feat: extract mcp tool usage (#31802)

Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>

api/core/tools/mcp_tool/tool.py

@@ -3,8 +3,8 @@ from __future__ import annotations
 import base64
 import json
 import logging
-from collections.abc import Generator
+from collections.abc import Generator, Mapping
-from typing import Any
+from typing import Any, cast
 
 from core.mcp.auth_client import MCPClientWithAuthRetry
 from core.mcp.error import MCPConnectionError
@@ -17,6 +17,7 @@ from core.mcp.types import (
     TextContent,
     TextResourceContents,
 )
+from core.model_runtime.entities.llm_entities import LLMUsage, LLMUsageMetadata
 from core.tools.__base.tool import Tool
 from core.tools.__base.tool_runtime import ToolRuntime
 from core.tools.entities.tool_entities import ToolEntity, ToolInvokeMessage, ToolProviderType
@@ -46,6 +47,7 @@ class MCPTool(Tool):
         self.headers = headers or {}
         self.timeout = timeout
         self.sse_read_timeout = sse_read_timeout
+        self._latest_usage = LLMUsage.empty_usage()
 
     def tool_provider_type(self) -> ToolProviderType:
         return ToolProviderType.MCP
@@ -59,6 +61,10 @@ class MCPTool(Tool):
         message_id: str | None = None,
     ) -> Generator[ToolInvokeMessage, None, None]:
         result = self.invoke_remote_mcp_tool(tool_parameters)
+
+        # Extract usage metadata from MCP protocol's _meta field
+        self._latest_usage = self._derive_usage_from_result(result)
+
         # handle dify tool output
         for content in result.content:
             if isinstance(content, TextContent):
@@ -120,6 +126,99 @@ class MCPTool(Tool):
         for item in json_list:
             yield self.create_json_message(item)
 
+    @property
+    def latest_usage(self) -> LLMUsage:
+        return self._latest_usage
+
+    @classmethod
+    def _derive_usage_from_result(cls, result: CallToolResult) -> LLMUsage:
+        """
+        Extract usage metadata from MCP tool result's _meta field.
+
+        The MCP protocol's _meta field (aliased as 'meta' in Python) can contain
+        usage information such as token counts, costs, and other metadata.
+
+        Args:
+            result: The CallToolResult from MCP tool invocation
+
+        Returns:
+            LLMUsage instance with values from meta or empty_usage if not found
+        """
+        # Extract usage from the meta field if present
+        if result.meta:
+            usage_dict = cls._extract_usage_dict(result.meta)
+            if usage_dict is not None:
+                return LLMUsage.from_metadata(cast(LLMUsageMetadata, cast(object, dict(usage_dict))))
+
+        return LLMUsage.empty_usage()
+
+    @classmethod
+    def _extract_usage_dict(cls, payload: Mapping[str, Any]) -> Mapping[str, Any] | None:
+        """
+        Recursively search for usage dictionary in the payload.
+
+        The MCP protocol's _meta field can contain usage data in various formats:
+        - Direct usage field: {"usage": {...}}
+        - Nested in metadata: {"metadata": {"usage": {...}}}
+        - Or nested within other fields
+
+        Args:
+            payload: The payload to search for usage data
+
+        Returns:
+            The usage dictionary if found, None otherwise
+        """
+        # Check for direct usage field
+        usage_candidate = payload.get("usage")
+        if isinstance(usage_candidate, Mapping):
+            return usage_candidate
+
+        # Check for metadata nested usage
+        metadata_candidate = payload.get("metadata")
+        if isinstance(metadata_candidate, Mapping):
+            usage_candidate = metadata_candidate.get("usage")
+            if isinstance(usage_candidate, Mapping):
+                return usage_candidate
+
+        # Check for common token counting fields directly in payload
+        # Some MCP servers may include token counts directly
+        if "total_tokens" in payload or "prompt_tokens" in payload or "completion_tokens" in payload:
+            usage_dict: dict[str, Any] = {}
+            for key in (
+                "prompt_tokens",
+                "completion_tokens",
+                "total_tokens",
+                "prompt_unit_price",
+                "completion_unit_price",
+                "total_price",
+                "currency",
+                "prompt_price_unit",
+                "completion_price_unit",
+                "prompt_price",
+                "completion_price",
+                "latency",
+                "time_to_first_token",
+                "time_to_generate",
+            ):
+                if key in payload:
+                    usage_dict[key] = payload[key]
+            if usage_dict:
+                return usage_dict
+
+        # Recursively search through nested structures
+        for value in payload.values():
+            if isinstance(value, Mapping):
+                found = cls._extract_usage_dict(value)
+                if found is not None:
+                    return found
+            elif isinstance(value, list) and not isinstance(value, (str, bytes, bytearray)):
+                for item in value:
+                    if isinstance(item, Mapping):
+                        found = cls._extract_usage_dict(item)
+                        if found is not None:
+                            return found
+        return None
+
     def fork_tool_runtime(self, runtime: ToolRuntime) -> MCPTool:
         return MCPTool(
             entity=self.entity,

