langroid 0.53.16__py3-none-any.whl → 0.54.1__py3-none-any.whl

langroid/agent/task.py

@@ -104,6 +104,10 @@ class TaskConfig(BaseModel):
             to use string-based signaling, and it is recommended to use the
             new Orchestration tools instead (see agent/tools/orchestration.py),
             e.g. DoneTool, SendTool, etc.
+        done_if_tool (bool): whether to consider the task done if the pending message
+            contains a Tool attempt by the LLM
+            (including tools not handled by the agent).
+            Default is False.
 
     """
 
@@ -115,6 +119,7 @@ class TaskConfig(BaseModel):
     addressing_prefix: str = ""
     allow_subtask_multi_oai_tools: bool = True
     recognize_string_signals: bool = True
+    done_if_tool: bool = False
 
 
 class Task:
@@ -1828,6 +1833,14 @@ class Task:
         if self._is_kill():
             return (True, StatusCode.KILL)
         result = result or self.pending_message
+
+        # Check if task should be done if message contains a tool
+        if self.config.done_if_tool and result is not None:
+            if isinstance(result, ChatDocument) and self.agent.try_get_tool_messages(
+                result, all_tools=True
+            ):
+                return (True, StatusCode.DONE)
+
         allow_done_string = self.config.recognize_string_signals
         # An entity decided task is done, either via DoneTool,
         # or by explicitly saying DONE

langroid/agent/tools/mcp/fastmcp_client.py

@@ -277,8 +277,22 @@ class FastMCPClient:
         result: CallToolResult,
     ) -> List[str] | str | None:
         if result.isError:
-            self.logger.error(f"Error calling MCP tool {tool_name}")
-            return None
+            # Log more detailed error information
+            error_content = None
+            if result.content and len(result.content) > 0:
+                try:
+                    error_content = [
+                        item.text if hasattr(item, "text") else str(item)
+                        for item in result.content
+                    ]
+                except Exception as e:
+                    error_content = [f"Could not extract error content: {str(e)}"]
+
+            self.logger.error(
+                f"Error calling MCP tool {tool_name}. Details: {error_content}"
+            )
+            return f"ERROR: Tool call failed - {error_content}"
+
         has_nontext_results = any(
             not isinstance(item, TextContent) for item in result.content
         )

langroid/embedding_models/models.py

@@ -10,7 +10,7 @@ from openai import AzureOpenAI, OpenAI
 
 from langroid.embedding_models.base import EmbeddingModel, EmbeddingModelsConfig
 from langroid.exceptions import LangroidImportError
-from langroid.language_models.openai_gpt import LangDBParams
+from langroid.language_models.provider_params import LangDBParams
 from langroid.mytypes import Embeddings
 from langroid.parsing.utils import batched
 

langroid/language_models/openai_gpt.py

@@ -60,6 +60,11 @@ from langroid.language_models.prompt_formatter.hf_formatter import (
     HFFormatter,
     find_hf_formatter,
 )
+from langroid.language_models.provider_params import (
+    DUMMY_API_KEY,
+    LangDBParams,
+    PortkeyParams,
+)
 from langroid.language_models.utils import (
     async_retry_with_exponential_backoff,
     retry_with_exponential_backoff,
@@ -81,9 +86,7 @@ DEEPSEEK_BASE_URL = "https://api.deepseek.com/v1"
 OPENROUTER_BASE_URL = "https://openrouter.ai/api/v1"
 GEMINI_BASE_URL = "https://generativelanguage.googleapis.com/v1beta/openai"
 GLHF_BASE_URL = "https://glhf.chat/api/openai/v1"
-LANGDB_BASE_URL = "https://api.us-east-1.langdb.ai"
 OLLAMA_API_KEY = "ollama"
-DUMMY_API_KEY = "xxx"
 
 VLLM_API_KEY = os.environ.get("VLLM_API_KEY", DUMMY_API_KEY)
 LLAMACPP_API_KEY = os.environ.get("LLAMA_API_KEY", DUMMY_API_KEY)
@@ -183,24 +186,6 @@ def noop() -> None:
     return None
 
