Upload folder using huggingface_hub

app.py

@@ -44,61 +44,65 @@ def speak_fn(
         print(f"Too Long Text: {text}")
         if exceed_flag:
             text = "不要超过100字！"
-            return gr.update(value="./assets/audios/nomorethan100.wav"), get_character_html(text), False, gr.update(interactive=True)
+            audio_value = "./assets/audios/nomorethan100.wav"
         else:
             text = "这句太长了，憋坏我啦！"
-            return gr.update(value="./assets/audios/overlength.wav"), get_character_html(text), True, gr.update(interactive=True)
-    audio_list = []
-    if len(text) > 42:
-        print(f"Long Text: {text}")
-        para_list = re_matching.cut_para(text)
-        for p in para_list:
-            audio_list_sent = []
-            sent_list = re_matching.cut_sent(p)
-            for s in sent_list:
-                audio = infer(
-                    s,
-                    sdp_ratio=sdp_ratio,
-                    noise_scale=noise_scale,
-                    noise_scale_w=noise_scale_w,
-                    length_scale=length_scale,
-                    sid=speaker,
-                    language=language,
-                    hps=hps,
-                    net_g=net_g,
-                    device=device,
-                )
-                audio_list_sent.append(audio)
-                silence = np.zeros((int)(44100 * interval_between_sent))
-                audio_list_sent.append(silence)
-            if (interval_between_para - interval_between_sent) > 0:
-                silence = np.zeros((int)(44100 * (interval_between_para - interval_between_sent)))
-                audio_list_sent.append(silence)
-            audio16bit = gr.processing_utils.convert_to_16_bit_wav(np.concatenate(audio_list_sent))  # 对完整句子做音量归一
-            audio_list.append(audio16bit)
+            audio_value = "./assets/audios/overlength.wav"
+        exceed_flag = not exceed_flag
     else:
-        print(f"Short Text: {text}")
-        silence = np.zeros(hps.data.sampling_rate // 2, dtype=np.int16)
-        with torch.no_grad():
-            for piece in text.split("|"):
-                audio = infer(
-                    piece,
-                    sdp_ratio=sdp_ratio,
-                    noise_scale=noise_scale,
-                    noise_scale_w=noise_scale_w,
-                    length_scale=length_scale,
-                    sid=speaker,
-                    language=language,
-                    hps=hps,
-                    net_g=net_g,
-                    device=device,
-                )
-                audio16bit = gr.processing_utils.convert_to_16_bit_wav(audio)
+        audio_list = []
+        if len(text) > 42:
+            print(f"Long Text: {text}")
+            para_list = re_matching.cut_para(text)
+            for p in para_list:
+                audio_list_sent = []
+                sent_list = re_matching.cut_sent(p)
+                for s in sent_list:
+                    audio = infer(
+                        s,
+                        sdp_ratio=sdp_ratio,
+                        noise_scale=noise_scale,
+                        noise_scale_w=noise_scale_w,
+                        length_scale=length_scale,
+                        sid=speaker,
+                        language=language,
+                        hps=hps,
+                        net_g=net_g,
+                        device=device,
+                    )
+                    audio_list_sent.append(audio)
+                    silence = np.zeros((int)(44100 * interval_between_sent))
+                    audio_list_sent.append(silence)
+                if (interval_between_para - interval_between_sent) > 0:
+                    silence = np.zeros((int)(44100 * (interval_between_para - interval_between_sent)))
+                    audio_list_sent.append(silence)
+                audio16bit = gr.processing_utils.convert_to_16_bit_wav(np.concatenate(audio_list_sent))  # 对完整句子做音量归一
                 audio_list.append(audio16bit)
-                audio_list.append(silence)  # 将静音添加到列表中
-    
-    audio_concat = np.concatenate(audio_list)    
-    return (hps.data.sampling_rate, audio_concat), get_character_html(text), exceed_flag, gr.update(interactive=True)
+        else:
+            print(f"Short Text: {text}")
+            silence = np.zeros(hps.data.sampling_rate // 2, dtype=np.int16)
+            with torch.no_grad():
+                for piece in text.split("|"):
+                    audio = infer(
+                        piece,
+                        sdp_ratio=sdp_ratio,
+                        noise_scale=noise_scale,
+                        noise_scale_w=noise_scale_w,
+                        length_scale=length_scale,
+                        sid=speaker,
+                        language=language,
+                        hps=hps,
+                        net_g=net_g,
+                        device=device,
+                    )
+                    audio16bit = gr.processing_utils.convert_to_16_bit_wav(audio)
+                    audio_list.append(audio16bit)
+                    audio_list.append(silence)  # 将静音添加到列表中
+        
+        audio_concat = np.concatenate(audio_list)
+        audio_value = (hps.data.sampling_rate, audio_concat)
+        
+    return gr.update(value=audio_value, autoplay=True), get_character_html(text), exceed_flag, gr.update(interactive=True)
 
 
 def submit_lock_fn():    
@@ -112,7 +116,7 @@ def init_fn():
     index = random.randint(1,7)
     welcome_text = get_sentence("Welcome", index)
     
