first upload

text switch(with add) clip switch condition switch image switch model switch
2026-04-22 22:56:25 +08:00
commit 25399acf49
12 changed files with 380 additions and 0 deletions
@@ -0,0 +1,57 @@
+import torch
+import gc
+
+class Iter_ConditioningSwitch:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "output_index": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
+                "input_count": ("INT", {"default": 2, "min": 1, "max": 64}),
+                "use_last_input": ("BOOLEAN", {"default": True}),
+            },
+        }
+
+    RETURN_TYPES = ("CONDITIONING",)
+    FUNCTION = "switch_conditioning"
+    CATEGORY = "CustomNodes/Iterators"
+
+    def switch_conditioning(self, output_index, input_count, use_last_input, **kwargs):
+        target_idx = output_index % input_count
+        res = None
+        for i in range(target_idx, -1, -1):
+            val = kwargs.get(f"cond_{i}")
+            if val is not None or not use_last_input:
+                res = val
+                break
+        #gc.collect()
+        #torch.cuda.empty_cache() 
+        return (res,)
+
+class Iter_CLIPSwitch:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "output_index": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
+                "input_count": ("INT", {"default": 2, "min": 1, "max": 64}),
+                "use_last_input": ("BOOLEAN", {"default": True}),
+            },
+        }
+
+    RETURN_TYPES = ("CLIP",)
+    FUNCTION = "switch_clip"
+    CATEGORY = "CustomNodes/Iterators"
+
+    def switch_clip(self, output_index, input_count, use_last_input, **kwargs):
+        target_idx = output_index % input_count
+        res = None
+        for i in range(target_idx, -1, -1):
+            val = kwargs.get(f"clip_{i}")
+            if val is not None or not use_last_input:
+                res = val
+                break
+        #gc.collect()
+        #torch.cuda.empty_cache()         
+        return (res,)
+    
@@ -0,0 +1,29 @@
+class Iter_FloatSwitch:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "output_index": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
+                "input_count": ("INT", {"default": 2, "min": 1, "max": 128, "step": 1}),
+                "use_last_input": ("BOOLEAN", {"default": True}),
+            },
+        }
+
+    RETURN_TYPES = ("FLOAT",)
+    FUNCTION = "switch_float"
+    CATEGORY = "CustomNodes/Iterators"
+
+    def switch_float(self, output_index, input_count, use_last_input, **kwargs):
+        # 1. 計算目標索引
+        target_idx = output_index % input_count
+        
+        # 2. 搜尋有效浮點數
+        res_float = None
+        for i in range(target_idx, -1, -1):
+            val = kwargs.get(f"float_{i}")
+            # 注意：在 Python 中 0.0 是 False，所以要檢查是否為 None
+            if val is not None or not use_last_input:
+                res_float = val
+                break
+        
+        return (float(res_float),)
@@ -0,0 +1,35 @@
+import torch
+import gc
+
+
+class Iter_ImageSwitch:
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "output_index": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
+                "input_count": ("INT", {"default": 2, "min": 1, "max": 64, "step": 1}),
+                "use_last_input": ("BOOLEAN", {"default": False}),
+            },
+        }
+
+    RETURN_TYPES = ("IMAGE",)
+    FUNCTION = "switch_image"
+    CATEGORY = "CustomNodes/Iterators"
+
+    def switch_image(self, output_index, input_count, use_last_input, **kwargs):
+        # 1. 取得目標索引（對 input_count 取模，確保不溢出）
+        # 這樣 index 0~n 就會嚴格對應到 image_0 ~ image_n
+        target_idx = output_index % input_count
+        
+        # 2. 從目標索引開始「往回」搜尋有效接點
+        res_image = None
+        for i in range(target_idx, -1, -1):
+            val = kwargs.get(f"image_{i}")
+            if val is not None or not use_last_input:
+                res_image = val
+                break
+        
+        gc.collect()
+        torch.cuda.empty_cache() 
+        return (res_image,)
@@ -0,0 +1,40 @@
+import torch
+import gc
+
+class Iter_ModelSwitch:
+    def __init__(self):
+        pass
+
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                "output_index": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
+                "input_count": ("INT", {"default": 2, "min": 1, "max": 64, "step": 1}),
+                "use_last_input": ("BOOLEAN", {"default": True}),
+            },
+        }
+
+    RETURN_TYPES = ("MODEL",)
+    FUNCTION = "switch_model"
+    CATEGORY = "CustomNodes/Iterators"
+
+    def switch_model(self, output_index, input_count, use_last_input, **kwargs):
+        # 1. 確定目標索引 (嚴格對應編號)
+        target_idx = output_index % input_count
+        
+        # 2. 往回搜尋有效模型
+        res_model = None
+        for i in range(target_idx, -1, -1):
