Update pattern_analyzer.py

pattern_analyzer.py

@@ -2,29 +2,30 @@ import os
 os.environ['HF_HOME'] = '/tmp/huggingface'
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
+import json
+import pandas as pd
+from pattern_logic import PatternLogic
 
 class PatternAnalyzer:
     def __init__(self):
-        # Check if CUDA is available
-        if torch.cuda.is_available():
-            self.model = AutoModelForCausalLM.from_pretrained(
-                "tmmdev/codellama-pattern-analysis",
-                load_in_8bit=True,
-                device_map="auto",
-                torch_dtype="auto"
-            )
-        else:
-            # CPU fallback configuration
-            self.model = AutoModelForCausalLM.from_pretrained(
-                "tmmdev/codellama-pattern-analysis",
-                device_map="auto",
-                torch_dtype=torch.float32,
-                low_cpu_mem_usage=True
-            )
+        model_kwargs = {
+            "device_map": "auto",
+            "torch_dtype": torch.float32,
+            "low_cpu_mem_usage": True,
+            "max_memory": {"cpu": "4GB"},
+            "offload_folder": "/tmp/offload"
+        }
+        
+        self.model = AutoModelForCausalLM.from_pretrained(
+            "tmmdev/codellama-pattern-analysis",
+            **model_kwargs
+        )
+        
+        self.tokenizer = AutoTokenizer.from_pretrained(
+            "tmmdev/codellama-pattern-analysis",
+            use_fast=True
+        )
         
-        self.tokenizer = AutoTokenizer.from_pretrained("tmmdev/codellama-pattern-analysis")
-
-
         self.basic_patterns = {
             'channel': {'min_points': 4, 'confidence_threshold': 0.7},
             'triangle': {'min_points': 3, 'confidence_threshold': 0.75},
@@ -36,7 +37,7 @@ class PatternAnalyzer:
         self.pattern_logic = PatternLogic()
 
     def analyze_data(self, ohlcv_data):
-        data_prompt = f"""TASK: Identify high-confidence technical patterns only. 
+        data_prompt = f"""TASK: Identify high-confidence technical patterns only.
         Minimum confidence threshold: 0.8
         Required pattern criteria:
         1. Channel: Must have at least 3 touching points
@@ -54,7 +55,7 @@ class PatternAnalyzer:
         analysis = self.tokenizer.decode(outputs[0])
 
         return self.parse_analysis(analysis)
-        
+
     def parse_analysis(self, analysis_text):
         try:
             json_start = analysis_text.find('{')
@@ -66,7 +67,6 @@ class PatternAnalyzer:
             
             for pattern in analysis_data.get('patterns', []):
                 pattern_type = pattern.get('type')
-                
                 if pattern_type in self.basic_patterns:
                     threshold = self.basic_patterns[pattern_type]['confidence_threshold']
                     if pattern.get('confidence', 0) >= threshold: