Update src/core.py

src/core.py

@@ -22,52 +22,56 @@ log = logging.getLogger(__name__)
 DEFAULT_MODEL_ID = os.environ.get("GEMINI_IMAGE_MODEL", "gemini-3-pro-image-preview")
 DEFAULT_PROMPT = os.environ.get(
     "GEMINI_IMAGE_PROMPT",
-    (
-        "TASK: STRICT IMAGE INPAINTING — MASK-BOUNDED OBJECT REMOVAL ONLY\n"
+    ("TASK: STRICT IMAGE INPAINTING — MASK-BOUNDED OBJECT REMOVAL ONLY\n\n"
     "You are given:\n"
     "1) **Image A**: The original image, which contains various elements and objects.\n"
-    "2) **Image B**: A binary mask that highlights the areas to be removed from the original image.\n"
-    "\n"
-    "**Step 1: Image and Mask Classification**:\n"
-    "- First, **classify** the two images:\n"
-    "  - Image A is the **original image** with visible content that may include objects, background, and scene elements.\n"
-    "  - Image B is a **binary mask**, where:\n"
-    "    - **White pixels (#FFFFFF)** indicate areas of Image A that need to be removed.\n"
-    "    - **Black pixels (#000000)** indicate areas that must **remain unchanged** from Image A.\n"
-    "\n"
-    "**Step 2: Understand What to Remove**:\n"
-    "- **Examine Image B carefully** to identify **all white areas** (the regions that need removal).\n"
-    "- **Understand the objects and parts of objects** that the white mask covers. If the mask partially overlaps any object, only the covered area should be removed.\n"
-    "- The model must **recognize and understand** the object(s) within the white-masked regions before proceeding with the removal.\n"
-    "- **Do NOT attempt to remove any content outside the white mask**, even if it appears similar to the content inside the mask. Only remove what is explicitly masked in Image B.\n"
+    "2) **Image B**: A binary mask that highlights the areas to be removed from the original image.\n\n"
+    
+    "MASK INTERPRETATION (CRITICAL):\n\n"
+    "• **White pixels (#FFFFFF)** in Image B represent **the exact regions** in Image A that need to be removed. Only these regions can be modified or deleted.\n"
+    "• **Black pixels (#000000)** in Image B must **remain completely unchanged** in Image A. Do **NOT** modify or generate anything in these black areas.\n"
+    "• **Do NOT** modify or alter any other parts of Image A that are outside the white-masked region. Do not paint, blur, or alter any pixels that are not covered by the white mask.\n"
+    "• **Do NOT** attempt to remove or alter anything in the unmasked regions, even if the object in the unmasked area looks similar to the object inside the mask.\n"
+    "• **The white-masked region is the **only** area to be modified. The rest of the image must **remain intact**. No other object or area should be painted or modified.\n"
     "\n"
-    "**Step 3: Object Removal**:\n"
-    "- **Completely remove** everything inside the white-masked region. This includes:\n"
+    
+    "STEP 1: OBJECT REMOVAL ANALYSIS\n\n"
+    "• **Analyze the mask carefully** and **identify the exact areas** to be removed. Only focus on the white-masked region for removal.\n"
+    "• The model must recognize what needs to be removed based on the white mask and **do nothing to the black (unmasked) regions**.\n"
+    "• Do **NOT** modify, erase, or paint anything outside the white-masked region. Only remove what is explicitly covered by the white mask.\n\n"
+    
+    "STEP 2: OBJECT REMOVAL\n\n"
+    "• **Completely remove** everything inside the white-masked region in Image A. This includes:\n"
     "    - Erasing the object(s) entirely with **no visible remnants**, including shadows, outlines, or color traces.\n"
     "    - If the mask overlaps only part of an object, **only remove the pixels covered by the white mask**, leaving the rest of the object untouched.\n"
-    "- The goal is to **clear the masked region** entirely, **removing all visual content** from it.\n"
-    "\n"
-    "**Step 4: Inpainting and Filling the Masked Area**:\n"
-    "- After removal, **fill the masked area** using **only the immediate surrounding background pixels** from Image A.\n"
-    "- Continue existing **textures, colors, lighting, and geometry naturally**.\n"
-    "    - Ensure that the **lighting**, **brightness**, **contrast**, and **color temperature** of the inpainted region perfectly match the surrounding unmasked areas.\n"
-    "    - Do **not invent new structures, objects, shadows**, or **reflections** that were not originally present in the scene.\n"
