From 7750375cef1de9a6ef591e7d4afde0a12f57894e Mon Sep 17 00:00:00 2001
From: Fillip Isgro <fillipisgro@Fillips-MacBook-Air.local>
Date: Sun, 9 Feb 2025 16:02:57 -0500
Subject: [PATCH] added clip scanner node

---
 __init__.py             |  3 ++
 nodes/FL_ClipScanner.py | 87 +++++++++++++++++++++++++++++++++++++++++
 requirements.txt        |  1 +
 3 files changed, 91 insertions(+)
 create mode 100644 nodes/FL_ClipScanner.py

diff --git a/__init__.py b/__init__.py
index 7b117ba..3242c0c 100644
--- a/__init__.py
+++ b/__init__.py
@@ -89,6 +89,7 @@ from .nodes.FL_API_ImageSaver import FL_API_ImageSaver
 from .nodes.FL_GoogleDriveImageDownloader import FL_GoogleDriveImageDownloader
 from .nodes.FL_AnimeLineExtractor import FL_AnimeLineExtractor
 from .nodes.FL_HunyuanDelight import FL_HunyuanDelight
+from .nodes.FL_ClipScanner import FL_ClipScanner
 
 
 NODE_CLASS_MAPPINGS = {
@@ -184,6 +185,7 @@ NODE_CLASS_MAPPINGS = {
     "FL_GoogleDriveImageDownloader": FL_GoogleDriveImageDownloader,
     "FL_AnimeLineExtractor": FL_AnimeLineExtractor,
     "FL_HunyuanDelight": FL_HunyuanDelight,
+    "FL_ClipScanner": FL_ClipScanner,
 
 }
 
@@ -280,6 +282,7 @@ NODE_DISPLAY_NAME_MAPPINGS = {
     "FL_GoogleDriveImageDownloader": "FL Google Drive Image Downloader",
     "FL_AnimeLineExtractor": "FL Anime Line Extractor",
     "FL_HunyuanDelight": "FL Hunyuan Delight",
+    "FL_ClipScanner": "FL Clip Scanner (Kytra)",
 
 }
 
diff --git a/nodes/FL_ClipScanner.py b/nodes/FL_ClipScanner.py
new file mode 100644
index 0000000..046ac97
--- /dev/null
+++ b/nodes/FL_ClipScanner.py
@@ -0,0 +1,87 @@
+import open_clip
+
+class FL_ClipScanner:
+    def __init__(self):
+        self.current_model = None
+        self.current_pretrained = None
+        self.tokenizer = None
+        
+        # Model configurations
+        self.model_configs = {
+            "SDXL (ViT-G/14)": {
+                "model": "ViT-g-14",
+                "pretrained": "laion2b_s12b_b42k"
+            },
+            "SD 1.5 (ViT-L/14)": {
+                "model": "ViT-L-14",
+                "pretrained": "openai"
+            },
+            "FLUX (ViT-L/14)": {
+                "model": "ViT-L-14",
+                "pretrained": "openai"
+            }
+        }
+
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "model_type": (list(cls().model_configs.keys()),),
+                "text": ("STRING", {"default": "", "multiline": True}),
+            },
+        }
+
+    RETURN_TYPES = ("STRING",)
+    FUNCTION = "analyze_tokens"
+    CATEGORY = "🏵️Fill Nodes/Analysis"
+    OUTPUT_NODE = True
+
+    def initialize_model(self, model_type):
+        if (self.current_model != self.model_configs[model_type]["model"] or
+            self.current_pretrained != self.model_configs[model_type]["pretrained"]):
+            try:
+                config = self.model_configs[model_type]
+                self.tokenizer = open_clip.get_tokenizer(config["model"])
+                self.current_model = config["model"]
+                self.current_pretrained = config["pretrained"]
+                return True
+            except Exception as e:
+                print(f"Error initializing CLIP tokenizer: {str(e)}")
+                self.tokenizer = None
+                return False
+        return True
+
+    def analyze_tokens(self, model_type: str, text: str) -> tuple[str]:
+        if not self.initialize_model(model_type):
+            return ("Error: Failed to initialize CLIP tokenizer.",)
+
+        try:
+            # Tokenize the input text
+            tokens = self.tokenizer(text)
+            
+            # Convert token IDs to words
+            decoded_tokens = [self.tokenizer.decoder.get(tok, f"[{tok}]")
+                            for tok in tokens[0].tolist()]
+            
+            # Remove start, end, and padding tokens
+            filtered_tokens = [t for t in decoded_tokens
+                             if not t.startswith("<") and t != "!"]
+            
+            # Create formatted output
+            output = []
+            output.append("=" * 60)
+            output.append(f"Model: {model_type}")
+            output.append(f"CLIP: {self.current_model} ({self.current_pretrained})")
+            output.append(f"Prompt: \"{text}\"")
+            output.append(f"Tokenized Output: {filtered_tokens}")
+            output.append(f"Token Count: {len(filtered_tokens)}")
+            output.append("=" * 60)
+            
+            return ("\n".join(output),)
+
+        except Exception as e:
+            return (f"Error analyzing tokens: {str(e)}",)
+
+    @classmethod
+    def IS_CHANGED(cls, model_type, text):
+        return float("NaN")
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index 28acfca..869d4f5 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -19,3 +19,4 @@ ollama
 kornia
 opencv-python
 gdown
+open_clip_torch
\ No newline at end of file