 
-class LangDBParams(BaseSettings):
-    """
-    Parameters specific to LangDB integration.
-    """
-
-    api_key: str = DUMMY_API_KEY
-    project_id: str = ""
-    label: Optional[str] = None
-    run_id: Optional[str] = None
-    thread_id: Optional[str] = None
-    base_url: str = LANGDB_BASE_URL
-
-    class Config:
-        # allow setting of fields via env vars,
-        # e.g. LANGDB_PROJECT_ID=1234
-        env_prefix = "LANGDB_"
-
-
 class OpenAICallParams(BaseModel):
     """
     Various params that can be sent to an OpenAI API chat-completion call.
@@ -289,6 +274,7 @@ class OpenAIGPTConfig(LLMConfig):
     formatter: str | None = None
     hf_formatter: HFFormatter | None = None
     langdb_params: LangDBParams = LangDBParams()
+    portkey_params: PortkeyParams = PortkeyParams()
     headers: Dict[str, str] = {}
 
     def __init__(self, **kwargs) -> None:  # type: ignore
@@ -535,6 +521,7 @@ class OpenAIGPT(LanguageModel):
         self.is_glhf = self.config.chat_model.startswith("glhf/")
         self.is_openrouter = self.config.chat_model.startswith("openrouter/")
         self.is_langdb = self.config.chat_model.startswith("langdb/")
+        self.is_portkey = self.config.chat_model.startswith("portkey/")
         self.is_litellm_proxy = self.config.chat_model.startswith("litellm-proxy/")
 
         if self.is_groq:
@@ -610,6 +597,26 @@ class OpenAIGPT(LanguageModel):
                         self.config.headers["x-run-id"] = params.run_id
                     if params.thread_id:
                         self.config.headers["x-thread-id"] = params.thread_id
+            elif self.is_portkey:
+                # Parse the model string and extract provider/model
+                provider, model = self.config.portkey_params.parse_model_string(
+                    self.config.chat_model
+                )
+                self.config.chat_model = model
+                if provider:
+                    self.config.portkey_params.provider = provider
+
+                # Set Portkey base URL
+                self.api_base = self.config.portkey_params.base_url + "/v1"
+
+                # Set API key - use provider's API key from env if available
+                if self.api_key == OPENAI_API_KEY:
+                    self.api_key = self.config.portkey_params.get_provider_api_key(
+                        self.config.portkey_params.provider, DUMMY_API_KEY
+                    )
+
+                # Add Portkey-specific headers
+                self.config.headers.update(self.config.portkey_params.get_headers())
 
