AIFSH · peizhiluo007 · Oct 24, 2024 · Oct 24, 2024
diff --git a/.gitignore b/.gitignore
diff --git a/LICENSE b/LICENSE
diff --git a/README.md b/README.md
diff --git a/__init__.py b/__init__.py
@@ -1,10 +1,12 @@
-from .nodes import TextNode, CosyVoiceNode, LoadSRT, CosyVoiceDubbingNode
+from .nodes import TextNode,CosyVoiceNode,CosyVoiceDubbingNode,LoadSRT,PreviewAudio,LoadAudio
 
 WEB_DIRECTORY = "./web"
 
 NODE_CLASS_MAPPINGS = {
-    "LoadSRT":LoadSRT,
     "TextNode": TextNode,
     "CosyVoiceNode": CosyVoiceNode,
-    "CosyVoiceDubbingNode":CosyVoiceDubbingNode
+    "CosyVoiceDubbingNode":CosyVoiceDubbingNode,
+    "LoadSRT":LoadSRT,
+    "LoadAudio": LoadAudio,
+    "PreviewAudio": PreviewAudio,
 }
diff --git a/cosyvoice/__init__.py b/cosyvoice/__init__.py
diff --git a/cosyvoice/bin/export_jit.py b/cosyvoice/bin/export_jit.py
diff --git a/cosyvoice/bin/export_trt.py b/cosyvoice/bin/export_trt.py
diff --git a/cosyvoice/bin/inference.py b/cosyvoice/bin/inference.py
diff --git a/cosyvoice/bin/train.py b/cosyvoice/bin/train.py
diff --git a/cosyvoice/cli/__init__.py b/cosyvoice/cli/__init__.py
diff --git a/cosyvoice/cli/cosyvoice.py b/cosyvoice/cli/cosyvoice.py
diff --git a/cosyvoice/cli/frontend.py b/cosyvoice/cli/frontend.py
diff --git a/cosyvoice/cli/model.py b/cosyvoice/cli/model.py
diff --git a/cosyvoice/cli/zh_normalization/README.md b/cosyvoice/cli/zh_normalization/README.md
diff --git a/cosyvoice/cli/zh_normalization/__init__.py b/cosyvoice/cli/zh_normalization/__init__.py
diff --git a/cosyvoice/cli/zh_normalization/char_convert.py b/cosyvoice/cli/zh_normalization/char_convert.py
diff --git a/cosyvoice/cli/zh_normalization/chronology.py b/cosyvoice/cli/zh_normalization/chronology.py
diff --git a/cosyvoice/cli/zh_normalization/constants.py b/cosyvoice/cli/zh_normalization/constants.py
diff --git a/cosyvoice/cli/zh_normalization/num.py b/cosyvoice/cli/zh_normalization/num.py
diff --git a/cosyvoice/cli/zh_normalization/phonecode.py b/cosyvoice/cli/zh_normalization/phonecode.py
diff --git a/cosyvoice/cli/zh_normalization/quantifier.py b/cosyvoice/cli/zh_normalization/quantifier.py
diff --git a/cosyvoice/cli/zh_normalization/text_normlization.py b/cosyvoice/cli/zh_normalization/text_normlization.py
diff --git a/cosyvoice/dataset/__init__.py b/cosyvoice/dataset/__init__.py
diff --git a/cosyvoice/dataset/dataset.py b/cosyvoice/dataset/dataset.py
diff --git a/cosyvoice/dataset/processor.py b/cosyvoice/dataset/processor.py
diff --git a/cosyvoice/flow/decoder.py b/cosyvoice/flow/decoder.py
diff --git a/cosyvoice/flow/flow.py b/cosyvoice/flow/flow.py
diff --git a/cosyvoice/flow/flow_matching.py b/cosyvoice/flow/flow_matching.py
diff --git a/cosyvoice/flow/length_regulator.py b/cosyvoice/flow/length_regulator.py
diff --git a/cosyvoice/hifigan/f0_predictor.py b/cosyvoice/hifigan/f0_predictor.py
diff --git a/cosyvoice/hifigan/generator.py b/cosyvoice/hifigan/generator.py
diff --git a/cosyvoice/llm/llm.py b/cosyvoice/llm/llm.py
diff --git a/cosyvoice/transformer/__init__.py b/cosyvoice/transformer/__init__.py
diff --git a/cosyvoice/transformer/activation.py b/cosyvoice/transformer/activation.py
diff --git a/cosyvoice/transformer/attention.py b/cosyvoice/transformer/attention.py
diff --git a/cosyvoice/transformer/convolution.py b/cosyvoice/transformer/convolution.py
diff --git a/cosyvoice/transformer/decoder.py b/cosyvoice/transformer/decoder.py
diff --git a/cosyvoice/transformer/decoder_layer.py b/cosyvoice/transformer/decoder_layer.py
diff --git a/cosyvoice/transformer/embedding.py b/cosyvoice/transformer/embedding.py
