mojentic 0.7.2__py3-none-any.whl → 0.7.3__py3-none-any.whl

mojentic/llm/gateways/anthropic.py

@@ -29,7 +29,7 @@ class AnthropicGateway(LLMGateway):
         response = self.client.messages.create(
             **anthropic_args,
             temperature=args.get('temperature', 1.0),
-            max_tokens=args.get('num_predict', 2000),
+            max_tokens=args.get('max_tokens', args.get('num_predict', 2000)),
             # thinking={
             #     "type": "enabled",
             #     "budget_tokens": 32768,

mojentic/llm/gateways/llm_gateway.py

@@ -19,7 +19,7 @@ class LLMGateway:
                  object_model: Optional[Type[BaseModel]] = None,
                  tools: Optional[List[LLMTool]] = None,
                  temperature: float = 1.0,
-                 num_ctx: int = 32768,
+                 num_ctx: int = 32768, max_tokens: int = 16384,
                  num_predict: int = -1) -> LLMGatewayResponse:
         """
         Complete the LLM request.
@@ -39,6 +39,8 @@ class LLMGateway:
             The temperature to use for the response. Defaults to 1.0.
         num_ctx : int
             The number of context tokens to use. Defaults to 32768.
+        max_tokens : int
+            The maximum number of tokens to generate. Defaults to 16384.
         num_predict : int
             The number of tokens to predict. Defaults to no limit.
 

mojentic/llm/gateways/ollama.py

@@ -35,6 +35,8 @@ class OllamaGateway(LLMGateway):
         )
         if args.get('num_predict', 0) > 0:
             options.num_predict = args['num_predict']
+        if 'max_tokens' in args:
+            options.num_predict = args['max_tokens']
         return options
 
     def complete(self, **args) -> LLMGatewayResponse:
@@ -56,6 +58,8 @@ class OllamaGateway(LLMGateway):
             The temperature to use for the response. Defaults to 1.0.
         num_ctx : int, optional
             The number of context tokens to use. Defaults to 32768.
+        max_tokens : int, optional
+            The maximum number of tokens to generate. Defaults to 16384.
         num_predict : int, optional
             The number of tokens to predict. Defaults to no limit.
 
@@ -120,6 +124,8 @@ class OllamaGateway(LLMGateway):
             The temperature to use for the response. Defaults to 1.0.
         num_ctx : int, optional
             The number of context tokens to use. Defaults to 32768.
+        max_tokens : int, optional
+            The maximum number of tokens to generate. Defaults to 16384.
         num_predict : int, optional
             The number of tokens to predict. Defaults to no limit.
 

mojentic/llm/gateways/openai.py

@@ -46,6 +46,8 @@ class OpenAIGateway(LLMGateway):
             The temperature to use for the response. Defaults to 1.0.
         num_ctx : int, optional
             The number of context tokens to use. Defaults to 32768.
+        max_tokens : int, optional
+            The maximum number of tokens to generate. Defaults to 16384.
         num_predict : int, optional
             The number of tokens to predict. Defaults to no limit.
 
@@ -68,6 +70,9 @@ class OpenAIGateway(LLMGateway):
         if 'tools' in args and args['tools'] is not None:
             openai_args['tools'] = [t.descriptor for t in args['tools']]
 
+        if 'max_tokens' in args:
+            openai_args['max_tokens'] = args['max_tokens']
+
         response = completion(**openai_args)
 
         object = None

mojentic/llm/llm_broker.py

@@ -5,18 +5,19 @@ from typing import List, Optional, Type
 import structlog
 from pydantic import BaseModel
 
-from mojentic.tracer.tracer_system import TracerSystem
 from mojentic.llm.gateways.llm_gateway import LLMGateway
 from mojentic.llm.gateways.models import MessageRole, LLMMessage, LLMGatewayResponse
 from mojentic.llm.gateways.ollama import OllamaGateway
 from mojentic.llm.gateways.tokenizer_gateway import TokenizerGateway
+from mojentic.tracer.tracer_system import TracerSystem
 
 logger = structlog.get_logger()
 
 
 class LLMBroker():
     """
-    This class is responsible for managing interaction with a Large Language Model. It abstracts the user
+    This class is responsible for managing interaction with a Large Language Model. It abstracts
+    the user
     from the specific mechanics of the LLM and provides a common interface for generating responses.
     """
 
@@ -25,7 +26,8 @@ class LLMBroker():
     model: str
     tracer: Optional[TracerSystem]
 