             self.client = OpenAI(
                 api_key=self.api_key,

langroid/language_models/provider_params.py

@@ -0,0 +1,171 @@
+"""
+Provider-specific parameter configurations for various LLM providers.
+"""
+
+from typing import Any, Dict, Optional
+
+from langroid.pydantic_v1 import BaseSettings
+
+# Constants
+LANGDB_BASE_URL = "https://api.us-east-1.langdb.ai"
+PORTKEY_BASE_URL = "https://api.portkey.ai"
+DUMMY_API_KEY = "xxx"
+
+
+class LangDBParams(BaseSettings):
+    """
+    Parameters specific to LangDB integration.
+    """
+
+    api_key: str = DUMMY_API_KEY
+    project_id: str = ""
+    label: Optional[str] = None
+    run_id: Optional[str] = None
+    thread_id: Optional[str] = None
+    base_url: str = LANGDB_BASE_URL
+
+    class Config:
+        # allow setting of fields via env vars,
+        # e.g. LANGDB_PROJECT_ID=1234
+        env_prefix = "LANGDB_"
+
+
+class PortkeyParams(BaseSettings):
+    """
+    Parameters specific to Portkey integration.
+
+    Portkey is an AI gateway that provides a unified API for multiple LLM providers,
+    with features like automatic retries, fallbacks, load balancing, and observability.
+
+    Example usage:
+        # Use Portkey with Anthropic
+        config = OpenAIGPTConfig(
+            chat_model="portkey/anthropic/claude-3-sonnet-20240229",
+            portkey_params=PortkeyParams(
+                api_key="your-portkey-api-key",
+                provider="anthropic"
+            )
+        )
+    """
+
+    api_key: str = DUMMY_API_KEY  # Portkey API key
+    provider: str = ""  # Required: e.g., "openai", "anthropic", "cohere", etc.
+    virtual_key: Optional[str] = None  # Optional: virtual key for the provider
+    trace_id: Optional[str] = None  # Optional: trace ID for request tracking
+    metadata: Optional[Dict[str, Any]] = None  # Optional: metadata for logging
+    retry: Optional[Dict[str, Any]] = None  # Optional: retry configuration
+    cache: Optional[Dict[str, Any]] = None  # Optional: cache configuration
+    cache_force_refresh: Optional[bool] = None  # Optional: force cache refresh
+    user: Optional[str] = None  # Optional: user identifier
+    organization: Optional[str] = None  # Optional: organization identifier
+    custom_headers: Optional[Dict[str, str]] = None  # Optional: additional headers
+    base_url: str = PORTKEY_BASE_URL
+
+    class Config:
+        # allow setting of fields via env vars,
+        # e.g. PORTKEY_API_KEY=xxx, PORTKEY_PROVIDER=anthropic
+        env_prefix = "PORTKEY_"
+
+    def get_headers(self) -> Dict[str, str]:
+        """Generate Portkey-specific headers from parameters."""
+        import json
+        import os
+
+        headers = {}
+
+        # API key - from params or environment
+        if self.api_key and self.api_key != DUMMY_API_KEY:
+            headers["x-portkey-api-key"] = self.api_key
+        else:
+            portkey_key = os.getenv("PORTKEY_API_KEY", "")
+            if portkey_key:
+                headers["x-portkey-api-key"] = portkey_key
+
+        # Provider
+        if self.provider:
+            headers["x-portkey-provider"] = self.provider
+
+        # Virtual key
+        if self.virtual_key:
+            headers["x-portkey-virtual-key"] = self.virtual_key
+
+        # Trace ID
+        if self.trace_id:
+            headers["x-portkey-trace-id"] = self.trace_id
+
+        # Metadata
+        if self.metadata:
+            headers["x-portkey-metadata"] = json.dumps(self.metadata)
+
+        # Retry configuration
+        if self.retry:
+            headers["x-portkey-retry"] = json.dumps(self.retry)
+
+        # Cache configuration
+        if self.cache:
+            headers["x-portkey-cache"] = json.dumps(self.cache)
+
+        # Cache force refresh
+        if self.cache_force_refresh is not None:
+            headers["x-portkey-cache-force-refresh"] = str(
+                self.cache_force_refresh
+            ).lower()
+
+        # User identifier
+        if self.user:
+            headers["x-portkey-user"] = self.user
+
+        # Organization identifier
+        if self.organization:
+            headers["x-portkey-organization"] = self.organization
+
+        # Add any custom headers
+        if self.custom_headers:
+            headers.update(self.custom_headers)
+
+        return headers
+
+    def parse_model_string(self, model_string: str) -> tuple[str, str]:
+        """
+        Parse a model string like "portkey/anthropic/claude-3-sonnet"
+        and extract provider and model name.
+
+        Returns:
+            tuple: (provider, model_name)
+        """
+        parts = model_string.split("/", 2)
+        if len(parts) >= 3 and parts[0] == "portkey":
+            _, provider, model = parts
+            return provider, model
+        else:
+            # Fallback: just remove "portkey/" prefix and return empty provider
+            model = model_string.replace("portkey/", "")
+            return "", model
+
+    def get_provider_api_key(
+        self, provider: str, default_key: str = DUMMY_API_KEY
+    ) -> str:
+        """
+        Get the API key for the provider from environment variables.
+
+        Args:
+            provider: The provider name (e.g., "anthropic", "openai")
+            default_key: Default key to return if not found
+
+        Returns:
+            The API key for the provider
+        """
+        import os
+
+        # Common environment variable patterns for different providers
+        env_patterns = [
+            f"{provider.upper()}_API_KEY",
+            f"{provider.upper()}_KEY",
+        ]
+
+        for pattern in env_patterns:
+            key = os.getenv(pattern, "")
+            if key:
+                return key
+
+        return default_key

{langroid-0.53.16.dist-info → langroid-0.54.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langroid
-Version: 0.53.16
+Version: 0.54.1
 Summary: Harness LLMs with Multi-Agent Programming
 Author-email: Prasad Chalasani <pchalasani@gmail.com>
 License: MIT
@@ -254,6 +254,10 @@ and works with [practically any LLM](https://langroid.github.io/langroid/tutoria
 :fire: Read the (WIP) [overview of the langroid architecture](https://langroid.github.io/langroid/blog/2024/08/15/overview-of-langroids-multi-agent-architecture-prelim/), 
  and a [quick tour of Langroid](https://langroid.github.io/langroid/tutorials/langroid-tour/).
 
