refactor(llm): route all tiers via OpenRouter, drop Anthropic SDK

Tutti i tier (S/A/B/C/D) ora passano per OpenRouter via OpenAI SDK.
Modelli Anthropic raggiungibili via prefisso `anthropic/...`.

- pyproject: rimosso `anthropic>=0.39` da deps + uv.lock
- config: rimosso `anthropic_api_key` field
- LLMClient: dispatch unico, single client OpenAI con base_url OpenRouter
- defaults S/A/B → `anthropic/claude-{opus-4-7,sonnet-4-6}`
- retry exceptions: solo openai.* (drop anthropic.*)
- test rinominati e adattati: tier S/A/B mockano OpenAI con prefisso `anthropic/`
- rimosso test `tier_S_without_anthropic_key_raises` (non più rilevante)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/multi_swarm/config.py

@@ -24,11 +24,10 @@ class Settings(BaseSettings):
     cerbero_bot_tag: str = "swarm-poc-phase1"
 
     openrouter_api_key: SecretStr
-    anthropic_api_key: SecretStr | None = None
 
-    llm_model_tier_s: str = "claude-opus-4-7"
-    llm_model_tier_a: str = "claude-sonnet-4-6"
-    llm_model_tier_b: str = "claude-sonnet-4-6"
+    llm_model_tier_s: str = "anthropic/claude-opus-4-7"
+    llm_model_tier_a: str = "anthropic/claude-sonnet-4-6"
+    llm_model_tier_b: str = "anthropic/claude-sonnet-4-6"
     llm_model_tier_c: str = "qwen/qwen-2.5-72b-instruct"
     llm_model_tier_d: str = "meta-llama/llama-3.3-70b-instruct"
     openrouter_base_url: str = "https://openrouter.ai/api/v1"

src/multi_swarm/llm/client.py

@@ -2,9 +2,7 @@ from __future__ import annotations
 
 from dataclasses import dataclass
 
-import anthropic
 import openai
-from anthropic import Anthropic
 from openai import OpenAI
 from tenacity import (
     retry,
@@ -15,12 +13,12 @@ from tenacity import (
 
 from ..genome.hypothesis import HypothesisAgentGenome, ModelTier
 
-# Modelli configurati per Phase 1
-MODEL_TIER_S = "claude-opus-4-7"  # via Anthropic
-MODEL_TIER_A = "claude-sonnet-4-6"  # via Anthropic (premium override)
-MODEL_TIER_B = "claude-sonnet-4-6"  # via Anthropic
-MODEL_TIER_C = "qwen/qwen-2.5-72b-instruct"  # via OpenRouter
-MODEL_TIER_D = "meta-llama/llama-3.3-70b-instruct"  # via OpenRouter
+# Modelli configurati per Phase 1 — tutti via OpenRouter
+MODEL_TIER_S = "anthropic/claude-opus-4-7"
+MODEL_TIER_A = "anthropic/claude-sonnet-4-6"
+MODEL_TIER_B = "anthropic/claude-sonnet-4-6"
+MODEL_TIER_C = "qwen/qwen-2.5-72b-instruct"
+MODEL_TIER_D = "meta-llama/llama-3.3-70b-instruct"
 OPENROUTER_BASE_URL = "https://openrouter.ai/api/v1"
 
 # Errori transient: retry. RateLimit/Auth/InvalidRequest: NO retry.
@@ -28,9 +26,6 @@ _RETRYABLE_EXCEPTIONS: tuple[type[BaseException], ...] = (
     openai.APIConnectionError,
     openai.APITimeoutError,
     openai.InternalServerError,
-    anthropic.APIConnectionError,
-    anthropic.APITimeoutError,
-    anthropic.InternalServerError,
 )
 
 
@@ -44,13 +39,9 @@ class CompletionResult:
 
 
 class LLMClient:
-    _ANTHROPIC_TIERS: tuple[ModelTier, ...] = (ModelTier.S, ModelTier.A, ModelTier.B)
-    _OPENROUTER_TIERS: tuple[ModelTier, ...] = (ModelTier.C, ModelTier.D)
-
     def __init__(
         self,
         openrouter_api_key: str,
-        anthropic_api_key: str | None = None,
         model_tier_s: str = MODEL_TIER_S,
         model_tier_a: str = MODEL_TIER_A,
         model_tier_b: str = MODEL_TIER_B,
@@ -71,8 +62,7 @@ class LLMClient:
             ModelTier.C: model_tier_c,
             ModelTier.D: model_tier_d,
         }
-        self._openrouter = OpenAI(api_key=openrouter_api_key, base_url=openrouter_base_url)
-        self._anthropic = Anthropic(api_key=anthropic_api_key) if anthropic_api_key else None
+        self._client = OpenAI(api_key=openrouter_api_key, base_url=openrouter_base_url)
 
     @retry(
         stop=stop_after_attempt(3),
@@ -88,19 +78,7 @@ class LLMClient:
         max_tokens: int = 2000,
     ) -> CompletionResult:
         model = self._tier_models[genome.model_tier]
-        if genome.model_tier in self._ANTHROPIC_TIERS:
-            return self._call_anthropic(genome, system, user, max_tokens, model)
-        return self._call_openrouter(genome, system, user, max_tokens, model)
-
-    def _call_openrouter(
-        self,
-        genome: HypothesisAgentGenome,
-        system: str,
-        user: str,
-        max_tokens: int,
-        model: str,
-    ) -> CompletionResult:
-        resp = self._openrouter.chat.completions.create(
+        resp = self._client.chat.completions.create(
             model=model,
             messages=[
                 {"role": "system", "content": system},
@@ -119,32 +97,3 @@ class LLMClient:
             tier=genome.model_tier,
             model=model,
         )
-
-    def _call_anthropic(
-        self,
-        genome: HypothesisAgentGenome,
-        system: str,
-        user: str,
-        max_tokens: int,
-        model: str,
-    ) -> CompletionResult:
-        if self._anthropic is None:
-            raise RuntimeError(
-                f"ANTHROPIC_API_KEY required for tier {genome.model_tier.value} genomes"
-            )
-        msg = self._anthropic.messages.create(
-            model=model,
-            system=system,
-            messages=[{"role": "user", "content": user}],
-            temperature=genome.temperature,
-            top_p=genome.top_p,
-            max_tokens=max_tokens,
-        )
-        text = "".join(block.text for block in msg.content if hasattr(block, "text"))
-        return CompletionResult(
-            text=text,
-            input_tokens=msg.usage.input_tokens,
-            output_tokens=msg.usage.output_tokens,
-            tier=genome.model_tier,
-            model=model,
-        )