[WIP] chore: add mocking examples to the calculator sample agent

pyproject.toml

@@ -1,11 +1,11 @@
 [project]
 name = "uipath-langchain"
-version = "0.0.139"
+version = "0.0.140"
 description = "UiPath Langchain"
 readme = { file = "README.md", content-type = "text/markdown" }
 requires-python = ">=3.10"
 dependencies = [
-    "uipath>=2.1.76, <2.2.0",
+    "uipath>=2.1.86, <2.2.0",
     "langgraph>=0.5.0, <0.7.0",
     "langchain-core>=0.3.34",
     "langgraph-checkpoint-sqlite>=2.0.3",

samples/calculator-agent/evals/eval-sets/default.json

@@ -0,0 +1,72 @@
+{
+  "fileName": "default.json",
+  "id": "default-eval-set-id",
+  "name": "Basic Calculator Evaluation Set",
+  "batchSize": 10,
+  "evaluatorRefs": [
+    "equality",
+    "llm-as-a-judge"
+  ],
+  "evaluations": [
+    {
+      "id": "test-addition",
+      "name": "Test Addition",
+      "inputs": {"a": 1, "b": 1, "operator":  "+"},
+      "expectedOutput": {"result": 2},
+      "expectedAgentBehavior": "",
+      "evalSetId": "default-eval-set-id",
+      "createdAt": "2025-09-04T18:54:58.378Z",
+      "updatedAt": "2025-09-04T18:55:55.416Z"
+    },
+    {
+      "id": "test-random-addition-using-mockito",
+      "name": "Test Random Addition Using Mockito",
+      "inputs": {"a": 1, "b": 1, "operator":  "random"},
+      "expectedOutput": {"result": 2},
+      "expectedAgentBehavior": "",
+      "mockingStrategy": {
+        "type": "mockito",
+        "behaviors": [
+          {
+            "function": "get_random_operator",
+            "arguments": {
+              "args": [],
+              "kwargs": {}
+            },
+            "then": [
+              {
+                "type": "return",
+                "value": {"result": "+"}
+              }
+            ]
+          }
+        ]
+      },
+      "evalSetId": "default-eval-set-id",
+      "createdAt": "2025-09-04T18:54:58.378Z",
+      "updatedAt": "2025-09-04T18:55:55.416Z"
+    },
+    {
+      "id": "test-random-addition-using-llm",
+      "name": "Test Random Addition Using LLM",
+      "inputs": {"a": 1, "b": 1, "operator":  "random"},
+      "expectedOutput": {"result": 2},
+      "expectedAgentBehavior": "",
+      "mockingStrategy": {
+        "type": "llm",
+        "prompt": "The random operator is '+'.",
+        "toolsToSimulate": [{"name":  "get_random_operator"}],
+        "model": {
+          "model": "gpt-4o-mini-2024-07-18",
+          "temperature": 0
+        }
+      },
+      "evalSetId": "default-eval-set-id",
+      "createdAt": "2025-09-04T18:54:58.378Z",
+      "updatedAt": "2025-09-04T18:55:55.416Z"
+    }
+  ],
+  "modelSettings": [],
+  "createdAt": "2025-09-04T18:54:58.379Z",
+  "updatedAt": "2025-09-04T18:55:55.416Z"
+}

samples/calculator-agent/evals/evaluators/equality.json

@@ -0,0 +1,11 @@
+{
+    "fileName": "equality.json",
+    "id": "equality",
+    "name": "Equality Evaluator",
+    "description": "An evaluator that judges the agent based on expected output.",
+    "category": 0,
+    "type": 1,
+    "targetOutputKey": "*",
+    "createdAt": "2025-06-26T17:45:39.651Z",
+    "updatedAt": "2025-06-26T17:45:39.651Z"
+}

