code-puppy 0.0.183__tar.gz → 0.0.185__tar.gz

{code_puppy-0.0.183 → code_puppy-0.0.185}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code-puppy
-Version: 0.0.183
+Version: 0.0.185
 Summary: Code generation agent
 Project-URL: repository, https://github.com/mpfaffenberger/code_puppy
 Project-URL: HomePage, https://github.com/mpfaffenberger/code_puppy

{code_puppy-0.0.183 → code_puppy-0.0.185}/code_puppy/agents/base_agent.py

@@ -729,6 +729,63 @@ class BaseAgent(ABC):
         manager = get_mcp_manager()
         return manager.get_servers_for_agent()
 
+    def _load_model_with_fallback(
+        self,
+        requested_model_name: str,
+        models_config: Dict[str, Any],
+        message_group: str,
+    ) -> Tuple[Any, str]:
+        """Load the requested model, applying a friendly fallback when unavailable."""
+        try:
+            model = ModelFactory.get_model(requested_model_name, models_config)
+            return model, requested_model_name
+        except ValueError as exc:
+            available_models = list(models_config.keys())
+            available_str = (
+                ", ".join(sorted(available_models))
+                if available_models
+                else "no configured models"
+            )
+            emit_warning(
+                (
+                    f"[yellow]Model '{requested_model_name}' not found. "
+                    f"Available models: {available_str}[/yellow]"
+                ),
+                message_group=message_group,
+            )
+
+            fallback_candidates: List[str] = []
+            global_candidate = get_global_model_name()
+            if global_candidate:
+                fallback_candidates.append(global_candidate)
+
+            for candidate in available_models:
+                if candidate not in fallback_candidates:
+                    fallback_candidates.append(candidate)
+
+            for candidate in fallback_candidates:
+                if not candidate or candidate == requested_model_name:
+                    continue
+                try:
+                    model = ModelFactory.get_model(candidate, models_config)
+                    emit_info(
+                        f"[bold cyan]Using fallback model: {candidate}[/bold cyan]",
+                        message_group=message_group,
+                    )
+                    return model, candidate
+                except ValueError:
+                    continue
+
+            friendly_message = (
+                "No valid model could be loaded. Update the model configuration or set "
+                "a valid model with `config set`."
+            )
+            emit_error(
+                f"[bold red]{friendly_message}[/bold red]",
+                message_group=message_group,
+            )
+            raise ValueError(friendly_message) from exc
+
     def reload_code_generation_agent(self, message_group: Optional[str] = None):
         """Force-reload the pydantic-ai Agent based on current config and model."""
         from code_puppy.tools import register_tools_for_agent
@@ -743,7 +800,11 @@ class BaseAgent(ABC):
             message_group=message_group,
         )
         models_config = ModelFactory.load_config()
-        model = ModelFactory.get_model(model_name, models_config)
+        model, resolved_model_name = self._load_model_with_fallback(
+            model_name,
+            models_config,
+            message_group,
+        )
 
         emit_info(
             f"[bold magenta]Loading Agent: {self.name}[/bold magenta]",
@@ -786,7 +847,7 @@ class BaseAgent(ABC):
         register_tools_for_agent(p_agent, agent_tools)
 
         self._code_generation_agent = p_agent
-        self._last_model_name = model_name
+        self._last_model_name = resolved_model_name
         # expose for run_with_mcp
         self.pydantic_agent = p_agent
         return self._code_generation_agent

{code_puppy-0.0.183 → code_puppy-0.0.185}/code_puppy/model_factory.py

@@ -31,6 +31,12 @@ from .round_robin_model import RoundRobinModel
 #   Example: "X-Api-Key": "$OPENAI_API_KEY" will use the value from os.environ.get("OPENAI_API_KEY")
 
