Skip to content

utils

create_agent_table(agents, agent_dict=None)

Create a Gradio Dataframe component for agent configuration.

Parameters:

Name Type Description Default
agents List[AgentConfig]

List of agent configurations

 required
agent_dict dict | None

An optional mapping from agent name to MCP server tool data produced by cycle_through_tools and is used to populate the MCP Server Tools column.

 None

Returns:

Type Description
Dataframe

The configured Gradio Dataframe component representing agents
Source code in src/mada/interfaces/gradio/utils.py 
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
def create_agent_table(
    agents: List[AgentConfig],
    agent_dict: dict | None = None,
) -> gr.Dataframe:
    """
    Create a Gradio Dataframe component for agent configuration.

    Args:
        agents: List of agent configurations
        agent_dict: An optional mapping from agent name to MCP server tool data
            produced by `cycle_through_tools` and is used to populate the MCP
            Server Tools column.

    Returns:
        The configured Gradio Dataframe component representing agents
    """

    # Headers for the agent table
    headers = [
        "Agent Name",
        "Description",
        "Domain",
        "MCP Servers",
        "MCP Server Tools",
        "Instructions",
    ]

    # Convert agents to table rows
    table_rows = []
    for agent in agents:
        row = [
            agent.agent_name,
            agent.description,
            getattr(agent, "domain", ""),
            ", ".join(agent.mcp_servers) if agent.mcp_servers else "",
            "Need to connect first..."
            if agent_dict is None
            else str(agent_dict[agent.agent_name]),
            agent.instructions,
        ]
        table_rows.append(row)

    return gr.Dataframe(
        headers=headers,
        datatype=["str"] * len(headers),
        row_count=len(table_rows),
        column_count=len(headers),
        label="Agents",
        value=table_rows,
        interactive=False,
        elem_id="my_df",
        column_widths=[f"{math.floor(100 / len(row))}%"] * len(row),
    )

cycle_through_tools(agents, sim_server=None, return_tool=False)

Build a mapping of agents to MCP servers and their available tool names.

Iterates through each agent, inspects its configured MCP tools, and collects the function names exposed by each MCP server. Optionally, the function can search for and return a specific tool function named in_situ_viz from the specified simulation server.

Parameters:

Name Type Description Default
agents List[AgentConfig]

List of agent configurations

 required
sim_server str | None

The name of the MCP server to inspect when return_tool is True. MCP servers with different names are skipped in that case.

 None
return_tool bool

If False, return a nested dictionary mapping agent names to MCP server names and tool names. If True, search for a tool function named in_situ_viz and return that function object if found.

 False

Returns:

Type Description
Any

If return_tool is False, returns a dictionary of the form:

{ agent_name: { mcp_server_name: [tool_name, ...] } }
Any

If return_tool is True, returns the function object whose name is
Any

in_situ_viz from the specified sim_server, or None if no such
Any

function is found.
Source code in src/mada/interfaces/gradio/utils.py 
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
def cycle_through_tools(
    agents: List[AgentConfig],
    sim_server: str | None = None,
    return_tool: bool = False,
) -> Any:
    """
    Build a mapping of agents to MCP servers and their available tool names.

    Iterates through each agent, inspects its configured MCP tools, and
    collects the function names exposed by each MCP server. Optionally,
    the function can search for and return a specific tool function named
    `in_situ_viz` from the specified simulation server.

    Args:
        agents: List of agent configurations
        sim_server: The name of the MCP server to inspect when `return_tool`
            is `True`. MCP servers with different names are skipped in that case.
        return_tool: If `False`, return a nested dictionary mapping agent
            names to MCP server names and tool names. If `True`, search for
            a tool function named `in_situ_viz` and return that function object
            if found.

    Returns:
        If `return_tool` is `False`, returns a dictionary of the form:

            {
                agent_name: {
                    mcp_server_name: [tool_name, ...]
                }
            }

        If `return_tool` is `True`, returns the function object whose name is
        `in_situ_viz` from the specified `sim_server`, or `None` if no such
        function is found.
    """
    agent_dict = {}
    for agent in agents:
        agent_name = getattr(agent, "name", None)
        agent_dict[agent_name] = {}

        for mcp_tool in getattr(agent, "mcp_tools", []):
            mcp_server_name = getattr(mcp_tool, "name", None)
            agent_dict[agent_name][mcp_server_name] = []
            if return_tool and mcp_server_name != sim_server:
                continue

            for fn in getattr(mcp_tool, "_functions", []):
                mcp_tool_name = getattr(fn, "name", None)
                agent_dict[agent_name][mcp_server_name].append(mcp_tool_name)
                if return_tool and mcp_tool_name == "in_situ_viz":
                    return fn

    if return_tool:
        return None
    else:
        return agent_dict