versionhq 1.2.2.4__tar.gz → 1.2.2.6__tar.gz

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/.env.sample

@@ -16,3 +16,7 @@ COMPOSIO_API_KEY=
 COMPOSIO_CLI_KEY=
 
 MEM0_API_KEY=
+
+AWS_ACCESS_KEY_ID=
+AWS_SECRET_ACCESS_KEY=
+AWS_REGION_NAME=

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/.github/workflows/run_tests.yml

@@ -17,6 +17,9 @@ env:
   DEFAULT_REDIRECT_URL:  ${{ secrets.DEFAULT_REDIRECT_URL }}
   DEFAULT_USER_ID:  ${{ secrets.DEFAULT_USER_ID }}
   MEM0_API_KEY: ${{ secrets.MEM0_API_KEY }}
+  AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
+  AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+  AWS_REGION_NAME: ${{ secrets.AWS_REGION_NAME }}
 
 jobs:
   run_test:

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: versionhq
-Version: 1.2.2.4
+Version: 1.2.2.6
 Summary: An agentic orchestration framework for building agent networks that handle task automation.
 Author-email: Kuriko Iwai <kuriko@versi0n.io>
 License: MIT License
@@ -66,6 +66,7 @@ Requires-Dist: envoy>=0.0.3
 Requires-Dist: composio-core==0.7.0
 Requires-Dist: networkx>=3.4.2
 Requires-Dist: matplotlib>=3.10.0
+Requires-Dist: boto3>=1.37.1
 Provides-Extra: docling
 Requires-Dist: docling>=2.17.0; extra == "docling"
 Provides-Extra: mem0ai
@@ -141,7 +142,7 @@ Agentic orchestration framework for multi-agent networks and task graphs for com
 
 `versionhq` is a Python framework for agent networks that handle complex task automation without human interaction.
 
-Agents are model-agnostic, and will improve task output, while oprimizing token cost and job latency, by sharing their memory, knowledge base, and RAG tools with other agents in the network.
+Agents are model-agnostic, and will improve task output, while optimizing token cost and job latency, by sharing their memory, knowledge base, and RAG tools with other agents in the network.
 
 
 ###  Agent Network
@@ -227,10 +228,7 @@ The following code snippet demonstrates agent customization:
 ```python
 import versionhq as vhq
 
-agent = vhq.Agent(
-   role="Marketing Analyst",
-   goal="my amazing goal"
-) # assuming this agent was created during the network formation
+agent = vhq.Agent(role="Marketing Analyst")
 
 # update the agent
 agent.update(
@@ -321,8 +319,8 @@ To create an agent network with one or more manager agents, designate members us
 ```python
 import versionhq as vhq
 
-agent_a = vhq.Agent(role="agent a", goal="My amazing goals", llm="llm-of-your-choice")
-agent_b = vhq.Agent(role="agent b", goal="My amazing goals", llm="llm-of-your-choice")
+agent_a = vhq.Agent(role="agent a", llm="llm-of-your-choice")
+agent_b = vhq.Agent(role="agent b", llm="llm-of-your-choice")
 
 task_1 = vhq.Task(
    description="Analyze the client's business model.",

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/README.md

@@ -55,7 +55,7 @@ Agentic orchestration framework for multi-agent networks and task graphs for com
 
 `versionhq` is a Python framework for agent networks that handle complex task automation without human interaction.
 
-Agents are model-agnostic, and will improve task output, while oprimizing token cost and job latency, by sharing their memory, knowledge base, and RAG tools with other agents in the network.
+Agents are model-agnostic, and will improve task output, while optimizing token cost and job latency, by sharing their memory, knowledge base, and RAG tools with other agents in the network.
 
