lm-deluge 0.0.56__py3-none-any.whl → 0.0.69__py3-none-any.whl

lm_deluge/__init__.py

@@ -1,7 +1,14 @@
 from .client import APIResponse, LLMClient, SamplingParams
 from .file import File
 from .prompt import Conversation, Message
-from .tool import Tool
+from .tool import Tool, ToolParams
+
+try:
+    from .mock_openai import MockAsyncOpenAI  # noqa
+
+    _has_openai = True
+except ImportError:
+    _has_openai = False
 
 # dotenv.load_dotenv() - don't do this, fucks with other packages
 
@@ -12,5 +19,9 @@ __all__ = [
     "Conversation",
     "Message",
     "Tool",
+    "ToolParams",
     "File",
 ]
+
+if _has_openai:
+    __all__.append("MockAsyncOpenAI")

lm_deluge/api_requests/anthropic.py

@@ -42,6 +42,14 @@ def _build_anthropic_request(
         "content-type": "application/json",
     }
 
+    # Check if any messages contain uploaded files (file_id)
+    # If so, add the files-api beta header
+    for msg in prompt.messages:
+        for file in msg.files:
+            if file.is_remote and file.remote_provider == "anthropic":
+                _add_beta(base_headers, "files-api-2025-04-14")
+                break
+
     request_json = {
         "model": model.name,
         "messages": messages,
@@ -72,7 +80,7 @@ def _build_anthropic_request(
         request_json["system"] = system_message
 
     # handle temp + top_p for opus 4.1/sonnet 4.5
-    if model.name in ["claude-sonnet-4-5-20250929", "claude-opus-4-1-20250805"]:
+    if "4-1" in model.name or "4-5" in model.name:
         if "temperature" in request_json and "top_p" in request_json:
             request_json.pop("top_p")
 
@@ -82,6 +90,9 @@ def _build_anthropic_request(
         for tool in tools:
             if isinstance(tool, Tool):
                 tool_definitions.append(tool.dump_for("anthropic"))
+            elif isinstance(tool, dict) and "url" in tool:
+                _add_beta(base_headers, "mcp-client-2025-04-04")
+                mcp_servers.append(tool)
             elif isinstance(tool, dict):
                 tool_definitions.append(tool)
                 # add betas if needed

lm_deluge/api_requests/base.py

@@ -1,4 +1,5 @@
 import asyncio
+import time
 import traceback
 from abc import ABC, abstractmethod
 
@@ -6,6 +7,7 @@ import aiohttp
 from aiohttp import ClientResponse
 
 from ..errors import raise_if_modal_exception
+from ..models.openai import OPENAI_MODELS
 from ..request_context import RequestContext
 from .response import APIResponse
 
@@ -82,15 +84,95 @@ class APIRequestBase(ABC):
         if self.context.status_tracker:
             self.context.status_tracker.task_succeeded(self.context.task_id)
 
+    async def _execute_once_background_mode(self) -> APIResponse:
+        """
+        ONLY for OpenAI responses API. Implement the
+        start -> poll -> result style of request.
+        """
+        assert self.context.status_tracker, "no status tracker"
+        start_time = time.time()
+        async with aiohttp.ClientSession() as session:
+            last_status: str | None = None
+
+            try:
+                self.context.status_tracker.total_requests += 1
+                assert self.url is not None, "URL is not set"
+                async with session.post(
+                    url=self.url,
+                    headers=self.request_header,
+                    json=self.request_json,
+                ) as http_response:
+                    # make sure we created the Response object
+                    http_response.raise_for_status()
+                    data = await http_response.json()
+                    response_id = data["id"]
+                    last_status = data["status"]
+
+                while True:
+                    if time.time() - start_time > self.context.request_timeout:
+                        # cancel the response
+                        async with session.post(
+                            url=f"{self.url}/{response_id}/cancel",
+                            headers=self.request_header,
+                        ) as http_response:
+                            http_response.raise_for_status()
+
+                        return APIResponse(
+                            id=self.context.task_id,
+                            model_internal=self.context.model_name,
+                            prompt=self.context.prompt,
+                            sampling_params=self.context.sampling_params,
+                            status_code=None,
+                            is_error=True,
+                            error_message="Request timed out (terminated by client).",
+                            content=None,
+                            usage=None,
+                        )
+                    # poll for the response
+                    await asyncio.sleep(5.0)
+                    async with session.get(
+                        url=f"{self.url}/{response_id}",
+                        headers=self.request_header,
+                    ) as http_response:
+                        http_response.raise_for_status()
+                        data = await http_response.json()
+
+                        if data["status"] != last_status:
+                            print(
+                                f"Background req {response_id} status updated to: {data['status']}"
+                            )
+                            last_status = data["status"]
+                        if last_status not in ["queued", "in_progress"]:
+                            return await self.handle_response(http_response)
+
+            except Exception as e:
+                raise_if_modal_exception(e)
+                tb = traceback.format_exc()
+                print(tb)
+                return APIResponse(
+                    id=self.context.task_id,
+                    model_internal=self.context.model_name,
+                    prompt=self.context.prompt,
+                    sampling_params=self.context.sampling_params,
+                    status_code=None,
+                    is_error=True,
+                    error_message=f"Unexpected {type(e).__name__}: {str(e) or 'No message.'}",
+                    content=None,
+                    usage=None,
+                )
+
     async def execute_once(self) -> APIResponse:
         """Send the HTTP request once and return the parsed APIResponse."""
         await self.build_request()
         assert self.context.status_tracker
-        # try:
-        #     dumped = json.dumps(self.request_json)
-        # except Exception:
-        #     print("couldn't serialize request json")
-        #     print(self.request_json)
+
+        if (
+            self.context.background
+            and self.context.use_responses_api
+            and self.context.model_name in OPENAI_MODELS
+        ):
+            return await self._execute_once_background_mode()
+
         try:
             self.context.status_tracker.total_requests += 1
             timeout = aiohttp.ClientTimeout(total=self.context.request_timeout)