-    def __init__(self, model: str, gateway: Optional[LLMGateway] = None, tokenizer: Optional[TokenizerGateway] = None,
+    def __init__(self, model: str, gateway: Optional[LLMGateway] = None,
+                 tokenizer: Optional[TokenizerGateway] = None,
                  tracer: Optional[TracerSystem] = None):
         """
         Create an instance of the LLMBroker.
@@ -35,10 +37,12 @@ class LLMBroker():
         model
             The name of the model to use.
         gateway
-            The gateway to use for communication with the LLM. If None, a gateway is created that will utilize a local
+            The gateway to use for communication with the LLM. If None, a gateway is created that
+            will utilize a local
             Ollama server.
         tokenizer
-            The gateway to use for tokenization. This is used to log approximate token counts for the LLM calls. If
+            The gateway to use for tokenization. This is used to log approximate token counts for
+            the LLM calls. If
             None, `mxbai-embed-large` is used on a local Ollama server.
         tracer
             Optional tracer system to record LLM calls and responses.
@@ -58,8 +62,9 @@ class LLMBroker():
         else:
             self.adapter = gateway
 
-    def generate(self, messages: List[LLMMessage], tools=None, temperature=1.0, num_ctx=32768, num_predict=-1, 
-              correlation_id: str = None) -> str:
+    def generate(self, messages: List[LLMMessage], tools=None, temperature=1.0, num_ctx=32768,
+                 num_predict=-1, max_tokens=16384,
+                 correlation_id: str = None) -> str:
         """
         Generate a text response from the LLM.
 
@@ -68,7 +73,8 @@ class LLMBroker():
         messages : LLMMessage
             A list of messages to send to the LLM.
         tools : List[Tool]
-            A list of tools to use with the LLM. If a tool call is requested, the tool will be called and the output
+            A list of tools to use with the LLM. If a tool call is requested, the tool will be
+            called and the output
             will be included in the response.
         temperature : float
             The temperature to use for the response. Defaults to 1.0
@@ -91,10 +97,11 @@ class LLMBroker():
         messages_for_tracer = [m.model_dump() for m in messages]
 
         # Record LLM call in tracer
-        tools_for_tracer = [{"name": t.name, "description": t.description} for t in tools] if tools else None
+        tools_for_tracer = [{"name": t.name, "description": t.description} for t in
+                            tools] if tools else None
         self.tracer.record_llm_call(
-            self.model, 
-            messages_for_tracer, 
+            self.model,
+            messages_for_tracer,
             temperature,
             tools=tools_for_tracer,
             source=type(self),
@@ -110,12 +117,14 @@ class LLMBroker():
             tools=tools,
             temperature=temperature,
             num_ctx=num_ctx,
-            num_predict=num_predict)
+            num_predict=num_predict,
+            max_tokens=max_tokens)
 
         call_duration_ms = (time.time() - start_time) * 1000
 
         # Record LLM response in tracer
-        tool_calls_for_tracer = [tc.model_dump() for tc in result.tool_calls] if result.tool_calls else None
+        tool_calls_for_tracer = [tc.model_dump() for tc in
+                                 result.tool_calls] if result.tool_calls else None
         self.tracer.record_llm_response(
             self.model,
             result.content,
@@ -153,13 +162,17 @@ class LLMBroker():
                     logger.info('Function output', output=output)
                     messages.append(LLMMessage(role=MessageRole.Assistant, tool_calls=[tool_call]))
                     messages.append(
-                        LLMMessage(role=MessageRole.Tool, content=json.dumps(output), tool_calls=[tool_call]))
-                    # {'role': 'tool', 'content': str(output), 'name': tool_call.name, 'tool_call_id': tool_call.id})
-                    return self.generate(messages, tools, temperature, num_ctx, num_predict, correlation_id=correlation_id)
+                        LLMMessage(role=MessageRole.Tool, content=json.dumps(output),
+                                   tool_calls=[tool_call]))
+                    # {'role': 'tool', 'content': str(output), 'name': tool_call.name,
+                    # 'tool_call_id': tool_call.id})
+                    return self.generate(messages, tools, temperature, num_ctx, num_predict,
+                                         correlation_id=correlation_id)
                 else:
                     logger.warn('Function not found', function=tool_call.name)
                     logger.info('Expected usage of missing function', expected_usage=tool_call)
-                    # raise Exception('Unknown tool function requested:', requested_tool.function.name)
+                    # raise Exception('Unknown tool function requested:',
+                    # requested_tool.function.name)
 
         return result.content
 
@@ -170,8 +183,9 @@ class LLMBroker():
                 content += message.content
         return content
 
-    def generate_object(self, messages: List[LLMMessage], object_model: Type[BaseModel], temperature=1.0, num_ctx=32768,
-                        num_predict=-1, correlation_id: str = None) -> BaseModel:
+    def generate_object(self, messages: List[LLMMessage], object_model: Type[BaseModel],
+                        temperature=1.0, num_ctx=32768, num_predict=-1, max_tokens=16384,
+                        correlation_id: str = None) -> BaseModel:
         """
         Generate a structured response from the LLM and return it as an object.
 
