Skip to content

datarobot_genai.drmcputils.exceptions

exceptions

Custom exceptions for DataRobot tools.

ToolErrorKind

Bases: StrEnum

High-level category for tool failures (logging and MCP-friendly).

SCHEMA — tool call arguments do not match the tool input schema (MCP/JSON shape, types, required fields). Use VALIDATION for domain rules enforced inside the tool body.

Source code in datarobot_genai/drmcputils/exceptions.py 
20
21
22
23
24
25
26
27
28
29
30
31
32
class ToolErrorKind(StrEnum):
    """High-level category for tool failures (logging and MCP-friendly).

    ``SCHEMA`` — tool call arguments do not match the tool input schema (MCP/JSON shape, types,
    required fields). Use ``VALIDATION`` for domain rules enforced inside the tool body.
    """

    SCHEMA = "schema"
    VALIDATION = "validation"
    AUTHENTICATION = "authentication"
    NOT_FOUND = "not_found"
    UPSTREAM = "upstream"
    INTERNAL = "internal"

ToolError

Bases: Exception

Drop-in replacement for fastmcp.exceptions.ToolError.

This exception provides full compatibility with fastmcp's ToolError while removing the dependency on the fastmcp package. It should be raised when a tool encounters an error during execution that prevents it from completing its task.

The MCP framework expects tools to raise ToolError for user-facing errors that should be reported back to the client. This is distinct from other exceptions which might indicate programming errors or system failures.

Attributes
message: The error message to be displayed to the user.
kind: Semantic category of the failure (defaults to INTERNAL). Use ``ToolErrorKind.SCHEMA``
    when MCP tool-call JSON or types do not match the declared input schema.
args: Tuple containing the error message (for compatibility with base Exception).
Examples
>>> raise ToolError("Failed to create deployment: No prediction servers available")
>>> raise ToolError(f"Job failed with status {job.status}", kind=ToolErrorKind.UPSTREAM)
>>> try:
...     risky_operation()
... except ValueError as e:
...     raise ToolError(f"Invalid input: {e}", kind=ToolErrorKind.VALIDATION)
>>> raise ToolError("Arguments do not match tool schema", kind=ToolErrorKind.SCHEMA)
Source code in datarobot_genai/drmcputils/exceptions.py 
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
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
class ToolError(Exception):
    """Drop-in replacement for fastmcp.exceptions.ToolError.

    This exception provides full compatibility with fastmcp's ToolError while
    removing the dependency on the fastmcp package. It should be raised when
    a tool encounters an error during execution that prevents it from completing
    its task.

    The MCP framework expects tools to raise ToolError for user-facing errors
    that should be reported back to the client. This is distinct from other
    exceptions which might indicate programming errors or system failures.

    Attributes
    ----------
        message: The error message to be displayed to the user.
        kind: Semantic category of the failure (defaults to INTERNAL). Use ``ToolErrorKind.SCHEMA``
            when MCP tool-call JSON or types do not match the declared input schema.
        args: Tuple containing the error message (for compatibility with base Exception).

    Examples
    --------
        >>> raise ToolError("Failed to create deployment: No prediction servers available")
        >>> raise ToolError(f"Job failed with status {job.status}", kind=ToolErrorKind.UPSTREAM)
        >>> try:
        ...     risky_operation()
        ... except ValueError as e:
        ...     raise ToolError(f"Invalid input: {e}", kind=ToolErrorKind.VALIDATION)
        >>> raise ToolError("Arguments do not match tool schema", kind=ToolErrorKind.SCHEMA)
    """

    def __init__(self, message: str, *, kind: ToolErrorKind = ToolErrorKind.INTERNAL) -> None:
        """Initialize the ToolError with a descriptive message.

        Args:
            message: A clear, user-friendly description of what went wrong.
                This message will be displayed to the end user.
            kind: Category used for routing, metrics, and MCP error strings.
        """
        self.message = message
        self.kind = kind
        super().__init__(message)

    def __str__(self) -> str:
        """Return the error message."""
        return self.message

    def __repr__(self) -> str:
        """Return a string representation of the exception."""
        return f"{self.__class__.__name__}({self.message!r}, kind={self.kind!r})"

__init__ 

__init__(message: str, *, kind: ToolErrorKind = ToolErrorKind.INTERNAL) -> None

Initialize the ToolError with a descriptive message.

Parameters:

Name Type Description Default
message str

A clear, user-friendly description of what went wrong. This message will be displayed to the end user.

 required
kind ToolErrorKind

Category used for routing, metrics, and MCP error strings.

 INTERNAL
Source code in datarobot_genai/drmcputils/exceptions.py 
65
66
67
68
69
70
71
72
73
74
75
def __init__(self, message: str, *, kind: ToolErrorKind = ToolErrorKind.INTERNAL) -> None:
    """Initialize the ToolError with a descriptive message.

    Args:
        message: A clear, user-friendly description of what went wrong.
            This message will be displayed to the end user.
        kind: Category used for routing, metrics, and MCP error strings.
    """
    self.message = message
    self.kind = kind
    super().__init__(message)

__str__ 

__str__() -> str

Return the error message.

Source code in datarobot_genai/drmcputils/exceptions.py 
77
78
79
def __str__(self) -> str:
    """Return the error message."""
    return self.message

__repr__ 

__repr__() -> str

Return a string representation of the exception.

Source code in datarobot_genai/drmcputils/exceptions.py 
81
82
83
def __repr__(self) -> str:
    """Return a string representation of the exception."""
    return f"{self.__class__.__name__}({self.message!r}, kind={self.kind!r})"