 
 ###  Agent Network
@@ -141,10 +141,7 @@ The following code snippet demonstrates agent customization:
 ```python
 import versionhq as vhq
 
-agent = vhq.Agent(
-   role="Marketing Analyst",
-   goal="my amazing goal"
-) # assuming this agent was created during the network formation
+agent = vhq.Agent(role="Marketing Analyst")
 
 # update the agent
 agent.update(
@@ -235,8 +232,8 @@ To create an agent network with one or more manager agents, designate members us
 ```python
 import versionhq as vhq
 
-agent_a = vhq.Agent(role="agent a", goal="My amazing goals", llm="llm-of-your-choice")
-agent_b = vhq.Agent(role="agent b", goal="My amazing goals", llm="llm-of-your-choice")
+agent_a = vhq.Agent(role="agent a", llm="llm-of-your-choice")
+agent_b = vhq.Agent(role="agent b", llm="llm-of-your-choice")
 
 task_1 = vhq.Task(
    description="Analyze the client's business model.",

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/docs/core/agent/config.md

@@ -11,11 +11,7 @@ By default, when the model provider name is provided, we will select the most co
 ```python
 import versionhq as vhq
 
-agent = vhq.Agent(
-	role="Marketing Analyst",
-	goal="Coping with price competition in saturated markets",
-	llm="gemini-2.0"
-)
+agent = vhq.Agent(role="Marketing Analyst", llm="gemini-2.0")
 ```
 
 <hr/>
@@ -34,7 +30,6 @@ import versionhq as vhq
 
 agent = vhq.Agent(
     role="Marketing Analyst",
-    goal="Coping with increased price competition in saturated markets.",
     respect_context_window=False,
     max_execution_time=60,
     max_rpm=5,
@@ -48,10 +43,10 @@ agent = vhq.Agent(
     )
 
 assert isinstance(agent.llm, vhq.LLM)
-assert agent.llm.temperature == 1
-assert agent.llm.top_p == 0.1
-assert agent.llm.n == 1
-assert agent.llm.stop == "answer"
+assert agent.llm.llm_config["temperature"] == 1
+assert agent.llm.llm_config["top_p"] == 0.1
+assert agent.llm.llm_config["n"] == 1
+assert agent.llm.llm_config["stop"] == "answer"
 ```
 
 <hr>

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/docs/core/agent/index.md

@@ -7,7 +7,7 @@ tags:
 
 <class>`class` versionhq.agent.model.<bold>Agent<bold></class>
 
-A Pydantic class to store `Agent` objects and handle `Task` execution as well as `LLM` configuration.
+A Pydantic class to store an `Agent` object that handles `Task` execution.
 
 
 ## Quick Start

versionhq-1.2.2.6/docs/core/llm/index.md

@@ -0,0 +1,73 @@
+---
+tags:
+  - Agent Network
+---
+
+# LLM
+
+<class>`class` versionhq.llm.model.<bold>LLM<bold></class>
+
+A Pydantic class to store LLM objects and its task handling rules.
+
+You can specify a model and integration platform from the list. Else, we'll use `gemini` or `gpt` via `LiteLLM` by default.
+
+
+**List of available models**
+
+```python
+"openai": [
+    "gpt-4",
+    "gpt-4o",
+    "gpt-4o-mini",
+    "o1-mini",
+    "o1-preview",
+],
+"gemini": [
+    "gemini/gemini-1.5-flash",
+    "gemini/gemini-1.5-pro",
+    "gemini/gemini-2.0-flash-exp",
+],
+"anthropic": [
+    "claude-3-7-sonnet-latest",
+    "claude-3-5-sonnet-20241022",
+    "claude-3-5-sonnet-20240620",
+    "claude-3-haiku-2024030",