diff --git a/cosyvoice/transformer/encoder.py b/cosyvoice/transformer/encoder.py
diff --git a/cosyvoice/transformer/encoder_layer.py b/cosyvoice/transformer/encoder_layer.py
diff --git a/cosyvoice/transformer/label_smoothing_loss.py b/cosyvoice/transformer/label_smoothing_loss.py
diff --git a/cosyvoice/transformer/positionwise_feed_forward.py b/cosyvoice/transformer/positionwise_feed_forward.py
diff --git a/cosyvoice/transformer/subsampling.py b/cosyvoice/transformer/subsampling.py
diff --git a/cosyvoice/utils/__init__.py b/cosyvoice/utils/__init__.py
diff --git a/cosyvoice/utils/class_utils.py b/cosyvoice/utils/class_utils.py
diff --git a/cosyvoice/utils/common.py b/cosyvoice/utils/common.py
diff --git a/cosyvoice/utils/executor.py b/cosyvoice/utils/executor.py
diff --git a/cosyvoice/utils/file_utils.py b/cosyvoice/utils/file_utils.py
diff --git a/cosyvoice/utils/frontend_utils.py b/cosyvoice/utils/frontend_utils.py
diff --git a/cosyvoice/utils/mask.py b/cosyvoice/utils/mask.py
diff --git a/cosyvoice/utils/scheduler.py b/cosyvoice/utils/scheduler.py
diff --git a/cosyvoice/utils/train_utils.py b/cosyvoice/utils/train_utils.py
diff --git a/matcha/VERSION b/matcha/VERSION
diff --git a/matcha/__init__.py b/matcha/__init__.py
diff --git a/matcha/app.py b/matcha/app.py
diff --git a/matcha/cli.py b/matcha/cli.py
diff --git a/matcha/data/__init__.py b/matcha/data/__init__.py
diff --git a/matcha/data/components/__init__.py b/matcha/data/components/__init__.py
diff --git a/matcha/data/text_mel_datamodule.py b/matcha/data/text_mel_datamodule.py
diff --git a/matcha/hifigan/LICENSE b/matcha/hifigan/LICENSE
diff --git a/matcha/hifigan/README.md b/matcha/hifigan/README.md
diff --git a/matcha/hifigan/__init__.py b/matcha/hifigan/__init__.py
diff --git a/matcha/hifigan/config.py b/matcha/hifigan/config.py
diff --git a/matcha/hifigan/denoiser.py b/matcha/hifigan/denoiser.py
diff --git a/matcha/hifigan/env.py b/matcha/hifigan/env.py
diff --git a/matcha/hifigan/meldataset.py b/matcha/hifigan/meldataset.py
diff --git a/matcha/hifigan/models.py b/matcha/hifigan/models.py
diff --git a/matcha/hifigan/xutils.py b/matcha/hifigan/xutils.py
diff --git a/matcha/models/__init__.py b/matcha/models/__init__.py
diff --git a/matcha/models/baselightningmodule.py b/matcha/models/baselightningmodule.py
diff --git a/matcha/models/components/__init__.py b/matcha/models/components/__init__.py
diff --git a/matcha/models/components/decoder.py b/matcha/models/components/decoder.py
diff --git a/matcha/models/components/flow_matching.py b/matcha/models/components/flow_matching.py
diff --git a/matcha/models/components/text_encoder.py b/matcha/models/components/text_encoder.py
diff --git a/matcha/models/components/transformer.py b/matcha/models/components/transformer.py
diff --git a/matcha/models/matcha_tts.py b/matcha/models/matcha_tts.py
diff --git a/matcha/onnx/__init__.py b/matcha/onnx/__init__.py
diff --git a/matcha/onnx/export.py b/matcha/onnx/export.py
diff --git a/matcha/onnx/infer.py b/matcha/onnx/infer.py
diff --git a/matcha/text/__init__.py b/matcha/text/__init__.py
diff --git a/matcha/text/cleaners.py b/matcha/text/cleaners.py
diff --git a/matcha/text/numbers.py b/matcha/text/numbers.py
diff --git a/matcha/text/symbols.py b/matcha/text/symbols.py
diff --git a/matcha/train.py b/matcha/train.py
diff --git a/matcha/utils/__init__.py b/matcha/utils/__init__.py
diff --git a/matcha/utils/audio.py b/matcha/utils/audio.py
diff --git a/matcha/utils/generate_data_statistics.py b/matcha/utils/generate_data_statistics.py
diff --git a/matcha/utils/get_durations_from_trained_model.py b/matcha/utils/get_durations_from_trained_model.py