-    return gr.update(value=f"./assets/audios/Welcome{index}.wav"), get_character_html(welcome_text)
+    return gr.update(value=f"./assets/audios/Welcome{index}.wav", autoplay=False), get_character_html(welcome_text)
 
 def get_sentence(category, index=-1):
     if index == -1:
@@ -129,14 +133,14 @@ with gr.Blocks(css=customCSS) as demo:
     tmp_string = gr.Textbox(value="", visible=False)
     character_area = gr.HTML(get_character_html("你好呀！"), elem_id="character_area")
     with gr.Tab("Speak", elem_id="tab-speak"):
-        speak_input = gr.Textbox(lines=1, label="Talking Flower will say:", elem_classes="wonder-card", elem_id="input_text")
+        speak_input = gr.Textbox(lines=1, label="Talking Flower will say:", elem_classes="wonder-card input_text", elem_id="speak_input")
         speak_button = gr.Button("Speak!", elem_id="speak_button", elem_classes="main-button wonder-card")
-        example_category = gr.Examples(["夸夸你 | Praise", "游戏台词 | Scripts", "玩梗 | Meme"], fn=get_sentence, inputs=[tmp_string], outputs=[speak_input], run_on_click=True, elem_id="examples")
+        example_category = gr.Examples(["夸夸你 | Praise", "游戏台词 | Scripts", "玩梗 | Meme"], inputs=[tmp_string], elem_id="examples")
     with gr.Tab("Chat", elem_id="tab-chat"):
-        chat_input = gr.Textbox(lines=1, placeholder="Coming Soon...", label="Chat to Talking Flower:", elem_classes="wonder-card", elem_id="input_text", interactive=False)
+        chat_input = gr.Textbox(lines=1, placeholder="Coming Soon...", label="Chat to Talking Flower:", elem_classes="wonder-card input_text", elem_id="chat_input", interactive=False)
         chat_button = gr.Button("Chat!", elem_id="chat_button", elem_classes="main-button wonder-card")
     with gr.Tab("Mimic", elem_id="tab-mimic"):
-        gr.Textbox(lines=1, placeholder="Coming Soon...", label="Choose sound to mimic:", elem_classes="wonder-card", elem_id="input_text", interactive=False)
+        gr.Textbox(lines=1, placeholder="Coming Soon...", label="Choose sound to mimic:", elem_classes="wonder-card input_text", elem_id="mimic_input", interactive=False)
         mimic_button = gr.Button("Mimic!", elem_id="mimic_button", elem_classes="main-button wonder-card")
     audio_output = gr.Audio(label="输出音频", show_label=False, autoplay=True, elem_id="audio_output", elem_classes="wonder-card")
     
@@ -163,7 +167,7 @@ if __name__ == "__main__":
     net_g = get_net_g(model_path=config.webui_config.model, version=version, device=device, hps=hps)
     reload_javascript()
     demo.launch(
-        allowed_paths=["./assets"],
+        allowed_paths=["./assets", "./javascript", "./css"],
         show_api=False,
         inbrowser=True,
     )

css/style.css

@@ -194,31 +194,31 @@ gradio-app {
     transition-property: transform,box-shadow !important;
 }
 
-#input_text label textarea {
-    width: 80%;
+.input_text label textarea {
+    width: 80% !important;
 }
-#input_text div.wrap {
-    display: none;
+.input_text div.wrap {
+    display: none !important;
 }
-#input_text {
-    display: flex;
+.input_text {
+    display: flex !important;
     background: #ffffff !important;
     height: 12rem !important;
 
 }
-#input_text label {
+.input_text label {
     margin-top: -2.25rem !important;
-    display: flex;
-    width: 100%;
-    align-content: center;
-    flex-direction: column;
-    flex-wrap: wrap;
-    justify-content: center;
-    align-items: center;
-}
-#input_text label span {
+    display: flex !important;
+    width: 100% !important;
+    align-content: center !important;
+    flex-direction: column !important;
+    flex-wrap: wrap !important;
+    justify-content: center !important;
+    align-items: center !important;
+}
+.input_text label span {
     font-family: var(--font-heading) !important;
-    font-size: 1.225rem;
+    font-size: 1.225rem !important;
 }
 