+            val = kwargs.get(f"model_{i}")
+            if val is not None or not use_last_input:
+                res_model = val
+                break
+    
+
+        # 4. VRAM 管理
+        #if force_gc == "enable":
+        #gc.collect()
+        #torch.cuda.empty_cache()
+
+        return (res_model,)
@@ -0,0 +1,68 @@
+import torch
+import gc
+
+class Iter_TextSwitch:
+    def __init__(self):
+        pass
+
+    @classmethod
+    def INPUT_TYPES(s):
+        return {
+            "required": {
+                # 預設為空字串，允許不接線
+                "add_text": ("STRING", {"forceInput": True, "default": ""}),
+                "output_index": ("INT", {"default": 0, "min": 0, "max": 0xffffffffffffffff}),
+                "input_count": ("INT", {"default": 2, "min": 1, "max": 128, "step": 1}),
+                "use_last_input": ("BOOLEAN", {"default": False}),
+            },
+            "optional": {
+                # 這裡不需要寫死 text_0，因為 JS 會動態產生所有輸入
+            }
+        }
+
+    RETURN_TYPES = ("STRING",)
+    FUNCTION = "switch_and_combine"
+    CATEGORY = "CustomNodes/Logic"
+
+    def switch_and_combine(self, add_text="", output_index=0, input_count=1, use_last_input=False, **kwargs):
+        """
+        # 1. 蒐集所有動態傳入的字串
+        texts = []
+        for i in range(input_count):
+            val = kwargs.get(f"text_{i}")
+            # 只有當該接點有連線且有值時才加入（避免計算空的接點）
+            if val is not None:
+                texts.append(val)
+        
+        # 2. 安全性檢查：如果完全沒有輸入接點有連線
+        if not texts:
+            # 如果連 add_text 都沒接，就回傳純空字串
+            return (add_text if add_text is not None else "",)
+
+        # 3. 根據索引選取文本（防止除以零）
+        target_index = output_index % len(texts)
+        selected_text = texts[target_index]
+
+        # 4. 串接並回傳
+        # 確保兩者皆為字串，避免 NoneType 相加報錯
+        final_header = add_text if add_text is not None else ""
+        final_selected = selected_text if selected_text is not None else ""
+        gc.collect()
+        torch.cuda.empty_cache() 
+        return (final_header + final_selected,)
+        """
+        # 1. 取得目標索引（對 input_count 取模，確保不溢出）
+        # 這樣 index 0~n 就會嚴格對應到 image_0 ~ image_n
+        target_idx = output_index % input_count
+        
+        # 2. 從目標索引開始「往回」搜尋有效接點
+        res_txt = None
+        for i in range(target_idx, -1, -1):
+            val = kwargs.get(f"text_{i}")
+            if val is not None or not use_last_input:
+                res_txt = val
+                break
+        
+        #gc.collect()
+        #torch.cuda.empty_cache() 
+        return (add_text + res_txt,)
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ChengJhong (younger15)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
@@ -0,0 +1,40 @@
+# ComfyUI-IterSwitch
+A set of dynamic switch nodes for ComfyUI, designed to simplify complex workflows by allowing flexible, iterative input selection. These nodes provide dynamic connection management and intelligent fallback logic.
+
+## Key Features
+Dynamic Inputs: Automatically resizes the number of input connections based on the input_count widget.
+
+## Smart Fallback: 
+With the use_last_input option enabled, the node can search backwards from the current index to find the nearest valid (connected) input, ensuring a result is always returned.
+
+## Type Support: 
+Includes switches for Text, Conditioning, CLIP, Image, Model, and Float types.
+
+### Example Usage: Iter_TextSwitch
+The Iter_TextSwitch node allows you to combine a base string with one of many dynamically connected inputs.
+
+1. Basic Connection
+The node takes an add_text input (base string) and allows you to dynamically define multiple text inputs. You can select the specific input using output_index.
+![text switch sample](assets/example_iterTextSwitch01.png)
+*Example using [waiIllustrious v16](https://civitai.com/models/827184?modelVersionId=2514310)*
+2. Fallback Logic
+When use_last_input is enabled, the node becomes more robust. If your current output_index points to a disconnected or empty input, the node will automatically search backwards through your inputs until it finds the most recent valid (connected) string.
+![will use last valid input if use_last_input checked](assets/example_iterTextSwitch02.png)
+*Example using [waiIllustrious v16](https://civitai.com/models/827184?modelVersionId=2514310)*
+3. Iterative Workflow
+By using the output_index, you can easily integrate this node into iterative loops, ensuring that your text prompts are dynamically selected or concatenated based on your workflow state.
+![text switch sample2](assets/example_iterTextSwitch03.png)
+*Example using [waiIllustrious v16](https://civitai.com/models/827184?modelVersionId=2514310)*
+
+
+
+
+## Installation
+Navigate to your ComfyUI custom_nodes directory.