diff --git a/matcha/utils/instantiators.py b/matcha/utils/instantiators.py
diff --git a/matcha/utils/logging_utils.py b/matcha/utils/logging_utils.py
diff --git a/matcha/utils/model.py b/matcha/utils/model.py
diff --git a/matcha/utils/monotonic_align/__init__.py b/matcha/utils/monotonic_align/__init__.py
diff --git a/matcha/utils/monotonic_align/core.pyx b/matcha/utils/monotonic_align/core.pyx
diff --git a/matcha/utils/monotonic_align/setup.py b/matcha/utils/monotonic_align/setup.py
diff --git a/matcha/utils/pylogger.py b/matcha/utils/pylogger.py
diff --git a/matcha/utils/rich_utils.py b/matcha/utils/rich_utils.py
diff --git a/matcha/utils/utils.py b/matcha/utils/utils.py
diff --git a/nodes.py b/nodes.py
@@ -6,12 +6,13 @@
 import numpy as np
 import os,sys
 import folder_paths
+
 now_dir = os.path.dirname(os.path.abspath(__file__))
 sys.path.append(now_dir)
 input_dir = folder_paths.get_input_directory()
 output_dir = os.path.join(folder_paths.get_output_directory(),"cosyvoice_dubb")
-pretrained_models = os.path.join(now_dir,"pretrained_models")
-
+#pretrained_models = os.path.join(now_dir,"pretrained_models")
+pretrained_models = os.path.join(folder_paths.models_dir, "CosyVoice")
 from modelscope import snapshot_download
 
 import ffmpeg
@@ -344,3 +345,45 @@ def INPUT_TYPES(s):
     def load_srt(self, srt):
         srt_path = folder_paths.get_annotated_filepath(srt)
         return (srt_path,)
+
+class LoadAudio:
+    @classmethod
+    def INPUT_TYPES(s):
+        files = [f for f in os.listdir(input_dir) if os.path.isfile(os.path.join(input_dir, f)) and f.split('.')[-1] in ["wav", "mp3","WAV","flac","m4a"]]
+        return {"required":
+                    {"audio": (sorted(files),)},
+                }
+    CATEGORY = "AIFSH_CosyVoice"
+    RETURN_TYPES = ("AUDIO",)
+    FUNCTION = "load_audio"
+    def load_audio(self, audio):
+        audio_path = folder_paths.get_annotated_filepath(audio)
+        #读取文件
+        waveform, sample_rate = torchaudio.load(audio_path)
+        audio = {"waveform": waveform.unsqueeze(0),"sample_rate":sample_rate}
+        return (audio,)
+
+class PreviewAudio:
+    def __init__(self):
+        self.preview_count = 00000
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required":
+                    {"audio": ("AUDIO",),}
+                }
+    CATEGORY = "AIFSH_CosyVoice"
+    RETURN_TYPES = ()
+    OUTPUT_NODE = True
+    FUNCTION = "preview_audio"
+    def preview_audio(self, audio):
+        #保存文件
+        filename = "audio_out_"+ f"{self.preview_count:05}"+".wav"
+        self.preview_count += 1
+        audioname = os.path.join(folder_paths.get_output_directory(),filename)
+        torchaudio.save(audioname,audio["waveform"].squeeze(0),audio["sample_rate"])
+        #torchaudio.save(os.path.join(full_output_folder, file), waveform, audio["sample_rate"], format="FLAC")
+        #
+        audio_name = os.path.basename(audioname)
+        tmp_path = os.path.dirname(audioname)
+        audio_root = os.path.basename(tmp_path)
+        return {"ui": {"audio":[audio_name,audio_root]}}
diff --git a/requirements.txt b/requirements.txt
diff --git a/web/js/previewAudio.js b/web/js/previewAudio.js
@@ -0,0 +1,155 @@
+import { app } from "../../../scripts/app.js";
+import { api } from '../../../scripts/api.js'
+
+function fitHeight(node) {
+    node.setSize([node.size[0], node.computeSize([node.size[0], node.size[1]])[1]])
+    node?.graph?.setDirtyCanvas(true);
+}
+function chainCallback(object, property, callback) {
+    if (object == undefined) {
+        //This should not happen.