-    "    - **Do not smooth, blur, or stylize** the inpainted region. The transition should be smooth and seamless.\n"
-    "\n"
-    "STRICT CONSTRAINTS (NO EXCEPTIONS):\n"
-    "- **Do NOT modify any black (unmasked) pixels**. The unmasked areas must remain exactly as they are in Image A.\n"
-    "- **Do NOT remove entire objects** unless they are **fully masked**. Only the regions inside the white mask should be modified.\n"
-    "- **Do NOT alter the framing, perspective**, or **global appearance** of the image. The overall structure of the image must remain intact.\n"
-    "- **Do NOT apply any semantic cleanup** or scene optimization. This is a **strict removal task**; only the content inside the white mask should be removed.\n"
-    "\n"
-    "QUALITY REQUIREMENTS:\n"
-    "- The final result must look like the object never existed in the original image.\n"
-    "- The inpainted region should blend **seamlessly** with surrounding pixels.\n"
-    "- **No visible artifacts**, such as ghosting, halos, color shifts, or texture repetition, should be present.\n"
-    "- The transition between the inpainted region and the unmasked areas must be **natural** and **imperceptible**.\n"
-    "\n"
-    "IMPORTANT:\n"
-    "If a conflict exists between **semantic understanding** and the **mask**, **ALWAYS follow the mask exactly**. The mask is the **absolute guide** for removal, and no other interpretation or inference should take place. The model should strictly follow the instructions from the mask without altering the intended areas."
+    "• The goal is to **clear the masked region** entirely, **removing all visual content** from it.\n\n"
+    
+    "IMPORTANT NOTE ABOUT INPAINTING:\n\n"
+    "• **The inpainted region must NOT be filled with pure white (#FFFFFF)**. The model should **never** generate a blank, white, or featureless region.\n"
+    "• The white areas should be filled with **the natural surrounding background** of Image A. This means extending **existing textures, colors, lighting, and geometry naturally**, without introducing any artificial white filling.\n"
+    "• **Do not paint or alter any other region** in the image, including the unmasked areas. **Only the white-masked area** should be modified, and no other content should be painted or changed.\n\n"
+    
+    "INPAINTING RULES:\n\n"
+    "• After the object is removed, **fill the masked area** using **only the immediate surrounding pixels** from Image A.\n"
+    "• **Do not invent new objects, structures, or details**. Only extend or copy background textures, lighting, and geometry from the unmasked areas.\n"
+    "• The lighting, contrast, brightness, color temperature, and texture in the inpainted region must **perfectly match** the surrounding areas.\n"
+    "• **Do not add** shadows, reflections, or new structures to the inpainted region.\n"
+    "• The inpainted region should seamlessly blend with the surrounding pixels. It should appear as though the object was **never there**.\n\n"
+    
+    "STRICT CONSTRAINTS (NO EXCEPTIONS):\n\n"
+    "• **Do NOT modify any black (unmasked) pixels**. The unmasked areas must remain **exactly** as they are in Image A.\n"
+    "• **Do NOT remove entire objects** unless they are **fully and completely masked**. Only the regions inside the white mask should be modified.\n"
+    "• **Do NOT alter the original image's **framing**, **perspective**, or **composition**. Only the regions inside the white mask should be affected.\n"
+    "• **Do NOT apply any semantic cleanup** or scene optimization. This is a **strict removal task**; only the content inside the white mask should be removed.\n"
+    
+    "QUALITY REQUIREMENTS:\n\n"
+    "• The final result should look as though the object **never existed** in the original image.\n"
+    "• The inpainted region must blend **seamlessly** with surrounding pixels.\n"
+    "• **No visible artifacts**, such as ghosting, halos, color shifts, or texture repetition, should be present.\n"
+    "• The transition between the inpainted region and the unmasked areas must be **natural** and **imperceptible**.\n\n"
+    
+    "IMPORTANT:\n\n"
+    "If there is a conflict between **semantic understanding** and the **mask**, **ALWAYS follow the mask exactly**. The mask is the **absolute guide**, and no interpretation should deviate from the mask’s instructions.\n"
     ),
 )
 _GENAI_MODEL: genai.GenerativeModel | None = None