning/captioning_service.py

@@ -6,7 +6,7 @@
 from config.config import my_config
 from services.alinls.speech_process import AliRecognitionService
 from services.audio.tencent_recognition_service import TencentRecognitionService
-from services.captioning.azure_captioning_service import Captioning
+from services.captioning.common_captioning_service import Captioning
 import subprocess
 
 from tools.file_utils import generate_temp_filename
@@ -53,8 +53,9 @@ def generate_caption():
         print("selected_audio_provider: Tencent")
         tencent_service = TencentRecognitionService()
         result_list = tencent_service.process(get_session_option("audio_output_file"),
-                                              get_session_option("video_language"))
-        # print("result list:", result_list)
+                                              get_session_option("audio_language"))
+        if result_list is None:
+            return
         captioning._offline_results = result_list
     captioning.finish()
 

services/hunjian/hunjian_service.py

@@ -0,0 +1,89 @@
+import os
+import subprocess
+
+import streamlit as st
+
+from tools.file_utils import random_line_from_text_file
+from tools.utils import get_must_session_option, random_with_system_time, extent_audio, run_ffmpeg_command
+
+# 获取当前脚本的绝对路径
+script_path = os.path.abspath(__file__)
+
+# print("当前脚本的绝对路径是:", script_path)
+
+# 脚本所在的目录
+script_dir = os.path.dirname(script_path)
+# 音频输出目录
+audio_output_dir = os.path.join(script_dir, "../../work")
+audio_output_dir = os.path.abspath(audio_output_dir)
+
+
+def get_session_video_scene_text():
+    video_dir_list = []
+    video_text_list = []
+    for i in range(5):
+        if "video_scene_folder_" + str(i + 1) in st.session_state and st.session_state["video_scene_folder_" + str(i + 1)] is not None:
+            video_dir_list.append(st.session_state["video_scene_folder_" + str(i + 1)])
+            video_text_list.append(st.session_state["video_scene_text_" + str(i + 1)])
+    return video_dir_list, video_text_list
+
+
+def get_video_scene_text_list(video_text_list):
+    video_scene_text_list = []
+    for video_text in video_text_list:
+        if video_text is not None:
+            video_line = random_line_from_text_file(video_text)
+            video_scene_text_list.append(video_line)
+    return video_scene_text_list
+
+
+def get_video_text_from_list(video_scene_text_list):
+    return " ".join(video_scene_text_list)
+
+
+def get_audio_and_video_list(audio_service, audio_rate):
+    audio_output_file_list = []
+    video_dir_list, video_text_list = get_session_video_scene_text()
+    video_scene_text_list = get_video_scene_text_list(video_text_list)
+    audio_voice = get_must_session_option("audio_voice", "请先设置配音语音")
+    i = 0
+    for video_scene_text in video_scene_text_list:
+        temp_file_name = str(random_with_system_time()) + str(i)
+        i = i + 1
+        audio_output_file = os.path.join(audio_output_dir, str(temp_file_name) + ".wav")
+        audio_service.save_with_ssml(video_scene_text,
+                                     audio_output_file,
+                                     audio_voice,
+                                     audio_rate)
+        extent_audio(audio_output_file, 1)
+        audio_output_file_list.append(audio_output_file)
+    return audio_output_file_list, video_dir_list
+
+
+def get_video_text():
+    video_dir_list, video_text_list = get_session_video_scene_text()
+    video_scene_text_list = get_video_scene_text_list(video_text_list)
+    return get_video_text_from_list(video_scene_text_list)
+
+
+def concat_audio_list(audio_output_file_list):
+    temp_output_file_name = os.path.join(audio_output_dir, str(random_with_system_time()) + ".wav")
+    concat_audio_file = os.path.join(audio_output_dir, "concat_audio_file.txt")
+    with open(concat_audio_file, 'w', encoding='utf-8') as f:
+        for audio_file in audio_output_file_list:
+            f.write("file '{}'\n".format(os.path.abspath(audio_file)))
+    # 调用ffmpeg来合并音频
+    # 注意：这里假设ffmpeg在你的PATH中，否则你需要提供ffmpeg的完整路径
+    command = [
+        'ffmpeg',
+        '-f', 'concat',
+        '-safe', '0',
+        '-i', concat_audio_file,
+        '-c', 'copy',  # 如果可能，直接复制流而不是重新编码
+        temp_output_file_name
+    ]
+    run_ffmpeg_command(command)
+    # 完成后，删除临时文件（如果你不再需要它）
+    os.remove(concat_audio_file)
+    print(f"Audio files have been merged into {temp_output_file_name}")
+    return temp_output_file_name