Added support for Anthropic models

modules/textInferenceInputsImages.py

@@ -0,0 +1,50 @@
+"""Runware Text Inference Inputs Images — builds inputs.images from ComfyUI IMAGE tensors."""
+
+from typing import Any, Dict, List, Tuple
+
+from .utils import runwareUtils as rwUtils
+
+
+class RunwareTextInferenceInputsImages:
+    """Collect IMAGE tensors and convert them for textInference task.inputs.images.
+
+    Each connected IMAGE is converted via runwareUtils.convertTensor2IMG, which returns
+    either a cached mediaUUID (when caching is enabled) or a base64 data URI.
+    """
+
+    _MAX_IMAGES = 8
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        optional: Dict[str, tuple] = {}
+        for i in range(1, cls._MAX_IMAGES + 1):
+            ordinal = rwUtils.getOrdinal(i)
+            optional[f"Image {i}"] = ("IMAGE", {
+                "tooltip": (
+                    f"{ordinal.capitalize()} image for the model. Connect a Load Image (or any IMAGE source) node. "
+                    "Tensor is converted to base64 / mediaUUID before being sent to the API."
+                ),
+            })
+        return {"required": {}, "optional": optional}
+
+    RETURN_TYPES = ("RUNWARETEXTINFERENCEINPUTSIMAGES",)
+    RETURN_NAMES = ("images",)
+    FUNCTION = "createImages"
+    CATEGORY = "Runware/Text"
+    DESCRIPTION = (
+        "Collect one or more IMAGE tensors (Load Image, generated images, etc.) for Runware Text Inference "
+        "inputs.images. Connect the output to Runware Text Inference Inputs."
+    )
+
+    def createImages(self, **kwargs) -> Tuple[Dict[str, Any], ...]:
+        urls: List[str] = []
+        for i in range(1, self._MAX_IMAGES + 1):
+            tensor = kwargs.get(f"Image {i}", None)
+            if tensor is None:
+                continue
+            urls.append(rwUtils.convertTensor2IMG(tensor))
+
+        if not urls:
+            return ({},)
+
+        return ({"images": urls},)