fast-agent-mcp 0.2.0__tar.gz → 0.2.2__tar.gz

{fast_agent_mcp-0.2.0 → fast_agent_mcp-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fast-agent-mcp
-Version: 0.2.0
+Version: 0.2.2
 Summary: Define, Prompt and Test MCP enabled Agents and Workflows
 Author-email: Shaun Smith <fastagent@llmindset.co.uk>, Sarmad Qadri <sarmad@lastmileai.dev>
 License:                                  Apache License

{fast_agent_mcp-0.2.0 → fast_agent_mcp-0.2.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "fast-agent-mcp"
-version = "0.2.0"
+version = "0.2.2"
 description = "Define, Prompt and Test MCP enabled Agents and Workflows"
 readme = "README.md"
 license = { file = "LICENSE" }

{fast_agent_mcp-0.2.0 → fast_agent_mcp-0.2.2}/src/mcp_agent/cli/commands/bootstrap.py

@@ -20,7 +20,6 @@ EXAMPLE_TYPES = {
         "'Building Effective Agents' paper. Some agents use the 'fetch'\n"
         "and filesystem MCP Servers.",
         "files": [
-            "agent_build.py",
             "chaining.py",
             "evaluator.py",
             "human_input.py",
@@ -70,9 +69,33 @@ def copy_example_files(example_type: str, target_dir: Path, force: bool = False)
             mount_point_dir.mkdir(parents=True)
             console.print(f"Created mount-point directory: {mount_point_dir}")
 
-    # Use examples from top-level directory
-    package_dir = Path(__file__).parent.parent.parent.parent.parent
-    source_dir = package_dir / "examples" / ("workflows" if example_type == "workflow" else f"{example_type}")
+    # Try to use examples from the installed package first, or fall back to the top-level directory
+    from importlib.resources import files
+
+    try:
+        # First try to find examples in the package resources
+        source_dir = (
+            files("mcp_agent")
+            .joinpath("resources")
+            .joinpath("examples")
+            .joinpath("workflows" if example_type == "workflow" else f"{example_type}")
+        )
+        if not source_dir.is_dir():
+            # Fall back to the top-level directory for development mode
+            package_dir = Path(__file__).parent.parent.parent.parent.parent
+            source_dir = (
+                package_dir
+                / "examples"
+                / ("workflows" if example_type == "workflow" else f"{example_type}")
+            )
+    except (ImportError, ModuleNotFoundError, ValueError):
+        # Fall back to the top-level directory if the resource finding fails
+        package_dir = Path(__file__).parent.parent.parent.parent.parent
+        source_dir = (
+            package_dir
+            / "examples"
+            / ("workflows" if example_type == "workflow" else f"{example_type}")
+        )
 
     if not source_dir.exists():
         console.print(f"[red]Error: Source directory not found: {source_dir}[/red]")

{fast_agent_mcp-0.2.0 → fast_agent_mcp-0.2.2}/src/mcp_agent/config.py

@@ -136,6 +136,18 @@ class DeepSeekSettings(BaseModel):
     model_config = ConfigDict(extra="allow", arbitrary_types_allowed=True)
 
 
+class GenericSettings(BaseModel):
+    """
+    Settings for using OpenAI models in the fast-agent application.
+    """
+
+    api_key: str | None = None
+
+    base_url: str | None = None
+
+    model_config = ConfigDict(extra="allow", arbitrary_types_allowed=True)
+
+
 class TemporalSettings(BaseModel):
     """
     Temporal settings for the fast-agent application.
@@ -250,6 +262,9 @@ class Settings(BaseSettings):
     deepseek: DeepSeekSettings | None = None
     """Settings for using DeepSeek models in the fast-agent application"""
 
+    generic: GenericSettings | None = None
+    """Settings for using Generic models in the fast-agent application"""
+
     logger: LoggerSettings | None = LoggerSettings()
     """Logger settings for the fast-agent application"""
 

