The Lab via PyGEAI Code Samples - Tools

Below, you can find code snippets that show how to use PyGEAI The Lab's module to create, retrieve, list, update, publish and delete Tools.

1. Create Tool

Initiate AILabManager to create a new Tool. Define a set of parameters user input including key, data_type, description, is_required, internal configuration parameters (type, key, from_secret) and construct the Tool, assigning key metadata: name, description, scope.

from pygeai.lab.managers import AILabManager
from pygeai.lab.models import Tool, ToolParameter
parameters = 
    ToolParameter(
        key="input",
        data_type="String",
        description="some input that the tool needs.",
        is_required=True
    ),
    ToolParameter(
        key="some_nonsensitive_id",
        data_type="String",
        description="Configuration that is static, in the sense that whenever the tool is used, the value for this parameter is configured here. The llm will not know about it.",
        is_required=True,
        type="config",
        from_secret=False,
        value="b001e30b4016001f5f76b9ae9215ac40"
    ),
    ToolParameter(
        key="api_token",
        data_type="String",
        description="Configuration that is static, but it is sensitive information . The value is stored in secret-manager",
        is_required=True,
        type="config",
        value="0cd84dc7-f3f5-4a03-9288-cdfd8d72fde1"
    )

tool = Tool(
    name="sample_tool_v5",
    description="a builtin tool that does something but really does nothing cos it does not exist.",
    scope="builtin",
    parameters=parameters
)
manager = AILabManager()
result = manager.create_tool(
    tool=tool,
    automatic_publish=False
)
print(f"Created tool: {result.name}, ID: {result.id}")
print(f"Description: {result.description}")
print(f"Messages: {result.messages}")

2. Get Tool

Start AILabManager to Get a Tool by its ID (tool_id). Apply FilterSettings, including revision, version, allow_drafts to request the Tool.

from pygeai.lab.managers import AILabManager
from pygeai.lab.models import FilterSettings, Tool
manager = AILabManager()
filter_settings = FilterSettings(
    revision="0",
    version="0",
    allow_drafts=True
)
result = manager.get_tool(
    tool_id="affd8ede-97c6-4083-b1f6-2b463ad4891e",
    filter_settings=filter_settings
)
print(f"Retrieved tool: {result.name}, ID: {result.id}")
print(f"Description: {result.description}")
print(f"Messages: {result.messages}")

3. List Tool

Use AILabManager to list a Tool with specific filter settings: id, count, access_scope, allow_drafts, scope, allow_external.

You can retrieve up to 100 tools that are public, built-in.

from pygeai.lab.managers import AILabManager
from pygeai.lab.models import FilterSettings
manager = AILabManager()
filter_settings = FilterSettings(
    id="",
    count="100",
    access_scope="public",
    allow_drafts=True,
    scope="builtin",
    allow_external=True
)
result = manager.list_tools(
    filter_settings=filter_settings
)
print(f"Found {len(result)} tools:")
for tool in result:
    print(f"Tool: {tool.name}, ID: {tool.id}, Scope: {tool.scope}")

4. Update Tool

Start AILabManager to update an existing Tool. Import the information and then define parameters key, data_type, description, is_required, internal configuration parameters (type, key, from_secret) and Tool metadad name, description, scope. Update the Tool in the system.

from pygeai.lab.managers import AILabManager
from pygeai.lab.models import Tool, ToolParameter
parameters = 
    ToolParameter(
        key="input",
        data_type="String",
        description="some input that the tool needs. ",
        is_required=True
    ),
    ToolParameter(
        key="some_nonsensitive_id",
        data_type="String",
        description="Configuration that is static, in the sense that whenever the tool is used, the value for this parameter is configured here. The llm will not know about it.",
        is_required=True,
        type="config",
        from_secret=False,
        value="b001e30b4016001f5f76b9ae9215ac40"
    ),
    ToolParameter(
        key="api_token",
        data_type="String",
        description="Configuration that is static, but it is sensitive information . The value is stored in secret-manager",
        is_required=True,
        type="config",
        from_secret=True,
        value="0cd84dc7-f3f5-4a03-9288-cdfd8d72fde1"
    )

tool = Tool(
    id="affd8ede-97c6-4083-b1f6-2b463ad4891e",
    name="sample tool V5",
    description="a builtin tool that does something but really does nothing cos it does not exist. UPDATED",
    scope="builtin",
    parameters=parameters
)
manager = AILabManager()
result = manager.update_tool(
    tool=tool,
    automatic_publish=False,
    upsert=False
)
print(f"Updated tool: {result.name}, ID: {result.id}")
print(f"Description: {result.description}")
print(f"Messages: {result.messages}")

5. Publish Tool

Initiate AILabManager interface to publish a specific revision of an existing Tool. Call manager.publish_tool_revision with the tool_id and its revision number to activate it.

from pygeai.lab.managers import AILabManager
from pygeai.lab.models import Tool
manager = AILabManager()
result = manager.publish_tool_revision(
    tool_id="affd8ede-97c6-4083-b1f6-2b463ad4891e",
    revision="1"
)
print(f"Published tool: {result.name}, ID: {result.id}, Revision: {result.revision}")

6. Delete a Tool

Use AILabManager to proceed to delete a Tool.

from pygeai.core.base.responses import EmptyResponse
from pygeai.lab.managers import AILabManager
manager = AILabManager(api_key="your-api-key")
result = manager.delete_tool(
    tool_id="affd8ede-97c6-4083-b1f6-2b463ad4891e"
)
if isinstance(result, EmptyResponse):
    print("Tool deleted successfully")
else:
    print("Errors:", result.errors)
result = manager.delete_tool(
    tool_name="sample tool V5"
)
print("Tool deleted successfully")