Add widget mappings for all WanVideo workflow nodes

The convert_frontend_to_api() function was missing mappings for most
node types, causing "Required input is missing" errors when the API
received workflows in frontend format.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

handler.py

@@ -122,46 +122,63 @@ def convert_frontend_to_api(frontend_workflow: dict) -> dict:
         link_id, src_node, src_slot, dst_node, dst_slot, link_type = link[:6]
         link_map[link_id] = (str(src_node), src_slot)
 
+    # Widget mappings for each node type: list of input names in order
+    WIDGET_MAPPINGS = {
+        "LoadImage": ["image", "upload"],
+        "CLIPTextEncode": ["text"],
+        "easy int": ["value"],
+        "INTConstant": ["value"],
+        "SaveVideo": ["filename_prefix", "format", "codec"],
+        "CreateVideo": ["fps"],
+        "RIFE VFI": ["ckpt_name", "clear_cache_after_n_frames", "multiplier", "fast_mode", "ensemble", "scale_factor"],
+        "CLIPLoader": ["clip_name", "type", "device"],
+        "GetImageSize": [],
+        "WanVideoTorchCompileSettings": [
+            "backend", "fullgraph", "mode", "dynamic", "dynamo_cache_size_limit",
+            "compile_transformer_blocks_only", "compile_cache_max_entries",
+            "compile_single_blocks", "compile_double_blocks"
+        ],
+        "WanVideoBlockSwap": [
+            "blocks_to_swap", "offload_txt_emb", "offload_img_emb", "offload_txt_clip_emb",
+            "offload_modulation", "cpu_offload_streams", "use_async_transfer"
+        ],
+        "WanVideoModelLoader": [
+            "model", "base_precision", "quantization", "load_device", "attention_mode", "lora_scale_mode"
+        ],
+        "WanVideoLoraSelect": ["lora", "strength", "use_lora_sparse", "sparse_lora_blocks"],
+        "WanVideoVAELoader": ["model_name", "precision"],
+        "WanVideoTextEmbedBridge": [],
+        "WanVideoImageToVideoEncode": [
+            "width", "height", "num_frames", "noise_aug_strength",
+            "start_latent_strength", "end_latent_strength", "image_noise_aug",
+            "mask_noise_aug", "force_offload"
+        ],
+        "WanVideoSampler": [
+            "shift", "cfg", "steps", "seed", "seed_mode", "force_offload", "scheduler",
+            "riflex_freq_index", "riflex_freq_dim", "riflex_freq_scale",
+            "use_comfy_pbar", "start_step", "end_step", "denoise"
+        ],
+        "WanVideoDecode": [
+            "enable_tiling", "tile_x", "tile_y", "tile_stride_x", "tile_stride_y", "tiling_decoder"
+        ],
+        "MathExpression|pysssss": ["expression"],
+    }
+
     for node in nodes:
         node_id = str(node["id"])
         class_type = node.get("type", "")
 
         inputs = {}
 
-        # Process widget values
+        # Process widget values using mappings
         widgets_values = node.get("widgets_values", [])
+        widget_names = WIDGET_MAPPINGS.get(class_type, [])
 
-        # Map widget values based on class type
+        for i, value in enumerate(widgets_values):
-        # This is a simplified mapping - specific nodes may need custom handling
+            if i < len(widget_names) and widget_names[i]:
-        if class_type == "LoadImage" and len(widgets_values) >= 1:
+                inputs[widget_names[i]] = value
-            inputs["image"] = widgets_values[0]
-            if len(widgets_values) >= 2:
-                inputs["upload"] = widgets_values[1]
 
-        elif class_type == "CLIPTextEncode" and len(widgets_values) >= 1:
+        # Process node inputs (connections) - these override widget values
-            inputs["text"] = widgets_values[0]
-
-        elif class_type in ["easy int", "INTConstant"] and len(widgets_values) >= 1:
-            inputs["value"] = widgets_values[0]
-
-        elif class_type == "SaveVideo" and len(widgets_values) >= 1:
-            inputs["filename_prefix"] = widgets_values[0]
-            if len(widgets_values) >= 2:
-                inputs["format"] = widgets_values[1]
-            if len(widgets_values) >= 3:
-                inputs["codec"] = widgets_values[2]
-
-        elif class_type == "CreateVideo" and len(widgets_values) >= 1:
-            inputs["frame_rate"] = widgets_values[0]
-
-        elif class_type == "RIFE VFI" and len(widgets_values) >= 1:
-            inputs["ckpt_name"] = widgets_values[0]
-            if len(widgets_values) >= 2:
-                inputs["clear_cache_after_n_frames"] = widgets_values[1]
-            if len(widgets_values) >= 3:
-                inputs["multiplier"] = widgets_values[2]
-
-        # Process node inputs (connections)
         for inp in node.get("inputs", []):
             inp_name = inp["name"]
             link_id = inp.get("link")