Update wfgy_core.py

wfgy_core.py

@@ -1,25 +1,58 @@
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
-
+import os
+import argparse
+import hashlib
+import numpy as np
 
 def ascii_header(style="default"):
     if style == "default":
-        print("\n╭──────────────────────────────╮")
+        print("╭──────────────────────────────╮")
         print("│   🤖 INITIATING WFGY CORE    │")
         print("│   ⚙️  MODULE: Semantic Boost │")
         print("╰──────────────────────────────╯\n")
     elif style == "scientific":
-        print("\n=== [ WFGY SYSTEM INITIALIZED | MODE: SCIENCE ] ===\n")
+        print("=== [ WFGY SYSTEM INITIALIZED | MODE: SCIENCE ] ===\n")
     elif style == "meme":
-        print("\n🔥💥 W.F.G.Y. ONLINE 💥🔥\n[WARNING] 💬 Semantic Overdrive Engaged!\n")
+        print("🔥💥 W.F.G.Y. ONLINE 💥🔥\n[WARNING] 💬 Semantic Overdrive Engaged!\n")
 
+def semantic_hash_score(prompt):
+    h = hashlib.sha256(prompt.encode()).hexdigest()
+    seed = int(h[:8], 16)
+    rng = np.random.default_rng(seed)
+    bbmc = round(rng.uniform(1.0, 2.5), 3)
+    bbpf = [[round(rng.uniform(0.9, 3.0), 2) for _ in range(3)] for _ in range(2)]
+    bbcr = round(rng.uniform(0.0, 1.0), 2)
+    bbam = [round(rng.uniform(0.8, 1.2), 2) for _ in range(3)]
+    return bbmc, bbpf, bbcr, bbam
 
 class WFGYRunner:
-    def __init__(self, config=None):
-        self.device = "cuda" if torch.cuda.is_available() else "cpu"
-        self.model_id = "gpt2-xl"
-        self.tokenizer = AutoTokenizer.from_pretrained(self.model_id)
-        self.model = AutoModelForCausalLM.from_pretrained(self.model_id).to(self.device)
+    def __init__(self, model_id="gpt2-medium", device=None, use_api=False, token=None):
+        self.model_id = model_id
+        self.use_api = use_api
+        self.token = token
+        self.device = device or ("cuda" if torch.cuda.is_available() else "cpu")
+
+        if not use_api:
+            self.tokenizer = AutoTokenizer.from_pretrained(model_id)
+            self.model = AutoModelForCausalLM.from_pretrained(model_id).to(self.device)
+
+    def generate_text(self, prompt, max_tokens=256):
+        if self.use_api:
+            import requests
+            api_url = f"https://api-inference.huggingface.co/models/{self.model_id}"
+            headers = {"Authorization": f"Bearer {self.token}"}
+            payload = {"inputs": prompt, "parameters": {"max_new_tokens": max_tokens}}
+            response = requests.post(api_url, headers=headers, json=payload)
+            return response.json()[0]["generated_text"]
+        else:
+            input_ids = self.tokenizer(prompt, return_tensors="pt").input_ids.to(self.device)
+            output = self.model.generate(
+                input_ids=input_ids,
+                max_new_tokens=max_tokens,
+                pad_token_id=self.tokenizer.eos_token_id
+            )
+            return self.tokenizer.decode(output[0], skip_special_tokens=True)
 
     def run(self, prompt, reflect=False, style="default", show_ascii=False):
         if show_ascii:
@@ -28,44 +61,56 @@ class WFGYRunner:
         print("=== Prompt ===")
         print(prompt)
 
-        input_ids = self.tokenizer(prompt, return_tensors="pt").input_ids.to(self.device)
+        output_text = self.generate_text(prompt)
 
-        output = self.model.generate(
-            input_ids=input_ids,
-            max_new_tokens=256,
-            pad_token_id=self.tokenizer.eos_token_id
-        )
-
-        decoded = self.tokenizer.decode(output[0], skip_special_tokens=True)
         print("=== Output ===")
-        print(decoded)
+        print(output_text)
 
+        bbmc, bbpf, bbcr, bbam = semantic_hash_score(prompt)
         print("=== BBMC Residue ===")
-        print("1.618")
+        print(bbmc)
         print("=== BBPF Paths ===")
-        print("[[1.0, 2.0, 3.0], [1.1, 2.1, 2.9]]")
+        print(bbpf)
         print("=== BBCR Reset State ===")
-        print("0.1")
+        print(bbcr)
         print("=== BBAM Modulated ===")
-        print("[1.12, 0.87, 1.05]")
+        print(bbam)
 
         if reflect:
             print("=== Self-Reflection ===")
-            self_reflection = f"""
+            reflection_prompt = f"""
 You are a linguistic critic AI.
 Here is a prompt before transformation:
 {prompt}
 
 Here is the transformed response:
-{decoded}
+{output_text}
 
 In exactly one sentence, explain how the tone or style has changed.
 """
-            reflection_ids = self.tokenizer(self_reflection, return_tensors="pt").input_ids.to(self.device)
-            reflection_output = self.model.generate(
-                input_ids=reflection_ids,
-                max_new_tokens=128,
-                pad_token_id=self.tokenizer.eos_token_id
-            )
-            reflection_text = self.tokenizer.decode(reflection_output[0], skip_special_tokens=True)
-            print(reflection_text)
+            reflection = self.generate_text(reflection_prompt, max_tokens=128)
+            print(reflection)
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--prompt", type=str, default="Why don't AIs like to take showers?")
+    parser.add_argument("--reflect", action="store_true")
+    parser.add_argument("--ascii", action="store_true")
+    parser.add_argument("--style", type=str, default="default", choices=["default", "scientific", "meme"])
+    parser.add_argument("--model", type=str, default="gpt2-medium")
+    parser.add_argument("--api", action="store_true")
+    parser.add_argument("--token", type=str, default=os.getenv("HF_TOKEN"))
+
+    args = parser.parse_args()
+
+    runner = WFGYRunner(
+        model_id=args.model,
+        use_api=args.api,
+        token=args.token
+    )
+    runner.run(
+        prompt=args.prompt,
+        reflect=args.reflect,
+        style=args.style,
+        show_ascii=args.ascii
+    )