+    "claude-3-opus-20240229",
+    "claude-3-haiku-20240307",
+],
+"openrouter": [
+    "openrouter/deepseek/deepseek-r1",
+
+    "openrouter/qwen/qwen-2.5-72b-instruct",
+
+    "openrouter/google/gemini-2.0-flash-thinking-exp:free",
+    "openrouter/google/gemini-2.0-flash-thinking-exp-1219:free",
+    "openrouter/google/gemini-2.0-flash-001",
+
+    "openrouter/meta-llama/llama-3.3-70b-instruct",
+    "openrouter/mistralai/mistral-large-2411",
+    "openrouter/cohere/command-r-plus",
+],
+"bedrock": [
+    "bedrock/converse/us.meta.llama3-3-70b-instruct-v1:0",
+    "bedrock/us.meta.llama3-2-1b-instruct-v1:0",
+    "bedrock/us.meta.llama3-2-3b-instruct-v1:0",
+    "bedrock/us.meta.llama3-2-11b-instruct-v1:0",
+
+    "bedrock/mistral.mistral-7b-instruct-v0:2",
+    "bedrock/mistral.mixtral-8x7b-instruct-v0:1",
+    "bedrock/mistral.mistral-large-2407-v1:0",
+
+    "bedrock/amazon.titan-text-lite-v1",
+    "bedrock/amazon.titan-text-express-v1",
+    "bedrock/amazon.titan-text-premier-v1:0",
+
+    "bedrock/cohere.command-r-plus-v1:0",
+    "bedrock/cohere.command-r-v1:0",
+    "bedrock/cohere.command-text-v14",
+    "bedrock/cohere.command-light-text-v14",
+],
+"huggingface": [
+    "huggingface/qwen/qwen2.5-VL-72B-Instruct",
+]
+```

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/docs/core/task/task-strc-response.md

@@ -165,7 +165,7 @@ import versionhq as vhq
 from pydantic import BaseModel
 from typing import Any
 
-# 1. Define and execute a sub task with Pydantic output.
+# 1. Defines a sub task
 class Sub(BaseModel):
     sub1: list[dict[str, Any]]
     sub2: dict[str, Any]
@@ -176,27 +176,28 @@ sub_task = vhq.Task(
 )
 sub_res = sub_task.execute()
 
-# 2. Define a main task, callback function to format the final response.
+# 2. Defines a main task with callbacks
 class Main(BaseModel):
-    main1: list[Any] # <= assume expecting to store Sub object in this field.
-    # error_main1: list[InstanceOf[Sub]]  # as this will trigger 400 error!
+    main1: list[Any] # <= assume expecting to store Sub object.
     main2: dict[str, Any]
 
-def format_response(sub: InstanceOf[Sub], main1: list[Any], main2: dict[str, Any]) -> Main:
-    main1.append(sub)
+def format_response(sub, main1, main2) -> Main:
+    if main1:
+        main1.append(sub)
     main = Main(main1=main1, main2=main2)
     return main
 
-# 3. Execute
+# 3. Executes
 main_task = vhq.Task(
-    description="generate random values that strictly follows the given format",
+    description="generate random values that strictly follows the given format.",
     pydantic_output=Main,
     callback=format_response,
-    callback_kwargs=dict(sub=Sub(sub1=sub_res.pydantic.sub1, sub2=sub_res.pydantic.sub2)),
+    callback_kwargs=dict(sub=sub_res.json_dict),
 )
-res = main_task.execute(context=sub_res.raw) # [Optional] Adding sub_task as a context.
+res = main_task.execute(context=sub_res.raw) # [Optional] Adding sub_task's response as context.
 
-assert [item for item in res.callback_output.main1 if isinstance(item, Sub)]
+assert res.callback_output.main1 is not None
+assert res.callback_output.main2 is not None
 ```
 