 
+class ZaiChatModel(OpenAIChatModel):
+    def _process_response(self, response):
+        response.object = 'chat.completion'
+        return super()._process_response(response)
+
+
 def get_custom_config(model_config):
     custom_config = model_config.get("custom_endpoint", {})
     if not custom_config:
@@ -51,8 +57,23 @@ def get_custom_config(model_config):
                     f"Please set the environment variable: export {env_var_name}=your_value"
                 )
             value = resolved_value
+        elif "$" in value:
+            tokens = value.split(" ")
+            resolved_values = []
+            for token in tokens:
+                if token.startswith("$"):
+                    env_var = token[1:]
+                    resolved_value = os.environ.get(env_var)
+                    if resolved_value is None:
+                        raise ValueError(
+                            f"Environment variable '{env_var}' is required for custom endpoint headers but is not set. "
+                            f"Please set the environment variable: export {env_var}=your_value"
+                        )
+                    resolved_values.append(resolved_value)
+                else:
+                    resolved_values.append(token)
+            value = " ".join(resolved_values)
         headers[key] = value
-
     api_key = None
     if "api_key" in custom_config:
         if custom_config["api_key"].startswith("$"):
@@ -223,7 +244,24 @@ class ModelFactory:
             model = OpenAIChatModel(model_name=model_config["name"], provider=provider)
             setattr(model, "provider", provider)
             return model
-
+        elif model_type == "zai_coding":
+            zai_model = ZaiChatModel(
+                model_name=model_config["name"],
+                provider=OpenAIProvider(
+                    api_key=os.getenv('ZAI_API_KEY'),
+                    base_url='https://api.z.ai/api/coding/paas/v4'
+                )
+            )
+            return zai_model
+        elif model_type == "zai_api":
+            zai_model = ZaiChatModel(
+                model_name=model_config["name"],
+                provider=OpenAIProvider(
+                    api_key=os.getenv('ZAI_API_KEY'),
+                    base_url='https://api.z.ai/api/paas/v4/'
+                )
+            )
+            return zai_model
         elif model_type == "custom_gemini":
             url, headers, verify, api_key = get_custom_config(model_config)
             os.environ["GEMINI_API_KEY"] = api_key

code_puppy-0.0.185/code_puppy/models.json

@@ -0,0 +1,71 @@
+{
+  "gpt-5": {
+    "type": "openai",
+    "name": "gpt-5",
+    "context_length": 400000
+  },
+  "Cerebras-Qwen3-Coder-480b": {
+    "type": "cerebras",
+    "name": "qwen-3-coder-480b",
+    "custom_endpoint": {
+      "url": "https://api.cerebras.ai/v1",
+      "api_key": "$CEREBRAS_API_KEY"
+    },
+    "context_length": 131072
+  },
+  "Cerebras-Qwen3-235b-a22b-instruct-2507": {
+    "type": "cerebras",
+    "name": "qwen-3-235b-a22b-instruct-2507",
+    "custom_endpoint": {
+      "url": "https://api.cerebras.ai/v1",
+      "api_key": "$CEREBRAS_API_KEY"
+    },
+    "context_length": 64000
+  },
+  "Cerebras-gpt-oss-120b": {
+    "type": "cerebras",
+    "name": "gpt-oss-120b",
+    "custom_endpoint": {
+      "url": "https://api.cerebras.ai/v1",
+      "api_key": "$CEREBRAS_API_KEY"
+    },
+    "context_length": 131072
+  },
+  "Cerebras-Qwen-3-32b": {
+    "type": "cerebras",
+    "name": "qwen-3-32b",
+    "custom_endpoint": {
+      "url": "https://api.cerebras.ai/v1",
+      "api_key": "$CEREBRAS_API_KEY"
+    },
+    "context_length": 65536
+  },
+  "claude-4-0-sonnet": {
+    "type": "anthropic",
+    "name": "claude-sonnet-4-20250514",
+    "context_length": 200000
+  },
+  "claude-4-5-sonnet": {
+    "type": "anthropic",
+    "name": "claude-sonnet-4-5-20250929",
+    "context_length": 200000
+  },
+  "glm-4.5-coding": {
+    "type": "zai_coding",
+    "name": "glm-4.5"
+  },
+  "glm-4.6-coding": {
+    "type": "zai_coding",
+    "name": "glm-4.6",
+    "context_length": 200000
+  },
+  "glm-4.5": {
+    "type": "zai_api",
+    "name": "glm-4.5"
+  },
+  "glm-4.6": {
+    "type": "zai_api",
+    "name": "glm-4.6",
+    "context_length": 200000
+  }
+}