+:fire: MCP Support: Allow any LLM-Agent to leverage MCP Servers via Langroid's simple
+[MCP tool adapter](https://langroid.github.io/langroid/notes/mcp-tools/) that converts 
+the server's tools into Langroid's `ToolMessage` instances.
+
 📢 Companies are using/adapting Langroid in **production**. Here is a quote:
 
 >[Nullify](https://www.nullify.ai) uses AI Agents for secure software development. 
@@ -340,6 +344,9 @@ teacher_task.run()
 <details>
 <summary> <b>Click to expand</b></summary>
 
+- **Jun 2025:**
+  - [0.54.0](https://github.com/langroid/langroid/releases/tag/0.54.0) Portkey AI Gateway support - access 200+ models 
+    across providers through unified API with caching, retries, observability.
 - **Mar-Apr 2025:**
   - [0.53.0](https://github.com/langroid/langroid/releases/tag/0.53.0) MCP Tools Support.
   - [0.52.0](https://github.com/langroid/langroid/releases/tag/0.52.0) Multimodal support, i.e. allow PDF, image 

{langroid-0.53.16.dist-info → langroid-0.54.1.dist-info}/RECORD

@@ -8,7 +8,7 @@ langroid/agent/batch.py,sha256=wpE9RqCNDVDhAXkCB7wEqfCIEAi6qKcrhaZ-Zr9T4C0,21375
 langroid/agent/chat_agent.py,sha256=2HIYzYxkrGkRIS97ioKfIqjaW3RbX89M39LjzBobBEY,88381
 langroid/agent/chat_document.py,sha256=6O20Fp4QrquykaF2jFtwNHkvcoDte1LLwVZNk9mVH9c,18057
 langroid/agent/openai_assistant.py,sha256=JkAcs02bIrgPNVvUWVR06VCthc5-ulla2QMBzux_q6o,34340
-langroid/agent/task.py,sha256=dJ2nRKc9-ulZyzNaABMKb9PKJzWRdZTonH9TPswpdbc,90801
+langroid/agent/task.py,sha256=Ns9SoeMoAHDW7OEcPHJ22cgQKsHzWSt9UAWi9BFl4NM,91366
 langroid/agent/tool_message.py,sha256=BhjP-_TfQ2tgxuY4Yo_JHLOwwt0mJ4BwjPnREvEY4vk,14744
 langroid/agent/xml_tool_message.py,sha256=oeBKnJNoGaKdtz39XoWGMTNlVyXew2MWH5lgtYeh8wQ,15496
 langroid/agent/callbacks/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -56,13 +56,13 @@ langroid/agent/tools/segment_extract_tool.py,sha256=__srZ_VGYLVOdPrITUM8S0HpmX4q
 langroid/agent/tools/tavily_search_tool.py,sha256=soI-j0HdgVQLf09wRQScaEK4b5RpAX9C4cwOivRFWWI,1903
 langroid/agent/tools/mcp/__init__.py,sha256=DJNM0VeFnFS3pJKCyFGggT8JVjVu0rBzrGzasT1HaSM,387
 langroid/agent/tools/mcp/decorators.py,sha256=h7dterhsmvWJ8q4mp_OopmuG2DF71ty8cZwOyzdDZuk,1127
-langroid/agent/tools/mcp/fastmcp_client.py,sha256=0ohGFa47ZqMEFR0QNJbIgHC0M6WeWm-lepkW8S8oMQk,17440
+langroid/agent/tools/mcp/fastmcp_client.py,sha256=WF3MhksDH2MzwXZF8cilMhux0hUmj6Z0dDdBYQMZwRs,18008
 langroid/cachedb/__init__.py,sha256=G2KyNnk3Qkhv7OKyxTOnpsxfDycx3NY0O_wXkJlalNY,96
 langroid/cachedb/base.py,sha256=ztVjB1DtN6pLCujCWnR6xruHxwVj3XkYniRTYAKKqk0,1354
 langroid/cachedb/redis_cachedb.py,sha256=7kgnbf4b5CKsCrlL97mHWKvdvlLt8zgn7lc528jEpiE,5141
 langroid/embedding_models/__init__.py,sha256=KyYxR3jDFUCfYjSuCL86qjAmrq6mXXjOT4lFNOKVj6Y,955
