lm-deluge 0.0.30__tar.gz → 0.0.31__tar.gz

{lm_deluge-0.0.30/src/lm_deluge.egg-info → lm_deluge-0.0.31}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lm_deluge
-Version: 0.0.30
+Version: 0.0.31
 Summary: Python utility for using LLM API models.
 Author-email: Benjamin Anderson <ben@trytaylor.ai>
 Requires-Python: >=3.10

{lm_deluge-0.0.30 → lm_deluge-0.0.31}/pyproject.toml

@@ -3,7 +3,7 @@ requires = ["setuptools", "wheel"]
 
 [project]
 name = "lm_deluge"
-version = "0.0.30"
+version = "0.0.31"
 authors = [{ name = "Benjamin Anderson", email = "ben@trytaylor.ai" }]
 description = "Python utility for using LLM API models."
 readme = "README.md"

{lm_deluge-0.0.30 → lm_deluge-0.0.31}/src/lm_deluge/api_requests/bedrock.py

@@ -132,7 +132,7 @@ async def _build_anthropic_bedrock_request(
         if len(mcp_servers) > 0:
             request_json["mcp_servers"] = mcp_servers
 
-    return request_json, base_headers, auth, url
+    return request_json, base_headers, auth, url, region
 
 
 class BedrockRequest(APIRequestBase):
@@ -140,6 +140,7 @@ class BedrockRequest(APIRequestBase):
         super().__init__(context=context)
 
         self.model = APIModel.from_registry(self.context.model_name)
+        self.region = None  # Will be set during build_request
 
     async def build_request(self):
         self.url = f"{self.model.api_base}/messages"
@@ -153,6 +154,7 @@ class BedrockRequest(APIRequestBase):
             base_headers,
             self.auth,
             self.url,
+            self.region,
         ) = await _build_anthropic_bedrock_request(self.model, self.context)
         self.request_header = self.merge_headers(
             base_headers, exclude_patterns=["anthropic", "openai", "gemini", "mistral"]
@@ -160,6 +162,7 @@ class BedrockRequest(APIRequestBase):
 
     async def execute_once(self) -> APIResponse:
         """Override execute_once to handle AWS4Auth signing."""
+        await self.build_request()
         import aiohttp
 
         assert self.context.status_tracker

{lm_deluge-0.0.30 → lm_deluge-0.0.31}/src/lm_deluge/file.py

@@ -37,7 +37,10 @@ class File:
             header, encoded = self.data.split(",", 1)
             return base64.b64decode(encoded)
         else:
-            raise ValueError("unreadable file format")
+            err = f"unreadable file. self.data type: {type(self.data)}"
+            if isinstance(self.data, str) and len(self.data) < 1_000:
+                err += f". self.data: {len(self.data)}"
+            raise ValueError(err)
 
     def _mime(self) -> str:
         if self.media_type:

{lm_deluge-0.0.30 → lm_deluge-0.0.31}/src/lm_deluge/models.py

@@ -28,7 +28,7 @@ BUILTIN_MODELS = {
         "reasoning_model": False,
     },
     "llama-4-maverick": {
-        "id": "llama-4-scout",
+        "id": "llama-4-maverick",
         "name": "Llama-4-Maverick-17B-128E-Instruct-FP8",
         "api_base": "https://api.llama.com/compat/v1",
         "api_key_env_var": "META_API_KEY",

{lm_deluge-0.0.30 → lm_deluge-0.0.31/src/lm_deluge.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lm_deluge
-Version: 0.0.30
+Version: 0.0.31
 Summary: Python utility for using LLM API models.
 Author-email: Benjamin Anderson <ben@trytaylor.ai>
 Requires-Python: >=3.10