api/tests/unit_tests/tools/test_mcp_tool.py

@@ -1,4 +1,5 @@
 import base64
+from decimal import Decimal
 from unittest.mock import Mock, patch
 
 import pytest
@@ -9,8 +10,10 @@ from core.mcp.types import (
     CallToolResult,
     EmbeddedResource,
     ImageContent,
+    TextContent,
     TextResourceContents,
 )
+from core.model_runtime.entities.llm_entities import LLMUsage
 from core.tools.__base.tool_runtime import ToolRuntime
 from core.tools.entities.common_entities import I18nObject
 from core.tools.entities.tool_entities import ToolEntity, ToolIdentity, ToolInvokeMessage
@@ -120,3 +123,231 @@ class TestMCPToolInvoke:
         # Validate values
         values = {m.message.variable_name: m.message.variable_value for m in var_msgs}
         assert values == {"a": 1, "b": "x"}
+
+
+class TestMCPToolUsageExtraction:
+    """Test usage metadata extraction from MCP tool results."""
+
+    def test_extract_usage_dict_from_direct_usage_field(self) -> None:
+        """Test extraction when usage is directly in meta.usage field."""
+        meta = {
+            "usage": {
+                "prompt_tokens": 100,
+                "completion_tokens": 50,
+                "total_tokens": 150,
+                "total_price": "0.001",
+                "currency": "USD",
+            }
+        }
+        usage_dict = MCPTool._extract_usage_dict(meta)
+        assert usage_dict is not None
+        assert usage_dict["prompt_tokens"] == 100
+        assert usage_dict["completion_tokens"] == 50
+        assert usage_dict["total_tokens"] == 150
+        assert usage_dict["total_price"] == "0.001"
+        assert usage_dict["currency"] == "USD"
+
+    def test_extract_usage_dict_from_nested_metadata(self) -> None:
+        """Test extraction when usage is nested in meta.metadata.usage."""
+        meta = {
+            "metadata": {
+                "usage": {
+                    "prompt_tokens": 200,
+                    "completion_tokens": 100,
+                    "total_tokens": 300,
+                }
+            }
+        }
+        usage_dict = MCPTool._extract_usage_dict(meta)
+        assert usage_dict is not None
+        assert usage_dict["prompt_tokens"] == 200
+        assert usage_dict["total_tokens"] == 300
+
+    def test_extract_usage_dict_from_flat_token_fields(self) -> None:
+        """Test extraction when token counts are directly in meta."""
+        meta = {
+            "prompt_tokens": 150,
+            "completion_tokens": 75,
+            "total_tokens": 225,
+            "currency": "EUR",
+        }
+        usage_dict = MCPTool._extract_usage_dict(meta)
+        assert usage_dict is not None
+        assert usage_dict["prompt_tokens"] == 150
+        assert usage_dict["completion_tokens"] == 75
+        assert usage_dict["total_tokens"] == 225
+        assert usage_dict["currency"] == "EUR"
+
+    def test_extract_usage_dict_recursive(self) -> None:
+        """Test recursive search through nested structures."""
+        meta = {
+            "custom": {
+                "nested": {
+                    "usage": {
+                        "total_tokens": 500,
+                        "prompt_tokens": 300,
+                        "completion_tokens": 200,
+                    }
+                }
+            }
+        }
+        usage_dict = MCPTool._extract_usage_dict(meta)
+        assert usage_dict is not None
+        assert usage_dict["total_tokens"] == 500
+
+    def test_extract_usage_dict_from_list(self) -> None:
+        """Test extraction from nested list structures."""
+        meta = {
+            "items": [
+                {"usage": {"total_tokens": 100}},
+                {"other": "data"},
+            ]
+        }
+        usage_dict = MCPTool._extract_usage_dict(meta)
+        assert usage_dict is not None
+        assert usage_dict["total_tokens"] == 100
+
+    def test_extract_usage_dict_returns_none_when_missing(self) -> None:
+        """Test that None is returned when no usage data is present."""
+        meta = {"other": "data", "custom": {"nested": {"value": 123}}}
+        usage_dict = MCPTool._extract_usage_dict(meta)
+        assert usage_dict is None
+
+    def test_extract_usage_dict_empty_meta(self) -> None:
+        """Test with empty meta dict."""
+        usage_dict = MCPTool._extract_usage_dict({})
+        assert usage_dict is None
+
+    def test_derive_usage_from_result_with_meta(self) -> None:
+        """Test _derive_usage_from_result with populated meta."""
+        meta = {
+            "usage": {
+                "prompt_tokens": 100,
+                "completion_tokens": 50,
+                "total_tokens": 150,
+                "total_price": "0.0015",
+                "currency": "USD",
+            }
+        }
+        result = CallToolResult(content=[], _meta=meta)
+        usage = MCPTool._derive_usage_from_result(result)
+
+        assert isinstance(usage, LLMUsage)
+        assert usage.prompt_tokens == 100
+        assert usage.completion_tokens == 50