 To automate these manual setups, refer to <a href="/core/agent-network">AgentNetwork</a> class.

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/docs/index.md

@@ -25,7 +25,7 @@ A Python framework for agentic orchestration that handles complex task automatio
 
 `versionhq` is a Python framework for agent networks that handle complex task automation without human interaction.
 
-Agents are model-agnostic, and will improve task output, while oprimizing token cost and job latency, by sharing their memory, knowledge base, and RAG tools with other agents in the network.
+Agents are model-agnostic, and will improve task output, while optimizing token cost and job latency, by sharing their memory, knowledge base, and RAG tools with other agents in the network.
 
 
 ### Agent Network
@@ -113,10 +113,7 @@ The following code snippet demonstrates agent customization:
 ```python
 import versionhq as vhq
 
-agent = vhq.Agent(
-   role="Marketing Analyst",
-   goal="my amazing goal"
-) # assuming this agent was created during the network formation
+agent = vhq.Agent(role="Marketing Analyst")
 
 # update the agent
 agent.update(

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/docs/quickstart.md

@@ -48,7 +48,7 @@ def dummy_func(message: str, test1: str, test2: list[str]) -> str:
    return f"""{message}: {test1}, {", ".join(test2)}"""
 
 
-agent = vhq.Agent(role="demo", goal="amazing project goal")
+agent = vhq.Agent(role="demo manager")
 
 task = vhq.Task(
    description="Amazing task",

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/pyproject.toml

@@ -15,7 +15,7 @@ exclude = ["test*", "__pycache__", "*.egg-info"]
 
 [project]
 name = "versionhq"
-version = "1.2.2.4"
+version = "1.2.2.6"
 authors = [{ name = "Kuriko Iwai", email = "kuriko@versi0n.io" }]
 description = "An agentic orchestration framework for building agent networks that handle task automation."
 readme = "README.md"
@@ -46,6 +46,7 @@ dependencies = [
   "composio-core==0.7.0",
   "networkx>=3.4.2",
   "matplotlib>=3.10.0",
+  "boto3>=1.37.1",
 ]
 classifiers = [
   "Programming Language :: Python",

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/src/versionhq/__init__.py

@@ -32,7 +32,7 @@ from versionhq.agent_network.formation import form_agent_network
 from versionhq.task_graph.draft import workflow
 
 
-__version__ = "1.2.2.4"
+__version__ = "1.2.2.6"
 __all__ = [
     "Agent",
 

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/src/versionhq/agent/inhouse_agents.py

@@ -38,7 +38,7 @@ vhq_formation_planner = Agent(
         "Solo is a formation where a single agent with tools, knowledge, and memory handles tasks indivudually. When self-learning mode is on - it will turn into Random formation. Typical usecase is an email agent drafts promo message for the given audience using their own knowledge.",
         "Supervising is a formation where the leader agent gives directions, while sharing its knowledge and memory with subbordinates.Subordinates can be solo agents or networks. Typical usecase is that the leader agent strategizes an outbound campaign plan and assigns components such as media mix or message creation to subordinate agents.",
         "Network is a formation where multple agents can share tasks, knowledge, and memory among network members without hierarchy. Typical usecase is that an email agent and social media agent share the product knowledge and deploy multi-channel outbound campaign. ",
-        "Random is a formation where a single agent handles tasks, asking help from other agents without sharing its memory or knowledge. Typical usecase is that an email agent drafts promo message for the given audience, asking insights on tones from other email agents which oversee other customer clusters, or an agent calls the external, third party agent to deploy the campaign. ",
+        "Random is a formation where a single agent handles tasks, asking help from other agents without sharing its memory or knowledge. Typical usecase is that an email agent drafts promo message for the given audience, asking insights on tones from other email agents which oversee other customer clusters, or an agent calls the external, third party agent to deploy the campaign.",
     ]
 )
 

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/src/versionhq/agent/model.py