@@ -203,8 +217,8 @@ class LLMBroker():
 
         # Record LLM call in tracer
         self.tracer.record_llm_call(
-            self.model, 
-            messages_for_tracer, 
+            self.model,
+            messages_for_tracer,
             temperature,
             tools=None,
             source=type(self),
@@ -214,14 +228,18 @@ class LLMBroker():
         # Measure call duration for audit
         start_time = time.time()
 
-        result = self.adapter.complete(model=self.model, messages=messages, object_model=object_model,
-                                       temperature=temperature, num_ctx=num_ctx, num_predict=num_predict)
+        result = self.adapter.complete(model=self.model, messages=messages,
+                                       object_model=object_model,
+                                       temperature=temperature, num_ctx=num_ctx,
+                                       num_predict=num_predict, max_tokens=max_tokens)
 
         call_duration_ms = (time.time() - start_time) * 1000
 
         # Record LLM response in tracer with object representation
         # Convert object to string for tracer
-        object_str = str(result.object.model_dump()) if hasattr(result.object, "model_dump") else str(result.object)
+        object_str = str(result.object.model_dump()) if hasattr(result.object,
+                                                                "model_dump") else str(
+            result.object)
         self.tracer.record_llm_response(
             self.model,
             f"Structured response: {object_str}",

{mojentic-0.7.2.dist-info → mojentic-0.7.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mojentic
-Version: 0.7.2
+Version: 0.7.3
 Summary: Mojentic is an agentic framework that aims to provide a simple and flexible way to assemble teams of agents to solve complex problems.
 Author-email: Stacey Vetzal <stacey@vetzal.com>
 Project-URL: Homepage, https://github.com/svetzal/mojentic