lm_deluge/api_requests/bedrock.py

@@ -1,10 +1,11 @@
 import asyncio
 import json
 import os
-import warnings
 
 from aiohttp import ClientResponse
 
+from lm_deluge.warnings import maybe_warn
+
 try:
     from requests_aws4auth import AWS4Auth
 except ImportError:
@@ -187,9 +188,7 @@ async def _build_openai_bedrock_request(
     # Note: GPT-OSS on Bedrock doesn't support response_format parameter
     # Even though the model supports JSON, we can't use the response_format parameter
     if sampling_params.json_mode and model.supports_json:
-        warnings.warn(
-            f"JSON mode requested for {model.name} but response_format parameter not supported on Bedrock"
-        )
+        maybe_warn("WARN_JSON_MODE_UNSUPPORTED", model_name=model.name)
 
     if tools:
         request_tools = []

lm_deluge/api_requests/chat_reasoning.py

@@ -0,0 +1,4 @@
+# this request type is for models that add "reasoning_content"
+# on top of the openai chat completions. it's important to be separate
+# for providers that expect you to provide back the reasoning content to
+# preserve best performance.

lm_deluge/api_requests/gemini.py

@@ -1,11 +1,12 @@
 import json
 import os
-import warnings
 from typing import Any
+
 from aiohttp import ClientResponse
 
 from lm_deluge.request_context import RequestContext
 from lm_deluge.tool import Tool
+from lm_deluge.warnings import maybe_warn
 
 from ..config import SamplingParams
 from ..models import APIModel
@@ -54,9 +55,7 @@ async def _build_gemini_request(
 
     else:
         if sampling_params.reasoning_effort:
-            warnings.warn(
-                f"Ignoring reasoning_effort param for non-reasoning model: {model.name}"
-            )
+            maybe_warn("WARN_REASONING_UNSUPPORTED", model_name=model.name)
 
     # Add tools if provided
     if tools:
@@ -76,8 +75,10 @@ class GeminiRequest(APIRequestBase):
 
         # Warn if cache is specified for Gemini model
         if self.context.cache is not None:
-            warnings.warn(
-                f"Cache parameter '{self.context.cache}' is not supported for Gemini models, ignoring for {self.context.model_name}"
+            maybe_warn(
+                "WARN_CACHING_UNSUPPORTED",
+                model_name=self.context.model_name,
+                cache_param=self.context.cache,
             )
 
         self.model = APIModel.from_registry(self.context.model_name)

lm_deluge/api_requests/mistral.py

@@ -1,9 +1,10 @@
 import json
 import os
-import warnings
 
 from aiohttp import ClientResponse
 
+from lm_deluge.warnings import maybe_warn
+
 from ..models import APIModel
 from ..prompt import Message
 from ..request_context import RequestContext
@@ -17,8 +18,10 @@ class MistralRequest(APIRequestBase):
 
         # Warn if cache is specified for non-Anthropic model
         if self.context.cache is not None:
-            warnings.warn(
-                f"Cache parameter '{self.context.cache}' is only supported for Anthropic models, ignoring for {self.context.model_name}"
+            maybe_warn(
+                "WARN_CACHING_UNSUPPORTED",
+                model_name=self.context.model_name,
+                cache_param=self.context.cache,
             )
         self.model = APIModel.from_registry(self.context.model_name)
 
@@ -38,13 +41,9 @@ class MistralRequest(APIRequestBase):
             "max_tokens": self.context.sampling_params.max_new_tokens,
         }
         if self.context.sampling_params.reasoning_effort:
-            warnings.warn(
-                f"Ignoring reasoning_effort param for non-reasoning model: {self.context.model_name}"
-            )
+            maybe_warn("WARN_REASONING_UNSUPPORTED", model_name=self.context.model_name)
         if self.context.sampling_params.logprobs:
-            warnings.warn(
-                f"Ignoring logprobs param for non-logprobs model: {self.context.model_name}"
-            )
+            maybe_warn("WARN_LOGPROBS_UNSUPPORTED", model_name=self.context.model_name)
         if self.context.sampling_params.json_mode and self.model.supports_json:
             self.request_json["response_format"] = {"type": "json_object"}