feat: add Pattern Extraction preset - 新增花纹提取预设，支持指定提取对象

nodes/tools/kontext_presets.py

@@ -131,6 +131,11 @@ class LoadKontextPresets_UTK:
                 "name": "Universal Editor (万能编辑)",
                 "brief": "You are a task-aware prompt translator for single-image editing using the Kontext format with the Flux model. Your role is to convert the user's natural language instruction into a clean, precise, and visually consistent image editing directive. Follow these strict rules: 1. Do not describe the original image - assume the model sees it. 2. Only describe what needs to change using clear verbs like replace, add, remove, insert, transform into, convert to. 3. Always state what must remain unchanged: for persons preserve facial features, expression, pose, hairstyle, skin tone, body proportion, clothing texture, and lighting; for objects preserve shape, scale, material, texture, surface details, lighting, reflections, and shadow behavior. 4. For background changes describe only the new scene, keep the main subject in exact same position, scale, lighting, and focus, ensure visual integration. 5. For style transfer name the visual style specifically and maintain original composition. 6. For text editing use quotes for target text and specify font retention if needed. 7. For composite edits ensure product stays visually dominant and unchanged with seamless integration. 8. Break complex changes into clear simple steps. 9. Output must be one paragraph in natural English starting with change description, including what to preserve, ending with visual consistency requirements. Based on the user's specific request: $user_prompt$, provide a precise editing instruction that follows these guidelines.",
             },
+            # === 花纹提取类 ===
+            {
+                "name": "Pattern Extraction (花纹提取)",
+                "brief": "Extract the visible pattern or logo from the $user_prompt$ in the image and convert it into a seamless flat texture. Remove all lighting, shading, wrinkles, folds, and perspective distortions. Ensure the output is a clean, high-resolution top-down view of the pattern with no character, background, or surrounding elements included. Preserve the original colors, fine details, and relative scale of the pattern as seen on the specified object. The extraction should isolate the pattern completely, eliminating any 3D effects, shadows, or environmental influences while maintaining the authentic visual characteristics and color palette of the original design.",
+            },
         ],
         "suffix": "Your response must consist of concise instruction ready for the image editing AI. Do not add any conversational text, explanations, or deviations; only the instructions.",
     }