perf: optimize cascade router memory — deduplicate provider loop, hoist constants (#1376)

src/infrastructure/router/cascade.py

@@ -10,6 +10,8 @@ models for image inputs and falls back through capability chains.
 
 import asyncio
 import logging
+import os
+import re
 from pathlib import Path
 from typing import TYPE_CHECKING, Any
 
@@ -28,6 +30,15 @@ try:
 except ImportError:
     requests = None  # type: ignore
 
+# Pre-compiled regex for env-var expansion (avoids re-compilation per call)
+_ENV_VAR_RE = re.compile(r"\$\{(\w+)\}")
+
+# Constant tuples for content-type detection (avoids per-call allocation)
+_IMAGE_EXTENSIONS = (".jpg", ".jpeg", ".png", ".gif", ".webp", ".bmp")
+
+# Constant set for cloud provider types (avoids per-call tuple creation)
+_CLOUD_PROVIDER_TYPES = frozenset(("anthropic", "openai", "grok"))
+
 # Re-export data models so existing ``from …cascade import X`` keeps working.
 # Mixins
 from .health import HealthMixin
@@ -156,20 +167,19 @@ class CascadeRouter(HealthMixin, ProviderCallsMixin):
 
         self.providers.sort(key=lambda p: p.priority)
 
-    def _expand_env_vars(self, content: str) -> str:
+    @staticmethod
+    def _expand_env_vars(content: str) -> str:
         """Expand ${VAR} syntax in YAML content.
 
         Uses os.environ directly (not settings) because this is a generic
         YAML config loader that must expand arbitrary variable references.
         """
-        import os
-        import re
 
         def replace_var(match: "re.Match[str]") -> str:
             var_name = match.group(1)
             return os.environ.get(var_name, match.group(0))
 
-        return re.sub(r"\$\{(\w+)\}", replace_var, content)
+        return _ENV_VAR_RE.sub(replace_var, content)
 
     def _check_provider_available(self, provider: Provider) -> bool:
         """Check if a provider is actually available."""
@@ -225,8 +235,7 @@ class CascadeRouter(HealthMixin, ProviderCallsMixin):
 
             # Check for image URLs in content
             if isinstance(content, str):
-                image_extensions = (".jpg", ".jpeg", ".png", ".gif", ".webp", ".bmp")
-                if any(ext in content.lower() for ext in image_extensions):
+                if any(ext in content.lower() for ext in _IMAGE_EXTENSIONS):
                     has_image = True
                 if content.startswith("data:image/"):
                     has_image = True
@@ -395,7 +404,7 @@ class CascadeRouter(HealthMixin, ProviderCallsMixin):
             return None
 
         # Metabolic protocol: skip cloud providers when quota is low
-        if provider.type in ("anthropic", "openai", "grok"):
+        if provider.type in _CLOUD_PROVIDER_TYPES:
             if not self._quota_allows_cloud(provider):
                 logger.info(
                     "Metabolic protocol: skipping cloud provider %s (quota too low)",
@@ -513,18 +522,6 @@ class CascadeRouter(HealthMixin, ProviderCallsMixin):
         providers = self._filter_providers(cascade_tier)
 
         for provider in providers:
-            if not self._is_provider_available(provider):
-                continue
-
-            # Metabolic protocol: skip cloud providers when quota is low
-            if provider.type in ("anthropic", "openai", "grok"):
-                if not self._quota_allows_cloud(provider):
-                    logger.info(
-                        "Metabolic protocol: skipping cloud provider %s (quota too low)",
-                        provider.name,
-                    )
-                    continue
-
             # Complexity-based model selection (only when no explicit model)
             effective_model = model
             if effective_model is None and complexity is not None:
@@ -537,33 +534,13 @@ class CascadeRouter(HealthMixin, ProviderCallsMixin):
                         effective_model,
                     )
 
-            selected_model, is_fallback_model = self._select_model(
-                provider, effective_model, content_type
+            result = await self._try_single_provider(
+                provider, messages, effective_model, temperature,
+                max_tokens, content_type, errors,
             )
-
-            try:
-                result = await self._attempt_with_retry(
-                    provider,
-                    messages,
-                    selected_model,
-                    temperature,
-                    max_tokens,
-                    content_type,
-                )
-            except RuntimeError as exc:
-                errors.append(str(exc))
-                self._record_failure(provider)
-                continue
-
-            self._record_success(provider, result.get("latency_ms", 0))
-            return {
-                "content": result["content"],
-                "provider": provider.name,
-                "model": result.get("model", selected_model or provider.get_default_model()),
-                "latency_ms": result.get("latency_ms", 0),
-                "is_fallback_model": is_fallback_model,
-                "complexity": complexity.value if complexity is not None else None,
-            }
+            if result is not None:
+                result["complexity"] = complexity.value if complexity is not None else None
+                return result
 
         raise RuntimeError(f"All providers failed: {'; '.join(errors)}")