+
+Run:
+
+Bash
+`git clone https://github.com/YOUR_USERNAME/ComfyUI-SwitchForIteration.git`
+Restart ComfyUI.
+
@@ -0,0 +1,28 @@
+from .Iter_TextSwitch import Iter_TextSwitch
+from .Iter_CondClipSwitch import Iter_ConditioningSwitch, Iter_CLIPSwitch
+from .Iter_ImageSwitch import Iter_ImageSwitch
+from .Iter_ModelSwitch import Iter_ModelSwitch
+from .Iter_FloatSwitch import Iter_FloatSwitch
+
+NODE_CLASS_MAPPINGS = {
+    "Iter_TextSwitch": Iter_TextSwitch,
+    "Iter_ConditioningSwitch": Iter_ConditioningSwitch,
+    "Iter_CLIPSwitch": Iter_CLIPSwitch,
+    "Iter_ImageSwitch": Iter_ImageSwitch,
+    "Iter_ModelSwitch": Iter_ModelSwitch,
+    "Iter_FloatSwitch": Iter_FloatSwitch
+}
+
+NODE_DISPLAY_NAME_MAPPINGS = {
+    "Iter_TextSwitch": "Iter_TextSwitch",
+    "Iter_ConditioningSwitch": "Iter_ConditioningSwitch",
+    "Iter_CLIPSwitch": "Iter_CLIPSwitch",
+    "Iter_ImageSwitch": "Iter_ImageSwitch",
+    "Iter_ModelSwitch": "Iter_ModelSwitch",
+    "Iter_FloatSwitch": "Iter_FloatSwitch"
+}
+
+# 確保 JS 資料夾被載入
+WEB_DIRECTORY = "./web"
+
+__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS", "WEB_DIRECTORY"]
@@ -0,0 +1,62 @@
+import { app } from "../../../scripts/app.js";
+
+// 通用的動態接點處理函數
+const setupSmartDynamicInputs = (nodeType, prefix, type) => {
+    const onNodeCreated = nodeType.prototype.onNodeCreated;
+    nodeType.prototype.onNodeCreated = function () {
+        onNodeCreated?.apply(this, arguments);
+        const inputCountWidget = this.widgets.find(w => w.name === "input_count");
+
+        inputCountWidget.callback = () => {
+            const targetCount = inputCountWidget.value;
+            
+            // 1. 找出目前所有屬於該 prefix 的輸入點
+            const currentInputs = this.inputs.filter(i => i.name.startsWith(`${prefix}_`));
+            const currentCount = currentInputs.length;
+
+            if (targetCount > currentCount) {
+                // 2. 如果目標數量較多：只新增缺少的接點，原有的接點不受影響
+                for (let i = currentCount; i < targetCount; i++) {
+                    this.addInput(`${prefix}_${i}`, type);
+                }
+            } else if (targetCount < currentCount) {
+                // 3. 如果目標數量較少：從最後面開始移除，保留前面的連線
+                for (let i = currentCount - 1; i >= targetCount; i--) {
+                    // 找到該名稱對應的索引並移除
+                    const inputIdx = this.inputs.findIndex(input => input.name === `${prefix}_${i}`);
+                    if (inputIdx !== -1) {
+                        this.removeInput(inputIdx);
+                    }
+                }
+            }
+
+            // 重新整理 UI 配置
+            this.setSize(this.computeSize());
+        };
+
+        // 延遲初始化，確保節點建立時 UI 正確
+        setTimeout(() => {
+            if (inputCountWidget.callback) inputCountWidget.callback();
+        }, 50);
+    };
+};
+
+app.registerExtension({
+    name: "CustomNodes.IterSwitches",
+    async beforeRegisterNodeDef(nodeType, nodeData, app) {
+        // 根據節點名稱套用不同的前綴與類型
+        if (nodeData.name === "Iter_TextSwitch") {
+            setupSmartDynamicInputs(nodeType, "text", "STRING");
+        } else if (nodeData.name === "Iter_ConditioningSwitch") {
+            setupSmartDynamicInputs(nodeType, "cond", "CONDITIONING");
+        } else if (nodeData.name === "Iter_CLIPSwitch") {
+            setupSmartDynamicInputs(nodeType, "clip", "CLIP");
+        } else if (nodeData.name === "Iter_ImageSwitch") {
+            setupSmartDynamicInputs(nodeType, "image", "IMAGE");
+        } else if (nodeData.name === "Iter_ModelSwitch") {
+            setupSmartDynamicInputs(nodeType, "model", "MODEL");
+        } else if (nodeData.name === "Iter_FloatSwitch") {
+            setupSmartDynamicInputs(nodeType, "float", "FLOAT");
+        }
+    }
+});