Add files via upload

Author: if-ai

IFChatPromptNode.py

@@ -777,7 +777,9 @@ async def process_chat(
                     else:
                         image_tensor, default_mask_tensor = process_images_for_comfy(
                             retrieved_image, 
-                            self.placeholder_image_path
+                            self.placeholder_image_path,
+                            response_key=None,
+                            field_name=None
                         )
                         mask_tensor = default_mask_tensor
 

IFDisplayTextWildcardNode.py

@@ -19,7 +19,7 @@ def __init__(self):
         self.wildcard_lock = threading.Lock()
 
         # Initialize paths
-        self.base_path = folder_paths.base_path
+        #self.base_path = folder_paths.base_path
         self.presets_dir = os.path.join(folder_paths.base_path, "custom_nodes", "ComfyUI-IF_AI_tools", "IF_AI", "presets")
         self.wildcards_dir = os.path.join(self.presets_dir, "wildcards") 
 

IFImagePromptNode.py

@@ -1,3 +1,4 @@
+# IFStepCounterNode.py
 class IFCounter:
     def __init__(self):
         self.counters = {}

IFStepCounterNode.py

@@ -1,7 +1,7 @@
 {
   "None": "",
   "AI_Fill": "lowres, lowquality, bad,",
-  "SD3x_q1": "deformed and malformed with worst qualities. distortions and artifacts, low quality, bad ratings ☆☆☆☆☆, 0/10 review, 0/5 review ↓",
+  "SD3x_q1": "deformed and malformed with worst qualities. distortions and artifacts, low quality, bad ratings, 0/10 review, 0/5 review ↓",
   "AtomeaseNoWaifus": "(3D, Adorable, Anime, CGI, Cartoon, Cartoonish, Childish, Computer Generated, Manga, Rendered)",
   "BadChill": "ng_deepnegative_v1_75t, bad-picture-chill-75v, bad_prompt, bad anatomy",
   "BadChill2": "bad-picture-chill-75v, ng_deepnegative_v1_75t, easynegative, verybadimagenegative_v1.3, negative_hand-neg",

IF_AI/presets/neg_prompts.json

@@ -4,16 +4,16 @@ base_ip: localhost
 port: '11434'
 llm_provider: openai
 llm_model: gpt-4o-mini
-temperature: 1.2000000000000002
-max_tokens: 2048
+temperature: 0.0
+max_tokens: 2061
 stop: null
 keep_alive: false
 top_k: 40
 top_p: 0.9
 repeat_penalty: 1.2
-seed: 558413258089916
-rag_folder_name: safiro1
-query_type: local
+seed: 197464410746652
+rag_folder_name: Safiro2
+query_type: global
 community_level: 2
 preset: Default
 external_llm_api_key: ''

IF_AI/rag/settings.yaml

@@ -4,7 +4,11 @@
 
 ComfyUI-IF_AI_tools is a set of custom nodes to Run Local and API LLMs and LMMs, Features OCR-RAG (Bialdy), nanoGraphRAG, Supervision Object Detection, supports Ollama, LlamaCPP LMstudio, Koboldcpp, TextGen, Transformers or via APIs Anthropic, Groq, OpenAI, Google Gemini, Mistral, xAI and create your own charcters assistants (SystemPrompts) with custom presets and muchmore
 
+################# ATENTION ####################
 
+   *COPY THE IF_AI FOLDER TO YOUR USER FOLDER*
+   
+###############################################
 # Prerequisite Installation (Poppler)
 
 To ensure compatibility and functionality with all tools, you may need `poppler` for PDF-related operations. Use `scoop` to install `poppler` on Windows:

README.md

@@ -25,8 +25,7 @@
 current_dir = os.path.dirname(os.path.abspath(__file__))
 if current_dir not in sys.path:
     sys.path.insert(0, current_dir)
-#print(f"Current directory: {current_dir}")
-#print(f"Files in current directory: {os.listdir(current_dir)}")
+
 try:
     from .omost import omost_function
     print("Successfully imported omost_function from omost.py in the current directory")

__init__.py

@@ -208,10 +208,25 @@ async def generate_image(
 
     async with aiohttp.ClientSession() as session:
         async with session.post(api_url, headers=headers, json=payload) as response:
-            response.raise_for_status()
+            if response.status != 200:
+                error_text = await response.text()
+                logger.error(f"OpenAI generate_image error {response.status}: {error_text}")
+                response.raise_for_status()
             data = await response.json()
-            images = [item["b64_json"] for item in data.get("data", [])]
-            return images
+            
+            # Handle different response structures
+            if "data" in data and isinstance(data["data"], list):
+                images = []
+                for item in data["data"]:
+                    if "b64_json" in item:
+                        images.append(item["b64_json"])
+                    elif "url" in item:
+                        # If response_format was changed or API returned URLs
+                        images.append(item["url"])
+                return images
+            else:
+                logger.error("Unexpected response format in generate_image")
+                raise ValueError("Unexpected response format")
 