 langroid/embedding_models/base.py,sha256=Ml7oA6PzQm0wZmIYn3fhF7dvZCi-amviWUwOeBegH3A,2562
-langroid/embedding_models/models.py,sha256=g6BdBon6JzCIg5hYN07y6aCcWRJqCpGwCLkvXDPEeew,20787
+langroid/embedding_models/models.py,sha256=52S7pZOXtocnAz50vL6MKBFxhJZZurdxSEWoqs7Qfow,20792
 langroid/embedding_models/remote_embeds.py,sha256=6_kjXByVbqhY9cGwl9R83ZcYC2km-nGieNNAo1McHaY,5151
 langroid/embedding_models/protoc/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 langroid/embedding_models/protoc/embeddings.proto,sha256=_O-SgFpTaylQeOTgSpxhEJ7CUw7PeCQQJLaPqpPYKJg,321
@@ -73,10 +73,10 @@ langroid/language_models/__init__.py,sha256=3aD2qC1lz8v12HX4B-dilv27gNxYdGdeu1Qv
 langroid/language_models/azure_openai.py,sha256=SW0Fp_y6HpERr9l6TtF6CYsKgKwjUf_hSL_2mhTV4wI,5034
 langroid/language_models/base.py,sha256=253xcwXZ0yxSQ1W4SR50tAPZKCDc35yyU1o35EqB9b8,28484
 langroid/language_models/config.py,sha256=9Q8wk5a7RQr8LGMT_0WkpjY8S4ywK06SalVRjXlfCiI,378
-langroid/language_models/mcp_client_lm.py,sha256=wyDvlc26E_En5u_ZNZxajCHm8KBNi4jzG-dL76QCdt4,4098
 langroid/language_models/mock_lm.py,sha256=tA9JpURznsMZ59iRhFYMmaYQzAc0D0BT-PiJIV58sAk,4079
 langroid/language_models/model_info.py,sha256=0e011vJZMi7XU9OkKT6doxlybrNJfMlP54klLDDNgFg,14939
-langroid/language_models/openai_gpt.py,sha256=F28jqTEerN32m14q3K0oc3vnvBT8J7Q9xqXGZNKUjKU,85938
+langroid/language_models/openai_gpt.py,sha256=Xyg2VHGmA3VgPIS5ppLZeeU2Aai0qMKF9ia-oIjqRNM,86616
+langroid/language_models/provider_params.py,sha256=fX25NAmYUIc1-nliMKpmTGZO6D6RpyTXtSDdZCZdb5w,5464
 langroid/language_models/utils.py,sha256=n55Oe2_V_4VNGhytvPWLYC-0tFS07RTjN83KWl-p_MI,6032
 langroid/language_models/prompt_formatter/__init__.py,sha256=2-5cdE24XoFDhifOLl8yiscohil1ogbP1ECkYdBlBsk,372
 langroid/language_models/prompt_formatter/base.py,sha256=eDS1sgRNZVnoajwV_ZIha6cba5Dt8xjgzdRbPITwx3Q,1221
@@ -133,7 +133,7 @@ langroid/vector_store/pineconedb.py,sha256=otxXZNaBKb9f_H75HTaU3lMHiaR2NUp5MqwLZ
 langroid/vector_store/postgres.py,sha256=wHPtIi2qM4fhO4pMQr95pz1ZCe7dTb2hxl4VYspGZoA,16104
 langroid/vector_store/qdrantdb.py,sha256=O6dSBoDZ0jzfeVBd7LLvsXu083xs2fxXtPa9gGX3JX4,18443
 langroid/vector_store/weaviatedb.py,sha256=Yn8pg139gOy3zkaPfoTbMXEEBCiLiYa1MU5d_3UA1K4,11847
-langroid-0.53.16.dist-info/METADATA,sha256=hGXx5kkmGfUFp2QEBTeGNGhaegoHB5z4Pwdi5WELxa0,64946
-langroid-0.53.16.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-langroid-0.53.16.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
-langroid-0.53.16.dist-info/RECORD,,
+langroid-0.54.1.dist-info/METADATA,sha256=uamTXuB82_ONKDncZCXu8Lz_Xu1VpL2DT56kaaFLbDY,65395
+langroid-0.54.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+langroid-0.54.1.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
+langroid-0.54.1.dist-info/RECORD,,