+        console.error("Tried to add callback to non-existant object")
+        return;
+    }
+    if (property in object) {
+        const callback_orig = object[property]
+        object[property] = function () {
+            const r = callback_orig.apply(this, arguments);
+            callback.apply(this, arguments);
+            return r
+        };
+    } else {
+        object[property] = callback;
+    }
+}
+
+function addPreviewOptions(nodeType) {
+    chainCallback(nodeType.prototype, "getExtraMenuOptions", function(_, options) {
+        // The intended way of appending options is returning a list of extra options,
+        // but this isn't used in widgetInputs.js and would require
+        // less generalization of chainCallback
+        let optNew = []
+        try {
+            const previewWidget = this.widgets.find((w) => w.name === "audiopreview");
+
+            let url = null
+            if (previewWidget.audioEl?.hidden == false && previewWidget.audioEl.src) {
+                //Use full quality audio
+                //url = api.apiURL('/view?' + new URLSearchParams(previewWidget.value.params));
+                url = previewWidget.audioEl.src
+            }
+            if (url) {
+                optNew.push(
+                    {
+                        content: "Open preview",
+                        callback: () => {
+                            window.open(url, "_blank")
+                        },
+                    },
+                    {
+                        content: "Save preview",
+                        callback: () => {
+                            const a = document.createElement("a");
+                            a.href = url;
+                            a.setAttribute("download", new URLSearchParams(previewWidget.value.params).get("filename"));
+                            document.body.append(a);
+                            a.click();
+                            requestAnimationFrame(() => a.remove());
+                        },
+                    }
+                );
+            }
+            if(options.length > 0 && options[0] != null && optNew.length > 0) {
+                optNew.push(null);
+            }
+            options.unshift(...optNew);
+
+        } catch (error) {
+            console.log(error);
+        }
+
+    });
+}
+function previewAudio(node,file,type){
+    var element = document.createElement("div");
+    const previewNode = node;
+    var previewWidget = node.addDOMWidget("audiopreview", "preview", element, {
+        serialize: false,
+        hideOnZoom: false,
+        getValue() {
+            return element.value;
+        },
+        setValue(v) {
+            element.value = v;
+        },
+    });
+    previewWidget.computeSize = function(width) {
+        if (this.aspectRatio && !this.parentEl.hidden) {
+            let height = (previewNode.size[0]-20)/ this.aspectRatio + 10;
+            if (!(height > 0)) {
+                height = 0;
+            }
+            this.computedHeight = height + 10;
+            return [width, height];
+        }
+        return [width, -4];//no loaded src, widget should not display
+    }
+    // element.style['pointer-events'] = "none"
+    previewWidget.value = {hidden: false, paused: false, params: {}}
+    previewWidget.parentEl = document.createElement("div");
+    previewWidget.parentEl.className = "audio_preview";
+    previewWidget.parentEl.style['width'] = "100%"
+    element.appendChild(previewWidget.parentEl);
+    previewWidget.audioEl = document.createElement("audio");
+    previewWidget.audioEl.controls = true;
+    previewWidget.audioEl.loop = false;
+    previewWidget.audioEl.muted = false;
+    previewWidget.audioEl.style['width'] = "100%"
+    previewWidget.audioEl.addEventListener("loadedmetadata", () => {
+
+        previewWidget.aspectRatio = previewWidget.audioEl.audioWidth / previewWidget.audioEl.audioHeight;
+        fitHeight(this);
+    });
+    previewWidget.audioEl.addEventListener("error", () => {
+        //TODO: consider a way to properly notify the user why a preview isn't shown.
+        previewWidget.parentEl.hidden = true;
+        fitHeight(this);
+    });
+
+    let params =  {
+        "filename": file,
+        "type": type,
+    }
+
+    previewWidget.parentEl.hidden = previewWidget.value.hidden;
+    previewWidget.audioEl.autoplay = !previewWidget.value.paused && !previewWidget.value.hidden;
+    let target_width = 256
+    if (element.style?.width) {
+        //overscale to allow scrolling. Endpoint won't return higher than native
+        target_width = element.style.width.slice(0,-2)*2;
+    }
+    if (!params.force_size || params.force_size.includes("?") || params.force_size == "Disabled") {
+        params.force_size = target_width+"x?"
+    } else {
+        let size = params.force_size.split("x")
+        let ar = parseInt(size[0])/parseInt(size[1])
+        params.force_size = target_width+"x"+(target_width/ar)
+    }
+
+    previewWidget.audioEl.src = api.apiURL('/view?' + new URLSearchParams(params));
+
+    previewWidget.audioEl.hidden = false;
+    previewWidget.parentEl.appendChild(previewWidget.audioEl)
+}
+
+app.registerExtension({
+	name: "CosyVoice.AudioPreviewer",
+	async beforeRegisterNodeDef(nodeType, nodeData, app) {
+		if (nodeData?.name == "PreviewAudio") {
+			nodeType.prototype.onExecuted = function (data) {
+				previewAudio(this, data.audio[0], data.audio[1]);
+			}
+            addPreviewOptions(nodeType)
+		}
+	}
+});