 async def edit_image(
     image_base64: str,

openai_api.py

@@ -1,7 +1,7 @@
 [project]
 name = "comfyui-if_ai_tools"
 description = "Run Local and API LLMs, Features OCR-RAG (Bialdy), nanoGraphRAG, Supervision Object Detection, Conditioning manipulation via Omost, supports Ollama, LlamaCPP LMstudio, Koboldcpp, TextGen, Transformers or via APIs Anthropic, Groq, OpenAI, Google Gemini, Mistral, xAI and create your own charcters assistants (SystemPrompts) with custom presets and muchmore"
-version = "1.0.7"
+version = "1.0.6"
 license = { file = "LICENSE.txt" } 
 dependencies = ["anthropic",
 "groq",

pyproject.toml

@@ -20,7 +20,7 @@
 from .vllm_api import send_vllm_request
 from .gemini_api import send_gemini_request
 from .transformers_api import TransformersModelManager  # Import the manager
-from .utils import format_images_for_provider, convert_images_for_api, format_response
+from .utils import  convert_images_for_api, format_response
 # Set up logging
 logging.basicConfig(level=logging.DEBUG, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
 logger = logging.getLogger(__name__)    
@@ -101,8 +101,6 @@ async def send_request(
         Union[str, Dict[str, Any]]: Unified response format.
     """
     try:
-        #formatted_images = format_images_for_provider(images, llm_provider) if images is not None else None
-        #formatted_mask = format_images_for_provider(mask, llm_provider) if mask is not None else None
         # Define aspect ratio to size mapping
         aspect_ratio_mapping = {
             "1:1": "1024x1024",
@@ -141,7 +139,7 @@ async def send_request(
         else:
             # For other providers, convert to base64 only if images exist
             formatted_images = convert_images_for_api(images, target_format='base64') if images is not None and len(images) > 0 else None
-            formatted_mask = convert_images_for_api(mask, target_format='base64') if mask is not None and len(mask) > 0 else None
+            #formatted_masks = convert_images_for_api(mask, target_format='base64') if mask is not None and len(mask) > 0 else None
 
             api_functions = {
                 "groq": send_groq_request,
@@ -233,39 +231,60 @@ async def send_request(
                     }      
                 elif llm_provider == "openai":
                     if llm_model.startswith("dall-e"):
-                        if strategy == "create":
-                            # Generate image
-                            generated_images = await generate_image(
-                                prompt=user_message,
-                                model=llm_model,
-                                n=batch_count,
-                                size=size,
-                                api_key=llm_api_key
-                            )
-                            return {"images": generated_images}
-                        elif strategy == "edit":
+                        try:
+                            # Handle image formatting for edit/variations
+                            formatted_image = None
+                            formatted_mask = None
+                            
+                            if images is not None and (strategy == "edit" or strategy == "variations"):
+                                # Convert to base64 and take first image only
+                                formatted_images = convert_images_for_api(images[0:1], target_format='base64')
+                                if formatted_images:
+                                    formatted_image = formatted_images[0]
 
-                            # Edit image
-                            edited_images = await edit_image(
-                                image_base64=formatted_images[0],
-                                mask_base64=formatted_mask,
-                                prompt=user_message,
-                                model=llm_model,
-                                n=batch_count,
-                                size=size,
-                                api_key=llm_api_key
-                            )
-                            return {"images": edited_images}
-                        elif strategy == "variations":
-                            # Generate variations
-                            variations_images = await generate_image_variations(
-                                image_base64=formatted_images[0],
-                                model=llm_model,
-                                n=batch_count,
-                                size=size,
-                                api_key=llm_api_key
-                            )
-                            return {"images": variations_images}
+                            # Handle mask for edit strategy
+                            if strategy == "edit" and mask is not None:
+                                formatted_masks = convert_images_for_api(mask[0:1], target_format='base64')
+                                if formatted_masks:
+                                    formatted_mask = formatted_masks[0]
+
+                            # Make appropriate API call based on strategy
+                            if strategy == "create":
+                                response = await generate_image(
+                                    prompt=user_message,
+                                    model=llm_model,
+                                    n=batch_count,
+                                    size=size,
+                                    api_key=llm_api_key
+                                )
+                            elif strategy == "edit":
+                                response = await edit_image(
+                                    image_base64=formatted_image,
+                                    mask_base64=formatted_mask,
+                                    prompt=user_message,
+                                    model=llm_model,
+                                    n=batch_count,
+                                    size=size,
+                                    api_key=llm_api_key
+                                )
+                            elif strategy == "variations":
+                                response = await generate_image_variations(
+                                    image_base64=formatted_image,
+                                    model=llm_model,
+                                    n=batch_count,
+                                    size=size,
+                                    api_key=llm_api_key
+                                )
+                            else:
+                                raise ValueError(f"Invalid strategy: {strategy}")
+
+                            # Return the response directly - it will be a list of base64 strings
+                            return {"images": response}
+                                
+                        except Exception as e:
+                            error_msg = f"Error in DALL·E {strategy}: {str(e)}"
+                            logger.error(error_msg)
+                            return {"error": error_msg}
                     else:
                         api_url = f"https://api.openai.com/v1/chat/completions"
                         kwargs = {