 /* Main Button */

javascript/main.js

@@ -0,0 +1,99 @@
+var key_down_history = [];
+var currentIndex = -1;
+
+var gradioContainer = null;
+
+var isInIframe = (window.self !== window.top);
+var currentTime = new Date().getTime();
+
+let windowWidth = window.innerWidth;
+let lines_json = []
+let lines_praise = []
+
+function addInit() {
+    return true;
+}
+
+function gradioApp() {
+    const elems = document.getElementsByTagName('gradio-app');
+    const elem = elems.length == 0 ? document : elems[0];
+
+    if (elem !== document) {
+        elem.getElementById = function(id) {
+            return document.getElementById(id);
+        };
+    }
+    return elem.shadowRoot ? elem.shadowRoot : elem;
+}
+
+function initialize() {
+    gradioObserver.observe(gradioApp(), { childList: true, subtree: true });
+
+    gradioContainer = gradioApp().querySelector(".gradio-container");
+
+    fetch('file=assets/lines.json')
+        .then(response => {
+            if (!response.ok) {
+                throw new Error('Network response was not ok');
+            }
+            return response.json();
+        })
+        .then(data => {
+            lines_json = data;
+            set_speak_examples();
+        })
+        .catch(error => {
+            console.error('There has been a problem with your fetch operation:', error);
+        });
+
+    
+
+    return true;
+}
+
+function set_speak_examples() {
+    buttons = gradioApp().querySelectorAll("#examples div button");
+    speak_input = gradioApp().querySelector("#speak_input label textarea");
+
+    let lines_praise = lines_json["夸夸你 | Praise"];
+    let lines_scripts = lines_json["游戏台词 | Scripts"];
+    let lines_meme = lines_json["玩梗 | Meme"];
+    let praiseArray = Object.values(lines_praise);
+    let scriptsArray = Object.values(lines_scripts);
+    let memeArray = Object.values(lines_meme);
+
+    buttons[0].addEventListener("click", function() {
+        const randomString = praiseArray[Math.floor(Math.random() * praiseArray.length)];
+        speak_input.value = randomString;
+    });
+    buttons[1].addEventListener("click", function() {
+        const randomString = scriptsArray[Math.floor(Math.random() * scriptsArray.length)];
+        speak_input.value = randomString;
+    });
+    buttons[2].addEventListener("click", function() {
+        const randomString = memeArray[Math.floor(Math.random() * memeArray.length)];
+        speak_input.value = randomString;
+    });
+
+}
+
+// 监视页面内部 DOM 变动
+var gradioObserver = new MutationObserver(function (mutations) {
+    for (var i = 0; i < mutations.length; i++) {
+        if (mutations[i].addedNodes.length) {
+            if (addInit()) {
+                gradioObserver.disconnect();
+                return;
+            }
+        }
+    }
+});
+
+// 监视页面变化
+window.addEventListener("DOMContentLoaded", function () {
+    windowWidth = window.innerWidth;
+    gradioApp().addEventListener("render", initialize);
+    isInIframe = (window.self !== window.top);
+});
+
+console.log("Welcome to TalkingFlower!");

tools/webui.py

@@ -1,8 +1,12 @@
+from collections import namedtuple
 import gradio as gr
 import os
 
+ScriptFile = namedtuple("ScriptFile", ["basedir", "filename", "path"])
+
 GradioTemplateResponseOriginal = gr.routes.templates.TemplateResponse
-root_path = os.path.dirname(os.path.realpath(__file__))
+current_dir = os.path.dirname(os.path.realpath(__file__))
+root_path = os.path.dirname(current_dir)
 
 def get_character_html(text):
     return f"""\
@@ -26,7 +30,7 @@ def list_scripts(scriptdirname, extension):
     scripts_dir = os.path.join(root_path, scriptdirname)
     if os.path.exists(scripts_dir):
         for filename in sorted(os.listdir(scripts_dir)):
-            scripts_list.append(ScriptFile(shared.assets_path, filename, os.path.join(scripts_dir, filename)))
+            scripts_list.append(ScriptFile(f"{root_path}/javascript", filename, os.path.join(scripts_dir, filename)))
     scripts_list = [x for x in scripts_list if os.path.splitext(x.path)[1].lower() == extension and os.path.isfile(x.path)]
     return scripts_list