Fix critical issues: Redis host/port parsing, Ollama URL sanitization, and Hugging Face provider initialization

core/providers/huggingface.py

@@ -25,7 +25,7 @@ class HuggingFaceProvider(LLMProvider):
         if not config.hf_token:
             raise ValueError("HF_TOKEN not set - required for Hugging Face provider")
             
-        # Fixed OpenAI client instantiation without deprecated parameters
+        # Fixed OpenAI client instantiation - removed deprecated proxies parameter
         self.client = OpenAI(
             base_url=config.hf_api_url,
             api_key=config.hf_token

core/redis_client.py

@@ -82,18 +82,28 @@ class RedisClient:
     
     def _parse_host_port(self, host_string: str, default_port: int) -> tuple:
         """Parse host and port from host string"""
-        # Remove any whitespace or control characters
+        if not host_string:
+            return "localhost", default_port
+            
+        # Remove any whitespace and control characters
         host_string = host_string.strip()
+        host_string = re.sub(r'[\r\n\t\0]+', '', host_string)
         
-        # Check if port is included in host (e.g., "host:port")
-        if ':' in host_string and not host_string.startswith('['):  # IPv6 addresses start with [
-            parts = host_string.rsplit(':', 1)  # Split only on the last colon
+        # Handle case where port is included in REDIS_HOST (e.g., "host:port")
+        if ':' in host_string and not host_string.startswith('['):  # Not IPv6
+            # Check if the part after : is a valid port number
+            parts = host_string.rsplit(':', 1)
             try:
                 host = parts[0]
                 port = int(parts[1])
-                return host, port
+                # Validate that this looks like a port (0-65535)
+                if 1 <= port <= 65535:
+                    return host, port
+                else:
+                    # Invalid port, use default
+                    return host_string, default_port
             except ValueError:
-                # Port is not a valid integer
+                # Port is not a valid integer, use default
                 return host_string, default_port
         
         return host_string, default_port

utils/config.py

@@ -43,14 +43,22 @@ class Config:
         # Remove leading/trailing whitespace and control characters
         url = url.strip()
         
-        # Remove any trailing newlines or control characters
-        url = re.sub(r'[\r\n\t\0]+', '', url)
+        # Remove any newlines, carriage returns, tabs, and null bytes
+        url = re.sub(r'[\r\n\t\0\x0b\x0c]+', '', url)
         
-        # Validate URL format
+        # Remove any trailing URL encoding artifacts
+        url = url.rstrip('%0a').rstrip('%0d')
+        
+        # Validate URL format - if it looks like a URL but missing scheme
         if url and not re.match(r'^https?://', url):
-            # If no scheme, add http:// by default
-            url = 'http://' + url
-            
+            # Check if it looks like a domain
+            if re.match(r'^[a-zA-Z0-9.-]+(?:\.[a-zA-Z]{2,})+', url) or 'ngrok' in url:
+                # Assume https for ngrok and domains
+                url = 'https://' + url
+            else:
+                # Otherwise default to http
+                url = 'http://' + url
+        
         return url
     
     def _sanitize_host(self, host: str) -> str:
@@ -61,8 +69,11 @@ class Config:
         # Remove leading/trailing whitespace and control characters
         host = host.strip()
         
-        # Remove any newlines or control characters
-        host = re.sub(r'[\r\n\t\0]+', '', host)
+        # Remove any newlines, carriage returns, tabs, and null bytes
+        host = re.sub(r'[\r\n\t\0\x0b\x0c]+', '', host)
+        
+        # Remove any trailing URL encoding artifacts
+        host = host.rstrip('%0a').rstrip('%0d')
         
         return host
 

utils/config_validator.py

@@ -17,9 +17,18 @@ class ConfigValidator:
             'warnings': []
         }
         
+        # Log current configuration for debugging
+        logger.info(f"Current configuration:")
+        logger.info(f"  OLLAMA_HOST: '{config.ollama_host}'")
+        logger.info(f"  REDIS_HOST: '{config.redis_host}'")
+        logger.info(f"  REDIS_PORT: {config.redis_port}")
+        
         # Validate Ollama host
         if config.ollama_host:
-            if not ConfigValidator._is_valid_url(config.ollama_host):
+            sanitized = config._sanitize_url(config.ollama_host)
+            if sanitized != config.ollama_host:
+                logger.info(f"Sanitized OLLAMA_HOST from '{config.ollama_host}' to '{sanitized}'")
+            if not ConfigValidator._is_valid_url(sanitized):
                 report['errors'].append(f"Invalid OLLAMA_HOST format: {config.ollama_host}")
                 report['valid'] = False
         else:
@@ -27,8 +36,9 @@ class ConfigValidator:
         
         # Validate Redis configuration
         if config.redis_host:
-            if not ConfigValidator._is_valid_host(config.redis_host):
-                report['warnings'].append(f"Redis host may be invalid: {config.redis_host}")
+            sanitized_host = config._sanitize_host(config.redis_host)
+            if sanitized_host != config.redis_host:
+                logger.info(f"Sanitized REDIS_HOST from '{config.redis_host}' to '{sanitized_host}'")
         else:
             report['warnings'].append("Redis not configured, using in-memory storage")