{fast_agent_mcp-0.2.0 → fast_agent_mcp-0.2.2}/src/mcp_agent/llm/model_factory.py

@@ -9,6 +9,7 @@ from mcp_agent.llm.augmented_llm_passthrough import PassthroughLLM
 from mcp_agent.llm.augmented_llm_playback import PlaybackLLM
 from mcp_agent.llm.providers.augmented_llm_anthropic import AnthropicAugmentedLLM
 from mcp_agent.llm.providers.augmented_llm_deepseek import DeepSeekAugmentedLLM
+from mcp_agent.llm.providers.augmented_llm_generic import GenericAugmentedLLM
 from mcp_agent.llm.providers.augmented_llm_openai import OpenAIAugmentedLLM
 from mcp_agent.mcp.interfaces import AugmentedLLMProtocol
 
@@ -32,6 +33,7 @@ class Provider(Enum):
     OPENAI = auto()
     FAST_AGENT = auto()
     DEEPSEEK = auto()
+    GENERIC = auto()
 
 
 class ReasoningEffort(Enum):
@@ -60,6 +62,7 @@ class ModelFactory:
         "openai": Provider.OPENAI,
         "fast-agent": Provider.FAST_AGENT,
         "deepseek": Provider.DEEPSEEK,
+        "generic": Provider.GENERIC,
     }
 
     # Mapping of effort strings to enum values
@@ -116,6 +119,7 @@ class ModelFactory:
         Provider.OPENAI: OpenAIAugmentedLLM,
         Provider.FAST_AGENT: PassthroughLLM,
         Provider.DEEPSEEK: DeepSeekAugmentedLLM,
+        Provider.GENERIC: GenericAugmentedLLM,
     }
 
     # Mapping of special model names to their specific LLM classes

fast_agent_mcp-0.2.2/src/mcp_agent/llm/providers/augmented_llm_generic.py

@@ -0,0 +1,46 @@
+import os
+
+from mcp_agent.core.request_params import RequestParams
+from mcp_agent.llm.providers.augmented_llm_openai import OpenAIAugmentedLLM
+
+DEFAULT_OLLAMA_BASE_URL = "http://localhost:11434/v1"
+DEFAULT_OLLAMA_MODEL = "llama3.2:latest"
+DEFAULT_OLLAMA_API_KEY = "ollama"
+
+
+class GenericAugmentedLLM(OpenAIAugmentedLLM):
+    def __init__(self, *args, **kwargs) -> None:
+        kwargs["provider_name"] = "GenericOpenAI"  # Set provider name in kwargs
+        super().__init__(*args, **kwargs)  # Properly pass args and kwargs to parent
+
+    def _initialize_default_params(self, kwargs: dict) -> RequestParams:
+        """Initialize Deepseek-specific default parameters"""
+        chosen_model = kwargs.get("model", DEFAULT_OLLAMA_MODEL)
+
+        return RequestParams(
+            model=chosen_model,
+            systemPrompt=self.instruction,
+            parallel_tool_calls=True,
+            max_iterations=10,
+            use_history=True,
+        )
+
+    def _api_key(self) -> str:
+        config = self.context.config
+        api_key = None
+
+        if config and config.generic:
+            api_key = config.generic.api_key
+            if api_key == "<your-api-key-here>":
+                api_key = None
+
+        if api_key is None:
+            api_key = os.getenv("GENERIC_API_KEY")
+
+        return api_key or "ollama"
+
+    def _base_url(self) -> str:
+        if self.context.config and self.context.config.deepseek:
+            base_url = self.context.config.deepseek.base_url
+
+        return base_url if base_url else DEFAULT_OLLAMA_BASE_URL

{fast_agent_mcp-0.2.0 → fast_agent_mcp-0.2.2}/src/mcp_agent/llm/providers/augmented_llm_openai.py