samples/calculator-agent/evals/evaluators/llm-as-a-judge.json

@@ -0,0 +1,13 @@
+{
+    "fileName": "llm-as-a-judge.json",
+    "id": "llm-as-a-judge",
+    "name": "LLMAsAJudge Evaluator",
+    "description": "An evaluator that judges the agent based on it's run history and expected behavior",
+    "category": 3,
+    "type": 7,
+    "prompt": "As an expert evaluator, determine how well the agent did on a scale of 0-100. Focus on if the simulation was successful and if the agent behaved according to the expected output accounting for alternative valid expressions, and reasonable variations in language while maintaining high standards for accuracy and completeness. Provide your score with a justification, explaining briefly and concisely why you gave that score.\n----\nUserOrSyntheticInputGivenToAgent:\n{{UserOrSyntheticInput}}\n----\nSimulationInstructions:\n{{SimulationInstructions}}\n----\nExpectedAgentBehavior:\n{{ExpectedAgentBehavior}}\n----\nAgentRunHistory:\n{{AgentRunHistory}}\n",
+    "targetOutputKey": "*",
+    "model": "gpt-4.1-2025-04-14",
+    "createdAt": "2025-08-21T15:12:58.695Z",
+    "updatedAt": "2025-08-21T15:12:58.695Z"
+}

samples/calculator-agent/graph.py

@@ -1,45 +1,78 @@
+import random
 from enum import Enum
-
 from langgraph.constants import START, END
 from langgraph.graph import StateGraph
 from pydantic.dataclasses import dataclass
 from uipath.tracing import traced
+from uipath.eval.mocks import mockable, ExampleCall
 
 
 class Operator(Enum):
     ADD = "+"
     SUBTRACT = "-"
     MULTIPLY = "*"
     DIVIDE = "/"
+    RANDOM = "random"
+
 
 @dataclass
 class CalculatorInput:
     a: float
     b: float
     operator: Operator
 
+
 @dataclass
 class CalculatorOutput:
     result: float
 
-@traced(name="postprocess")
+@dataclass
+class Wrapper:
+    result: Operator
+
+GET_RANDOM_OPERATOR_EXAMPLES = [ExampleCall(id="example", input="{}", output="{\"result\": \"*\"}")]
+
+@traced()
+@mockable(example_calls=GET_RANDOM_OPERATOR_EXAMPLES)
+async def get_random_operator() -> Wrapper:
+    """Return a random math operator."""
+    return Wrapper(result=random.choice([
+        Operator.ADD,
+        Operator.SUBTRACT,
+        Operator.MULTIPLY,
+        Operator.DIVIDE,
+    ]))
+
+
+@traced()
 async def postprocess(x: float) -> float:
-    """
-    Example of nested traced invocation.
-    """
+    """Example of nested traced invocation."""
     return x
 
-@traced(name="calculate")
+
+@traced()
 async def calculate(input: CalculatorInput) -> CalculatorOutput:
-    result = 0
-    match input.operator:
-        case Operator.ADD: result = input.a + input.b
-        case Operator.SUBTRACT: result = input.a - input.b
-        case Operator.MULTIPLY: result = input.a * input.b
-        case Operator.DIVIDE: result = input.a / input.b if input.b != 0 else 0
+    if input.operator == Operator.RANDOM:
+        operator = (await get_random_operator()).result
+    else:
+        operator = input.operator
+
+    match operator:
+        case Operator.ADD:
+            result = input.a + input.b
+        case Operator.SUBTRACT:
+            result = input.a - input.b
+        case Operator.MULTIPLY:
+            result = input.a * input.b
+        case Operator.DIVIDE:
+            result = input.a / input.b if input.b != 0 else 0
+        case _:
+            raise ValueError(f"Unknown operator: {operator}")
+
     result = await postprocess(result)
     return CalculatorOutput(result=result)
 
+
 builder = StateGraph(state_schema=CalculatorInput, input=CalculatorInput, output=CalculatorOutput)
 
 builder.add_node("calculate", calculate)

samples/calculator-agent/pyproject.toml

@@ -5,8 +5,8 @@ description = "Calculator Agent"
 authors = [{ name = "John Doe", email = "[email protected]" }]
 requires-python = ">=3.10"
 dependencies = [
-    "uipath-langchain>=0.0.130",
-    "uipath>=2.1.54",
+    "uipath-langchain>=0.0.138",
+    "uipath>=2.1.86",
 ]
 
 [project.optional-dependencies]