{mojentic-0.7.2.dist-info → mojentic-0.7.3.dist-info}/RECORD

@@ -69,21 +69,21 @@ mojentic/context/shared_working_memory.py,sha256=Zt9MNGErEkDIUAaHvyhEOiTaEobI9l0
 mojentic/llm/__init__.py,sha256=rTUTMjAe524evcH09wKSpL5ymnwHFlM1fWvGKd3mSis,414
 mojentic/llm/chat_session.py,sha256=H2gY0mZYVym8jC69VHsmKaRZ9T87Suyw0-TW5r850nA,3992
 mojentic/llm/chat_session_spec.py,sha256=8-jj-EHV2WwWuvo3t8I75kSEAYiG1nR-OEwkkLTi_z0,3872
-mojentic/llm/llm_broker.py,sha256=i9jhSulXV4lQX9G6YOQTiMZt4UMipqUfvl-P-gmXQXo,9200
+mojentic/llm/llm_broker.py,sha256=wNBGQD9GwuZoj3liZZEfCFN4sb2_TN8HDHZvZc7-PMQ,9726
 mojentic/llm/llm_broker_spec.py,sha256=40lzmYm_6Zje6z5MQ7_o3gSBThLsNW_l_1mZTUVll6A,5342
 mojentic/llm/message_composers.py,sha256=6_BbBGx-D77gtx97bN2sfftP8lMt0tlSs6GvT1sdVmU,12114
 mojentic/llm/message_composers_spec.py,sha256=RNW14Zb-kIBWT5Wy9cZQyxHPrcRIaBFjBYCR9N-m0kE,12109
 mojentic/llm/gateways/__init__.py,sha256=vz1FDJSaBm-WnCuIyXy5ybrU0X4RxVtnhtaCqHPwCSA,711
-mojentic/llm/gateways/anthropic.py,sha256=SsyNjq9QaXaqiMM43C9fwLp57hpgFtwNPJUnOAYVrtc,1788
+mojentic/llm/gateways/anthropic.py,sha256=_sX_sQpbESoMc-s02sPkgBwpp1WVsHtwJ7tf5ui6rBM,1812
 mojentic/llm/gateways/anthropic_messages_adapter.py,sha256=K6kEZeVt7E1endbGMLsh5l9SxC3Y5dnvbcejVqi_qUs,3003
 mojentic/llm/gateways/embeddings_gateway.py,sha256=kcOhiyHzOyQgKgwPDQJD5oVvfwk71GsBgMYJOIDv5NU,1347
 mojentic/llm/gateways/file_gateway.py,sha256=3bZpalSyl_R4016WzCmmjUBDtAgPsmx19eVGv6p1Ufk,1418
-mojentic/llm/gateways/llm_gateway.py,sha256=5BayP6VuMgMHdAzCFaXLRjRCWh-IOYnaq_s4LZ0_3x4,2559
+mojentic/llm/gateways/llm_gateway.py,sha256=v4Q8nOw9n_bef3Udnc40McUDwylyMzfkr_1x3Yqc688,2682
 mojentic/llm/gateways/models.py,sha256=OyIaMHKrrx6dHo5FbC8qOFct7PRql9wqbe_BJlgDSDE,3015
-mojentic/llm/gateways/ollama.py,sha256=fBnmG4B0Trq8BJa8eZgrYUEVgb3kiTPytW1e6aTIjj8,7607
+mojentic/llm/gateways/ollama.py,sha256=OUUImBNzPte52Gsf-e7TBjDHRvYW5flU9ddxwG2zlzk,7909
 mojentic/llm/gateways/ollama_messages_adapter.py,sha256=kUN_p2FyN88_trXMcL-Xsn9xPBU7pGKlJwTUEUCf6G4,1404
 mojentic/llm/gateways/ollama_messages_adapter_spec.py,sha256=gVRbWDrHOa1EiZ0CkEWe0pGn-GKRqdGb-x56HBQeYSE,4981
-mojentic/llm/gateways/openai.py,sha256=P40KKfNCmY_EKDP6VUKu7sM2GaqX_iUPx6_2_BYb80Q,5489
+mojentic/llm/gateways/openai.py,sha256=O2iMHGGnw96zgni10jE7L1c5HoskGOO5yMqlp8bIu8w,5690
 mojentic/llm/gateways/openai_message_adapter_spec.py,sha256=ITBSV5njldV_x0NPgjmg8Okf9KzevQJ8dTXM-t6ubcg,6612
 mojentic/llm/gateways/openai_messages_adapter.py,sha256=Scal68JKKdBHB35ok1c5DeWYdD6Wra5oXSsPxJyyXSQ,3947
 mojentic/llm/gateways/tokenizer_gateway.py,sha256=ztuqfunlJ6xmyUPPHcC_69-kegiNJD6jdSEde7hDh2w,485
@@ -131,8 +131,8 @@ mojentic/tracer/tracer_system.py,sha256=7CPy_2tlsHtXQ4DcO5oo52N9a9WS0GH-mjeINzu6
 mojentic/tracer/tracer_system_spec.py,sha256=TNm0f9LV__coBx0JGEKyzzNN9mFjCSG_SSrRISO8Xeg,8632
 mojentic/utils/__init__.py,sha256=lqECkkoFvHFttDnafRE1vvh0Dmna_lwupMToP5VvX5k,115
 mojentic/utils/formatting.py,sha256=bPrwwdluXdQ8TsFxfWtHNOeMWKNvAfABSoUnnA1g7c8,947
-mojentic-0.7.2.dist-info/licenses/LICENSE.md,sha256=txSgV8n5zY1W3NiF5HHsCwlaW0e8We1cSC6TuJUqxXA,1060
-mojentic-0.7.2.dist-info/METADATA,sha256=QzdW64AEvamEdUIKRySmkHJI0eJWLHVRmZ4xAHLU2JQ,5475
-mojentic-0.7.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-mojentic-0.7.2.dist-info/top_level.txt,sha256=Q-BvPQ8Eu1jnEqK8Xkr6A9C8Xa1z38oPZRHuA5MCTqg,19
-mojentic-0.7.2.dist-info/RECORD,,
+mojentic-0.7.3.dist-info/licenses/LICENSE.md,sha256=txSgV8n5zY1W3NiF5HHsCwlaW0e8We1cSC6TuJUqxXA,1060
+mojentic-0.7.3.dist-info/METADATA,sha256=YTPkYyv6DMo1lluiT6uY5t0xDz1wGyh7dsdRP5ZYTig,5475
+mojentic-0.7.3.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+mojentic-0.7.3.dist-info/top_level.txt,sha256=Q-BvPQ8Eu1jnEqK8Xkr6A9C8Xa1z38oPZRHuA5MCTqg,19
+mojentic-0.7.3.dist-info/RECORD,,