@@ -115,7 +115,6 @@ class OpenAIAugmentedLLM(AugmentedLLM[ChatCompletionMessageParam, ChatCompletion
         self,
         message,
         request_params: RequestParams | None = None,
-        response_model: Type[ModelT] | None = None,
     ) -> List[ChatCompletionMessage]:
         """
         Process a query using an LLM and available tools.
@@ -192,16 +191,9 @@ class OpenAIAugmentedLLM(AugmentedLLM[ChatCompletionMessageParam, ChatCompletion
             self.logger.debug(f"{arguments}")
             self._log_chat_progress(self.chat_turn(), model=model)
 
-            if response_model is None:
-                executor_result = await self.executor.execute(
-                    openai_client.chat.completions.create, **arguments
-                )
-            else:
-                executor_result = await self.executor.execute(
-                    openai_client.beta.chat.completions.parse,
-                    **arguments,
-                    response_format=response_model,
-                )
+            executor_result = await self.executor.execute(
+                openai_client.chat.completions.create, **arguments
+            )
 
             response = executor_result[0]
 

fast_agent_mcp-0.2.2/src/mcp_agent/resources/examples/data-analysis/analysis-campaign.py

@@ -0,0 +1,188 @@
+import asyncio
+
+from mcp_agent.core.fastagent import FastAgent
+from mcp_agent.llm.augmented_llm import RequestParams
+
+# Create the application
+fast = FastAgent("Data Analysis & Campaign Generator")
+
+
+# Original data analysis components
+@fast.agent(
+    name="data_analysis",
+    instruction="""
+You have access to a Python 3.12 interpreter and you can use this to analyse and process data. 
+Common analysis packages such as Pandas, Seaborn and Matplotlib are already installed. 
+You can add further packages if needed.
+Data files are accessible from the /mnt/data/ directory (this is the current working directory).
+Visualisations should be saved as .png files in the current working directory.
+Extract key insights that would be compelling for a social media campaign.
+""",
+    servers=["interpreter"],
+    request_params=RequestParams(maxTokens=8192),
+    model="sonnet",
+)
+@fast.agent(
+    "evaluator",
+    """You are collaborating with a Data Analysis tool that has the capability to analyse data and produce visualisations.
+    You must make sure that the tool has:
+     - Considered the best way for a Human to interpret the data
+     - Produced insightful visualisations.
+     - Provided a high level summary report for the Human.
+     - Has had its findings challenged, and justified
+     - Extracted compelling insights suitable for social media promotion
+    """,
+    request_params=RequestParams(maxTokens=8192),
+    model="gpt-4o",
+)
+@fast.evaluator_optimizer(
+    "analysis_tool",
+    generator="data_analysis",
+    evaluator="evaluator",
+    max_refinements=3,
+    min_rating="EXCELLENT",
+)
+# Research component using Brave search
+@fast.agent(
+    "context_researcher",
+    """You are a research specialist who provides cultural context for different regions.
+    For any given data insight and target language/region, research:
+    1. Cultural sensitivities related to presenting this type of data
+    2. Local social media trends and preferences
+    3. Region-specific considerations for marketing campaigns
+    
+    Always provide actionable recommendations for adapting content to each culture.
+    """,
+    servers=["fetch", "brave"],  # Using the fetch MCP server for Brave search
+    request_params=RequestParams(temperature=0.3),
+    model="gpt-4o",
+)
+# Social media content generator
+@fast.agent(
+    "campaign_generator",
+    """Generate engaging social media content based on data insights.
+    Create compelling, shareable content that:
+    - Highlights key research findings in an accessible way
+    - Uses appropriate tone for the platform (Twitter/X, LinkedIn, Instagram, etc.)
+    - Is concise and impactful
+    - Includes suggested hashtags and posting schedule
+    
+    Format your response with clear sections for each platform.
+    Save different campaign elements as separate files in the current directory.
+    """,
+    servers=["filesystem"],  # Using filesystem MCP server to save files
+    request_params=RequestParams(temperature=0.7),
+    model="sonnet",