+        assert usage.total_tokens == 150
+        assert usage.total_price == Decimal("0.0015")
+        assert usage.currency == "USD"
+
+    def test_derive_usage_from_result_without_meta(self) -> None:
+        """Test _derive_usage_from_result with no meta returns empty usage."""
+        result = CallToolResult(content=[], meta=None)
+        usage = MCPTool._derive_usage_from_result(result)
+
+        assert isinstance(usage, LLMUsage)
+        assert usage.total_tokens == 0
+        assert usage.prompt_tokens == 0
+        assert usage.completion_tokens == 0
+
+    def test_derive_usage_from_result_calculates_total_tokens(self) -> None:
+        """Test that total_tokens is calculated when missing."""
+        meta = {
+            "usage": {
+                "prompt_tokens": 100,
+                "completion_tokens": 50,
+                # total_tokens is missing
+            }
+        }
+        result = CallToolResult(content=[], _meta=meta)
+        usage = MCPTool._derive_usage_from_result(result)
+
+        assert usage.total_tokens == 150  # 100 + 50
+        assert usage.prompt_tokens == 100
+        assert usage.completion_tokens == 50
+
+    def test_invoke_sets_latest_usage_from_meta(self) -> None:
+        """Test that _invoke sets _latest_usage from result meta."""
+        tool = _make_mcp_tool()
+        meta = {
+            "usage": {
+                "prompt_tokens": 200,
+                "completion_tokens": 100,
+                "total_tokens": 300,
+                "total_price": "0.003",
+                "currency": "USD",
+            }
+        }
+        result = CallToolResult(content=[TextContent(type="text", text="test")], _meta=meta)
+
+        with patch.object(tool, "invoke_remote_mcp_tool", return_value=result):
+            list(tool._invoke(user_id="test_user", tool_parameters={}))
+
+        # Verify latest_usage was set correctly
+        assert tool.latest_usage.prompt_tokens == 200
+        assert tool.latest_usage.completion_tokens == 100
+        assert tool.latest_usage.total_tokens == 300
+        assert tool.latest_usage.total_price == Decimal("0.003")
+
+    def test_invoke_with_no_meta_returns_empty_usage(self) -> None:
+        """Test that _invoke returns empty usage when no meta is present."""
+        tool = _make_mcp_tool()
+        result = CallToolResult(content=[TextContent(type="text", text="test")], _meta=None)
+
+        with patch.object(tool, "invoke_remote_mcp_tool", return_value=result):
+            list(tool._invoke(user_id="test_user", tool_parameters={}))
+
+        # Verify latest_usage is empty
+        assert tool.latest_usage.total_tokens == 0
+        assert tool.latest_usage.prompt_tokens == 0
+        assert tool.latest_usage.completion_tokens == 0
+
+    def test_latest_usage_property_returns_llm_usage(self) -> None:
+        """Test that latest_usage property returns LLMUsage instance."""
+        tool = _make_mcp_tool()
+        assert isinstance(tool.latest_usage, LLMUsage)
+
+    def test_initial_usage_is_empty(self) -> None:
+        """Test that MCPTool is initialized with empty usage."""
+        tool = _make_mcp_tool()
+        assert tool.latest_usage.total_tokens == 0
+        assert tool.latest_usage.prompt_tokens == 0
+        assert tool.latest_usage.completion_tokens == 0
+        assert tool.latest_usage.total_price == Decimal(0)
+
+    @pytest.mark.parametrize(
+        "meta_data",
+        [
+            # Direct usage field
+            {"usage": {"prompt_tokens": 10, "completion_tokens": 5, "total_tokens": 15}},
+            # Nested metadata
+            {"metadata": {"usage": {"total_tokens": 100}}},
+            # Flat token fields
+            {"total_tokens": 50, "prompt_tokens": 30, "completion_tokens": 20},
+            # With price info
+            {
+                "usage": {
+                    "total_tokens": 150,
+                    "total_price": "0.002",
+                    "currency": "EUR",
+                }
+            },
+            # Deep nested
+            {"level1": {"level2": {"usage": {"total_tokens": 200}}}},
+        ],
+    )
+    def test_various_meta_formats(self, meta_data) -> None:
+        """Test that various meta formats are correctly parsed."""
+        result = CallToolResult(content=[], _meta=meta_data)
+        usage = MCPTool._derive_usage_from_result(result)
+
+        assert isinstance(usage, LLMUsage)
+        # Should have at least some usage data
+        if meta_data.get("usage", {}).get("total_tokens") or meta_data.get("total_tokens"):
+            expected_total = (
+                meta_data.get("usage", {}).get("total_tokens")
+                or meta_data.get("total_tokens")
+                or meta_data.get("metadata", {}).get("usage", {}).get("total_tokens")
+                or meta_data.get("level1", {}).get("level2", {}).get("usage", {}).get("total_tokens")
+            )
+            if expected_total:
+                assert usage.total_tokens == expected_total