@@ -35,8 +35,8 @@ class Agent(BaseModel):
     config: Optional[Dict[str, Any]] = Field(default=None, exclude=True, description="values to add to the Agent class")
 
     id: UUID4 = Field(default_factory=uuid.uuid4, frozen=True)
-    role: str = Field(description="role of the agent - used in summary and logs")
-    goal: str = Field(description="concise goal of the agent (details are set in the Task instance)")
+    role: str = Field(description="required. agent's role")
+    goal: Optional[str] = Field(default=None)
     backstory: Optional[str] = Field(default=None, description="developer prompt to the llm")
     skillsets: Optional[List[str]] = Field(default_factory=list)
     tools: Optional[List[Any]] = Field(default_factory=list)
@@ -68,10 +68,10 @@ class Agent(BaseModel):
     # llm settings cascaded to the LLM model
     llm: str | InstanceOf[LLM] | Dict[str, Any] = Field(default=None)
     func_calling_llm: str | InstanceOf[LLM] | Dict[str, Any] = Field(default=None)
-    respect_context_window: bool = Field(default=True,description="keep messages under the context window size")
+    respect_context_window: bool = Field(default=True, description="keep messages under the context window size")
     max_execution_time: Optional[int] = Field(default=None, description="max. task execution time in seconds")
     max_rpm: Optional[int] = Field(default=None, description="max. number of requests per minute")
-    llm_config: Optional[Dict[str, Any]] = Field(default=None, description="other llm config cascaded to the LLM model")
+    llm_config: Optional[Dict[str, Any]] = Field(default=None, description="other llm config cascaded to the LLM class")
 
     # # cache, error, ops handling
     # formatting_errors: int = Field(default=0, description="number of formatting errors.")
@@ -92,7 +92,7 @@ class Agent(BaseModel):
 
     @model_validator(mode="after")
     def validate_required_fields(self) -> Self:
-        required_fields = ["role", "goal"]
+        required_fields = ["role",]
         for field in required_fields:
             if getattr(self, field) is None:
                 raise ValueError(f"{field} must be provided either directly or through config")
@@ -172,7 +172,7 @@ class Agent(BaseModel):
             skills = ", ".join([item for item in self.skillsets]) if self.skillsets else ""
             tools = ", ".join([item.name for item in self.tools if hasattr(item, "name") and item.name is not None]) if self.tools else ""
             role = self.role.lower()
-            goal = self.goal.lower()
+            goal = self.goal.lower() if self.goal else ""
 
             if self.tools or self.skillsets:
                 backstory = BACKSTORY_FULL.format(role=role, goal=goal, skills=skills, tools=tools)
@@ -276,8 +276,8 @@ class Agent(BaseModel):
                 return self._set_llm_params(llm=llm, config=self.llm_config)
 
             case str():
-                llm_obj = LLM(model=llm)
-                return self._set_llm_params(llm=llm_obj, config=self.llm_config)
+                llm = LLM(model=llm)
+                return self._set_llm_params(llm=llm, config=self.llm_config)
 
             case dict():
                 model_name = llm.pop("model_name", llm.pop("deployment_name", str(llm)))
@@ -287,53 +287,21 @@ class Agent(BaseModel):
 
             case _:
                 model_name = (getattr(self.llm, "model_name") or getattr(self.llm, "deployment_name") or str(self.llm))
-                llm_obj = LLM(model=model_name if model_name else DEFAULT_MODEL_NAME)
+                llm = LLM(model=model_name if model_name else DEFAULT_MODEL_NAME)
                 llm_params = {
-                    "max_tokens": (getattr(llm, "max_tokens") or 3000),
                     "timeout": getattr(llm, "timeout", self.max_execution_time),
                     "callbacks": getattr(llm, "callbacks", None),
-                    "temperature": getattr(llm, "temperature", None),
-                    "logprobs": getattr(llm, "logprobs", None),
-                    "api_key": getattr(llm, "api_key", os.environ.get("LITELLM_API_KEY", None)),
+                    "llm_config": getattr(llm, "llm_config", None),
                     "base_url": getattr(llm, "base_url", None),