+    use_history=False,
+)
+# Translation agents with cultural adaptation
+@fast.agent(
+    "translate_fr",
+    """Translate social media content to French with cultural adaptation.
+    Consider French cultural norms, expressions, and social media preferences.
+    Ensure the translation maintains the impact of the original while being culturally appropriate.
+    Save the translated content to a file with appropriate naming.
+    """,
+    model="haiku",
+    use_history=False,
+    servers=["filesystem"],
+)
+@fast.agent(
+    "translate_es",
+    """Translate social media content to Spanish with cultural adaptation.
+    Consider Spanish-speaking cultural contexts, expressions, and social media preferences.
+    Ensure the translation maintains the impact of the original while being culturally appropriate.
+    Save the translated content to a file with appropriate naming.
+    """,
+    model="haiku",
+    use_history=False,
+    servers=["filesystem"],
+)
+@fast.agent(
+    "translate_de",
+    """Translate social media content to German with cultural adaptation.
+    Consider German cultural norms, expressions, and social media preferences.
+    Ensure the translation maintains the impact of the original while being culturally appropriate.
+    Save the translated content to a file with appropriate naming.
+    """,
+    model="haiku",
+    use_history=False,
+    servers=["filesystem"],
+)
+@fast.agent(
+    "translate_ja",
+    """Translate social media content to Japanese with cultural adaptation.
+    Consider Japanese cultural norms, expressions, and social media preferences.
+    Ensure the translation maintains the impact of the original while being culturally appropriate.
+    Save the translated content to a file with appropriate naming.
+    """,
+    model="haiku",
+    use_history=False,
+    servers=["filesystem"],
+)
+# Parallel workflow for translations
+@fast.parallel(
+    "translate_campaign",
+    instruction="Translates content to French, Spanish, German and Japanese. Supply the content to translate, translations will be saved to the filesystem.",
+    fan_out=["translate_fr", "translate_es", "translate_de", "translate_ja"],
+    include_request=True,
+)
+# Cultural sensitivity review agent
+@fast.agent(
+    "cultural_reviewer",
+    """Review all translated content for cultural sensitivity and appropriateness.
+    For each language version, evaluate:
+    - Cultural appropriateness
+    - Potential misunderstandings or sensitivities
+    - Effectiveness for the target culture
+    
+    Provide specific recommendations for any needed adjustments and save a review report.
+    """,
+    servers=["filesystem"],
+    request_params=RequestParams(temperature=0.2),
+)
+# Campaign optimization workflow
+@fast.evaluator_optimizer(
+    "campaign_optimizer",
+    generator="campaign_generator",
+    evaluator="cultural_reviewer",
+    max_refinements=2,
+    min_rating="EXCELLENT",
+)
+# Main workflow orchestration
+@fast.orchestrator(
+    "research_campaign_creator",
+    instruction="""
+    Create a complete multi-lingual social media campaign based on data analysis results.
+    The workflow will:
+    1. Analyze the provided data and extract key insights
+    2. Research cultural contexts for target languages
+    3. Generate appropriate social media content
+    4. Translate and culturally adapt the content
+    5. Review and optimize all materials
+    6. Save all campaign elements to files
+    """,
+    agents=[
+        "analysis_tool",
+        "context_researcher",
+        "campaign_optimizer",
+        "translate_campaign",
+    ],
+    model="sonnet",  # Using a more capable model for orchestration
+    request_params=RequestParams(maxTokens=8192),
+    plan_type="full",
+)
+async def main() -> None:
+    # Use the app's context manager
+    print(
+        "WARNING: This workflow will likely run for >10 minutes and consume a lot of tokens. Press Enter to accept the default prompt and proceed"
+    )
+
+    async with fast.run() as agent:
+        await agent.research_campaign_creator.prompt(
+            default_prompt="Analyze the CSV file in the current directory and create a comprehensive multi-lingual social media campaign based on the findings. Save all campaign elements as separate files."
+        )
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