langroid/language_models/mcp_client_lm.py

@@ -1,128 +0,0 @@
-"""
-An API for an Agent in an MCP Server to use for chat-completions
-"""
-
-from typing import Awaitable, Callable, Dict, List, Optional, Union
-
-from fastmcp.server import Context
-
-import langroid.language_models as lm
-from langroid.language_models import LLMResponse
-from langroid.language_models.base import (
-    LanguageModel,
-    LLMConfig,
-    OpenAIJsonSchemaSpec,
-    OpenAIToolSpec,
-    ToolChoiceTypes,
-)
-from langroid.utils.types import to_string
-
-
-def none_fn(x: str) -> None | str:
-    return None
-
-
-class MCPClientLMConfig(LLMConfig):
-    """
-    Mock Language Model Configuration.
-
-    Attributes:
-        response_dict (Dict[str, str]): A "response rule-book", in the form of a
-            dictionary; if last msg in dialog is x,then respond with response_dict[x]
-    """
-
-    response_dict: Dict[str, str] = {}
-    response_fn: Callable[[str], None | str] = none_fn
-    response_fn_async: Optional[Callable[[str], Awaitable[Optional[str]]]] = None
-    default_response: str = "Mock response"
-
-    type: str = "mock"
-
-
-class MockLM(LanguageModel):
-
-    def __init__(self, config: MockLMConfig = MockLMConfig()):
-        super().__init__(config)
-        self.config: MockLMConfig = config
-
-    def _response(self, msg: str) -> LLMResponse:
-        # response is based on this fallback order:
-        # - response_dict
-        # - response_fn
-        # - default_response
-        mapped_response = self.config.response_dict.get(
-            msg, self.config.response_fn(msg) or self.config.default_response
-        )
-        return lm.LLMResponse(
-            message=to_string(mapped_response),
-            cached=False,
-        )
-
-    async def _response_async(self, msg: str) -> LLMResponse:
-        # response is based on this fallback order:
-        # - response_dict
-        # - response_fn_async
-        # - response_fn
-        # - default_response
-        if self.config.response_fn_async is not None:
-            response = await self.config.response_fn_async(msg)
-        else:
-            response = self.config.response_fn(msg)
-
-        mapped_response = self.config.response_dict.get(
-            msg, response or self.config.default_response
-        )
-        return lm.LLMResponse(
-            message=to_string(mapped_response),
-            cached=False,
-        )
-
-    def chat(
-        self,
-        messages: Union[str, List[lm.LLMMessage]],
-        max_tokens: int = 200,
-        tools: Optional[List[OpenAIToolSpec]] = None,
-        tool_choice: ToolChoiceTypes | Dict[str, str | Dict[str, str]] = "auto",
-        functions: Optional[List[lm.LLMFunctionSpec]] = None,
-        function_call: str | Dict[str, str] = "auto",
-        response_format: Optional[OpenAIJsonSchemaSpec] = None,
-    ) -> lm.LLMResponse:
-        """
-        Mock chat function for testing
-        """
-        last_msg = messages[-1].content if isinstance(messages, list) else messages
-        return self._response(last_msg)
-
-    async def achat(
-        self,
-        messages: Union[str, List[lm.LLMMessage]],
-        max_tokens: int = 200,
-        tools: Optional[List[OpenAIToolSpec]] = None,
-        tool_choice: ToolChoiceTypes | Dict[str, str | Dict[str, str]] = "auto",
-        functions: Optional[List[lm.LLMFunctionSpec]] = None,
-        function_call: str | Dict[str, str] = "auto",
-        response_format: Optional[OpenAIJsonSchemaSpec] = None,
-    ) -> lm.LLMResponse:
-        """
-        Mock chat function for testing
-        """
-        last_msg = messages[-1].content if isinstance(messages, list) else messages
-        return await self._response_async(last_msg)
-
-    def generate(self, prompt: str, max_tokens: int = 200) -> lm.LLMResponse:
-        """
-        Mock generate function for testing
-        """
-        return self._response(prompt)
-
-    async def agenerate(self, prompt: str, max_tokens: int = 200) -> LLMResponse:
-        """
-        Mock generate function for testing
-        """
-        return await self._response_async(prompt)
-
-    def get_stream(self) -> bool:
-        return False
-
-    def set_stream(self, stream: bool) -> bool:
-        return False