                 }
                 config = llm_params.update(self.llm_config) if self.llm_config else llm_params
-                return self._set_llm_params(llm=llm_obj, config=config)
+                return self._set_llm_params(llm=llm, config=config)
 
 
     def _set_llm_params(self, llm: LLM, config: Dict[str, Any] = None) -> LLM:
         """
         Add valid params to the LLM object.
         """
-
-        import litellm
-        from versionhq.llm.llm_vars import PARAMS
-
-        valid_config = {k: v for k, v in config.items() if v} if config else {}
-
-        if valid_config:
-            valid_keys = list()
-            try:
-                valid_keys = litellm.get_supported_openai_params(model=llm.model, custom_llm_provider=self.endpoint_provider, request_type="chat_completion")
-                if not valid_keys:
-                    valid_keys = PARAMS.get("common")
-            except:
-                valid_keys = PARAMS.get("common")
-
-            valid_keys += PARAMS.get("litellm")
-
-            for key in valid_keys:
-                if key in valid_config and valid_config[key]:
-                    val = valid_config[key]
-                    if [key == k for k, v in LLM.model_fields.items()]:
-                        setattr(llm, key, val)
-                    else:
-                        llm.other_valid_config.update({ key: val})
-
-
-        llm.timeout = self.max_execution_time if llm.timeout is None else llm.timeout
-        # llm.max_tokens = self.max_tokens if self.max_tokens else llm.max_tokens
-
         if llm.provider is None:
             provider_name = llm.model.split("/")[0]
             valid_provider = provider_name if provider_name in PROVIDERS else None
@@ -346,6 +314,12 @@ class Agent(BaseModel):
         if self.respect_context_window == False:
             llm.context_window_size = DEFAULT_CONTEXT_WINDOW_SIZE
 
+        llm.timeout = self.max_execution_time if llm.timeout is None else llm.timeout
+
+        if config:
+            llm.llm_config = {k: v for k, v in config.items() if v or v == False}
+            llm.setup_config()
+
         return llm
 
 
@@ -494,7 +468,7 @@ class Agent(BaseModel):
         Defines and executes a task when it is not given and returns TaskOutput object.
         """
 
-        if not self.goal or not self.role:
+        if not self.role:
             return None
 
         from versionhq.task.model import Task
@@ -504,7 +478,7 @@ class Agent(BaseModel):
             steps: list[str]
 
         task = Task(
-            description=f"Generate a simple result in a sentence to achieve the goal: {self.goal}. If needed, list up necessary steps in concise manner.",
+            description=f"Generate a simple result in a sentence to achieve the goal: {self.goal if self.goal else self.role}. If needed, list up necessary steps in concise manner.",
             pydantic_output=Output,
             tool_res_as_final=tool_res_as_final,
         )
@@ -595,7 +569,7 @@ class Agent(BaseModel):
 
 
     def __repr__(self):
-        return f"Agent(role={self.role}, goal={self.goal}"
+        return f"Agent(role={self.role}, id={str(self.id)}"
 
     def __str__(self):
         return super().__str__()

{versionhq-1.2.2.4 → versionhq-1.2.2.6}/src/versionhq/llm/llm_vars.py

@@ -6,30 +6,16 @@ PROVIDERS = [
     "openai",
     "gemini",
     "openrouter",
-    "huggingface",
     "anthropic",
-    "sagemaker",
     "bedrock",
-    "ollama",
-    "watson",
-    "azure",
-    "cerebras",
-    "llama",
+    "bedrock/converse",
+    "huggingface",
 ]
 
 ENDPOINT_PROVIDERS = [
     "huggingface",
 ]
 
-"""
-List of models available on the framework.
-Model names align with the LiteLLM's key names defined in the JSON URL.
-Provider names align with the custom provider or model provider names.
--> model_key = custom_provider_name/model_name
-
-Option
-litellm.pick_cheapest_chat_models_from_llm_provider(custom_llm_provider: str, n=1)
-"""
 
 MODELS = {
     "openai": [
@@ -45,6 +31,7 @@ MODELS = {
         "gemini/gemini-2.0-flash-exp",
     ],
     "anthropic": [
+        "claude-3-7-sonnet-latest",
         "claude-3-5-sonnet-20241022",
         "claude-3-5-sonnet-20240620",
         "claude-3-haiku-2024030",
@@ -53,77 +40,52 @@ MODELS = {