{code_puppy-0.0.183 → code_puppy-0.0.185}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "code-puppy"
-version = "0.0.183"
+version = "0.0.185"
 description = "Code generation agent"
 readme = "README.md"
 requires-python = ">=3.11"

code_puppy-0.0.183/code_puppy/models.json

@@ -1,128 +0,0 @@
-{
-  "gpt-5": {
-    "type": "openai",
-    "name": "gpt-5",
-    "context_length": 400000
-  },
-  "Cerebras-Qwen3-Coder-480b": {
-    "type": "cerebras",
-    "name": "qwen-3-coder-480b",
-    "custom_endpoint": {
-      "url": "https://api.cerebras.ai/v1",
-      "api_key": "$CEREBRAS_API_KEY"
-    },
-    "context_length": 131072
-  },
-  "Cerebras-Qwen3-235b-a22b-instruct-2507": {
-    "type": "cerebras",
-    "name": "qwen-3-235b-a22b-instruct-2507",
-    "custom_endpoint": {
-      "url": "https://api.cerebras.ai/v1",
-      "api_key": "$CEREBRAS_API_KEY"
-    },
-    "context_length": 64000
-  },
-  "Cerebras-gpt-oss-120b": {
-    "type": "cerebras",
-    "name": "gpt-oss-120b",
-    "custom_endpoint": {
-      "url": "https://api.cerebras.ai/v1",
-      "api_key": "$CEREBRAS_API_KEY"
-    },
-    "context_length": 131072
-  },
-  "Cerebras-Qwen-3-32b": {
-    "type": "cerebras",
-    "name": "qwen-3-32b",
-    "custom_endpoint": {
-      "url": "https://api.cerebras.ai/v1",
-      "api_key": "$CEREBRAS_API_KEY"
-    },
-    "context_length": 65536
-  },
-  "claude-4-0-sonnet": {
-    "type": "anthropic",
-    "name": "claude-sonnet-4-20250514",
-    "context_length": 200000
-  },
-  "o3": {
-    "type": "openai",
-    "name": "o3",
-    "context_length": 200000
-  },
-  "grok-4": {
-    "type": "custom_openai",
-    "name": "grok-4",
-    "custom_endpoint": {
-      "url": "https://api.x.ai/v1",
-      "api_key": "$XAI_API_KEY"
-    },
-    "context_length": 256000
-  },
-    "grok-code-fast-1": {
-    "type": "custom_openai",
-    "name": "grok-code-fast-1",
-    "custom_endpoint": {
-      "url": "https://api.x.ai/v1",
-      "api_key": "$XAI_API_KEY"
-    },
-    "context_length": 256000
-  },
-  "gemini-2.5-flash-preview-05-20": {
-    "type": "gemini",
-    "name": "gemini-2.5-flash-preview-05-20",
-    "context_length": 1048576
-  },
-  "gpt-4.1": {
-    "type": "openai",
-    "name": "gpt-4.1",
-    "context_length": 1000000
-  },
-  "Qwen/Qwen3-235B-A22B-fp8-tput": {
-    "type": "custom_openai",
-    "name": "Qwen/Qwen3-235B-A22B-fp8-tput",
-    "custom_endpoint": {
-      "url": "https://api.together.xyz/v1",
-      "api_key": "$TOGETHER_API_KEY"
-    },
-    "context_length": 64000
-  },
-  "azure-gpt-4.1": {
-    "type": "azure_openai",
-    "name": "gpt-4.1",
-    "api_version": "2024-12-01-preview",
-    "api_key": "$AZURE_OPENAI_API_KEY",
-    "azure_endpoint": "$AZURE_OPENAI_ENDPOINT",
-    "context_length": 128000
-  },
-   "gpt-4.1-mini": {
-    "type": "openai",
-    "name": "gpt-4.1-mini",
-    "context_length": 128000
-  },
-  "gpt-4.1-nano": {
-    "type": "openai",
-    "name": "gpt-4.1-nano",
-    "context_length": 128000
-  },
-  "gpt-4.1-custom": {
-    "type": "custom_openai",
-    "name": "gpt-4.1-custom",
-    "custom_endpoint": {
-      "url": "https://my.cute.endpoint:8080",
-      "headers": {
-        "X-Api-Key": "$OPENAI_API_KEY"
-      },
-      "ca_certs_path": "/path/to/cert.pem"
-    },
-    "context_length": 128000
-  },
-  "ollama-llama3.3": {
-    "type": "custom_openai",
-    "name": "llama3.3",
-    "custom_endpoint": {
-      "url": "http://localhost:11434/v1"
-    },
-    "context_length": 8192
-  }
-}