fast_agent_mcp-0.2.2/src/mcp_agent/resources/examples/data-analysis/analysis.py

@@ -0,0 +1,65 @@
+import asyncio
+
+from mcp_agent.core.fastagent import FastAgent
+from mcp_agent.llm.augmented_llm import RequestParams
+
+# Create the application
+fast = FastAgent("Data Analysis (Roots)")
+
+
+# The sample data is under Database Contents License (DbCL) v1.0.
+# Available here : https://www.kaggle.com/datasets/pavansubhasht/ibm-hr-analytics-attrition-dataset
+
+
+@fast.agent(
+    name="data_analysis",
+    instruction="""
+You have access to a Python 3.12 interpreter and you can use this to analyse and process data. 
+Common analysis packages such as Pandas, Seaborn and Matplotlib are already installed. 
+You can add further packages if needed.
+Data files are accessible from the /mnt/data/ directory (this is the current working directory).
+Visualisations should be saved as .png files in the current working directory.
+""",
+    servers=["interpreter"],
+    request_params=RequestParams(maxTokens=8192),
+)
+async def main() -> None:
+    # Use the app's context manager
+    async with fast.run() as agent:
+        await agent(
+            "There is a csv file in the current directory. "
+            "Analyse the file, produce a detailed description of the data, and any patterns it contains.",
+        )
+        await agent(
+            "Consider the data, and how to usefully group it for presentation to a Human. Find insights, using the Python Interpreter as needed.\n"
+            "Use MatPlotLib to produce insightful visualisations. Save them as '.png' files in the current directory. Be sure to run the code and save the files.\n"
+            "Produce a summary with major insights to the data",
+        )
+        await agent()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())
+
+
+############################################################################################################
+# Example of evaluator/optimizer flow
+############################################################################################################
+# @fast.agent(
+#     "evaluator",
+#     """You are collaborating with a Data Analysis tool that has the capability to analyse data and produce visualisations.
+#     You must make sure that the tool has:
+#      - Considered the best way for a Human to interpret the data
+#      - Produced insightful visualasions.
+#      - Provided a high level summary report for the Human.
+#      - Has had its findings challenged, and justified
+#     """,
+#     request_params=RequestParams(maxTokens=8192),
+# )
+# @fast.evaluator_optimizer(
+#     "analysis_tool",
+#     generator="data_analysis",
+#     evaluator="evaluator",
+#     max_refinements=3,
+#     min_rating="EXCELLENT",
+# )

fast_agent_mcp-0.2.2/src/mcp_agent/resources/examples/data-analysis/fastagent.config.yaml

@@ -0,0 +1,41 @@
+default_model: sonnet
+
+# on windows, adjust the mount point to be the full path e.g. x:/temp/data-analysis/mount-point:/mnt/data/
+
+mcp:
+  servers:
+    interpreter:
+      command: "docker"
+      args:
+        [
+          "run",
+          "-i",
+          "--rm",
+          "--pull=always",
+          "-v",
+          "./mount-point:/mnt/data/",
+          "ghcr.io/evalstate/mcp-py-repl:latest",
+        ]
+      roots:
+        - uri: "file://./mount-point/"
+          name: "test_data"
+          server_uri_alias: "file:///mnt/data/"
+    filesystem:
+      # On windows update the command and arguments to use `node` and the absolute path to the server.
+      # Use `npm i -g @modelcontextprotocol/server-filesystem` to install the server globally.
+      # Use `npm -g root` to find the global node_modules path.`
+      # command: "node"
+      # args: ["c:/Program Files/nodejs/node_modules/@modelcontextprotocol/server-filesystem/dist/index.js","."]
+      command: "npx"
+      args: ["-y", "@modelcontextprotocol/server-filesystem", "./mount-point/"]
+    fetch:
+      command: "uvx"
+      args: ["mcp-server-fetch"]
+    brave:
+      # On windows replace the command and args line to use `node` and the absolute path to the server.
+      # Use `npm i -g @modelcontextprotocol/server-brave-search` to install the server globally.
+      # Use `npm -g root` to find the global node_modules path.`
+      # command: "node"
+      # args: ["c:/Program Files/nodejs/node_modules/@modelcontextprotocol/server-brave-search/dist/index.js"]
+      command: "npx"
+      args: ["-y", "@modelcontextprotocol/server-brave-search"]