     ],
     "openrouter": [
         "openrouter/deepseek/deepseek-r1",
+
         "openrouter/qwen/qwen-2.5-72b-instruct",
+
         "openrouter/google/gemini-2.0-flash-thinking-exp:free",
         "openrouter/google/gemini-2.0-flash-thinking-exp-1219:free",
         "openrouter/google/gemini-2.0-flash-001",
+
         "openrouter/meta-llama/llama-3.3-70b-instruct",
         "openrouter/mistralai/mistral-large-2411",
+        "openrouter/cohere/command-r-plus",
     ],
-    "huggingface": [
-        "huggingface/qwen/qwen2.5-VL-72B-Instruct",
-    ],
-    # "sagemaker": [
-    #     "sagemaker/huggingface-text2text-flan-t5-base",
-    #     "sagemaker/huggingface-llm-gemma-7b",
-    #     "sagemaker/jumpstart-dft-meta-textgeneration-llama-2-13b",
-    #     "sagemaker/jumpstart-dft-meta-textgeneration-llama-2-70b",
-    #     "sagemaker/jumpstart-dft-meta-textgeneration-llama-3-8b",
-    #     "sagemaker/jumpstart-dft-meta-textgeneration-llama-3-70b",
-    #     "sagemaker/huggingface-llm-mistral-7b"
-    # ], #https://docs.aws.amazon.com/sagemaker/latest/dg/jumpstart-foundation-models-latest.html
-    "ollama": [
-      "ollama/llama3.1",
-      "ollama/mixtral",
-      "ollama/mixtral-8x22B-Instruct-v0.1",
-    ],
-    # "watson": [
-    #     "watsonx/meta-llama/llama-3-1-70b-instruct",
-    #     "watsonx/meta-llama/llama-3-1-8b-instruct",
-    #     "watsonx/meta-llama/llama-3-2-11b-vision-instruct",
-    #     "watsonx/meta-llama/llama-3-2-1b-instruct",
-    #     "watsonx/meta-llama/llama-3-2-90b-vision-instruct",
-    #     "watsonx/meta-llama/llama-3-405b-instruct",
-    #     "watsonx/mistral/mistral-large",
-    #     "watsonx/ibm/granite-3-8b-instruct",
-    # ],
     "bedrock": [
-        "bedrock/anthropic.claude-3-5-sonnet-20240620-v1:0",
-        "bedrock/anthropic.claude-3-sonnet-20240229-v1:0",
-        "bedrock/anthropic.claude-3-haiku-20240307-v1:0",
-        "bedrock/anthropic.claude-3-opus-20240229-v1:0",
-        "bedrock/anthropic.claude-v2",
-        "bedrock/anthropic.claude-instant-v1",
-        "bedrock/meta.llama3-1-405b-instruct-v1:0",
-        "bedrock/meta.llama3-1-70b-instruct-v1:0",
-        "bedrock/meta.llama3-1-8b-instruct-v1:0",
-        "bedrock/meta.llama3-70b-instruct-v1:0",
-        "bedrock/meta.llama3-8b-instruct-v1:0",
+        "bedrock/converse/us.meta.llama3-3-70b-instruct-v1:0",
+        "bedrock/us.meta.llama3-2-1b-instruct-v1:0",
+        "bedrock/us.meta.llama3-2-3b-instruct-v1:0",
+        "bedrock/us.meta.llama3-2-11b-instruct-v1:0",
+
+        "bedrock/mistral.mistral-7b-instruct-v0:2",
+        "bedrock/mistral.mixtral-8x7b-instruct-v0:1",
+        "bedrock/mistral.mistral-large-2407-v1:0",
+
         "bedrock/amazon.titan-text-lite-v1",
         "bedrock/amazon.titan-text-express-v1",
+        "bedrock/amazon.titan-text-premier-v1:0",
+
+        "bedrock/cohere.command-r-plus-v1:0",
+        "bedrock/cohere.command-r-v1:0",
         "bedrock/cohere.command-text-v14",
-        "bedrock/ai21.j2-mid-v1",
-        "bedrock/ai21.j2-ultra-v1",
-        "bedrock/ai21.jamba-instruct-v1:0",
-        "bedrock/meta.llama2-13b-chat-v1",
-        "bedrock/meta.llama2-70b-chat-v1",
-        "bedrock/mistral.mistral-7b-instruct-v0:2",
-        "bedrock/mistral.mixtral-8x7b-instruct-v0:1",
+        "bedrock/cohere.command-light-text-v14",
+    ],
+    "huggingface": [
+        "huggingface/qwen/qwen2.5-VL-72B-Instruct",
     ],
 }
 
 
-
-KEYS = {
+ENV_VARS = {
     "openai": ["OPENAI_API_KEY"],
     "gemini": ["GEMINI_API_KEY"],
     "anthropic": ["ANTHROPIC_API_KEY"],
     "huggingface": ["HUGGINGFACE_API_KEY", ],
-    "sagemaker": ["AWS_ACCESS_KEY_ID", "ADW_SECURET_ACCESS_KEY", "AWS_REGION_NAME"],
+    "bedrock": ["AWS_ACCESS_KEY_ID", "AWS_SECRET_ACCESS_KEY", "AWS_REGION_NAME"],
+    "sagemaker": ["AWS_ACCESS_KEY_ID", "AWS_SECRET_ACCESS_KEY", "AWS_REGION_NAME"],
 }
 
 
-
 """
 Max input token size by the model.
 """
@@ -193,8 +155,8 @@ PARAMS = {
         "response_format",
         "n",
         "stop",
-        "base_url",
-        "api_key",
+        # "base_url",
+        # "api_key",
     ],
     "openai": [
         "timeout",
@@ -216,7 +178,10 @@ PARAMS = {
     ],
     "gemini": [
         "topK",
-    ]
+    ],
+    "bedrock": {
+        "top-k",
+    }
 }