feat: add evaluations

Author: radugheo

src/uipath/dev/__init__.py

@@ -1,12 +1,22 @@
 """UiPath Dev Console models module."""
 
+from uipath.dev.models.eval_run import EvalRun, EvaluationResult, EvaluatorResult
+from uipath.dev.models.evaluator_types import (
+    EVALUATOR_TYPES,
+    get_evaluator_type,
+)
 from uipath.dev.models.execution import ExecutionMode, ExecutionRun
 from uipath.dev.models.messages import ChatMessage, LogMessage, TraceMessage
 
 __all__ = [
-    "ExecutionRun",
-    "ExecutionMode",
+    "EVALUATOR_TYPES",
     "ChatMessage",
+    "EvalRun",
+    "EvaluationResult",
+    "EvaluatorResult",
+    "ExecutionMode",
+    "ExecutionRun",
     "LogMessage",
     "TraceMessage",
+    "get_evaluator_type",
 ]

src/uipath/dev/models/__init__.py

@@ -0,0 +1,247 @@
+"""Model for evaluation runs."""
+
+import os
+from dataclasses import dataclass, field
+from datetime import datetime
+from typing import TYPE_CHECKING, Any
+from uuid import uuid4
+
+from rich.text import Text
+from uipath.runtime.errors import UiPathErrorContract
+
+if TYPE_CHECKING:
+    from uipath.dev.models.messages import LogMessage, TraceMessage
+
+
+@dataclass
+class EvaluatorResult:
+    """Result from a single evaluator for a single evaluation."""
+
+    evaluator_id: str
+    evaluator_name: str
+    score: float
+    details: str = ""
+    evaluation_time: float = 0.0
+    justification: str = ""
+
+
+@dataclass
+class EvaluationResult:
+    """Result for a single evaluation."""
+
+    eval_id: str
+    eval_name: str
+    evaluator_results: list[EvaluatorResult] = field(default_factory=list)
+
+    @property
+    def passed(self) -> bool:
+        """Check if all evaluators passed (score == 1.0)."""
+        return all(r.score == 1.0 for r in self.evaluator_results)
+
+
+class EvalRun:
+    """A single evaluation run."""
+
+    def __init__(
+        self,
+        eval_set_path: str,
+        entrypoint: str,
+        *,
+        id: str | None = None,
+        name: str = "",
+        no_report: bool = False,
+        workers: int = 1,
+        eval_set_run_id: str | None = None,
+        enable_mocker_cache: bool = False,
+        eval_ids: list[str] | None = None,
+        report_coverage: bool = False,
+        output_file: str | None = None,
+        # For deserialization
+        status: str = "pending",
+        start_time: datetime | None = None,
+        end_time: datetime | None = None,
+        evaluator_refs: list[str] | None = None,
+        error: UiPathErrorContract | None = None,
+    ):
+        """Initialize an EvalRun instance."""
+        self.id = id if id is not None else str(uuid4())[:8]
+        self.eval_set_path = eval_set_path
+        self.entrypoint = entrypoint
+        self.name = name if name else f"Run: {self.id}"
+        self.status = status  # pending, running, completed, failed
+        self.start_time = start_time if start_time is not None else datetime.now()
+        self.end_time = end_time
+        self.evaluator_refs: list[str] = evaluator_refs if evaluator_refs is not None else []
+        self.evaluation_results: list[EvaluationResult] = []
+        self.error = error
+        self.logs: list["LogMessage"] = []
+        self.traces: list["TraceMessage"] = []
+        # Execution options
+        self.no_report = no_report
+        self.workers = workers
+        self.eval_set_run_id = eval_set_run_id
+        self.enable_mocker_cache = enable_mocker_cache
+        self.eval_ids: list[str] = eval_ids if eval_ids is not None else []
+        self.report_coverage = report_coverage
+        self.output_file = output_file
+
+    @property
+    def duration(self) -> str:
+        """Get the duration of the run as a formatted string."""
+        if self.end_time:
+            delta = self.end_time - self.start_time
+            return f"{delta.total_seconds():.1f}s"
+        elif self.start_time:
+            delta = datetime.now() - self.start_time
+            return f"{delta.total_seconds():.1f}s"
+        return "0.0s"
+
+    @property
+    def display_name(self) -> Text:
+        """Get formatted display name with status indicator."""
+        status_colors = {
+            "pending": "grey50",
+            "running": "yellow",
+            "completed": "green",
+            "failed": "red",
+        }
+
+        status_icon = {
+            "pending": "●",
+            "running": "▶",
+            "completed": "✔",
+            "failed": "✖",
+        }.get(self.status, "?")
+
+        eval_set_name = (
+            os.path.basename(self.eval_set_path).rsplit(".", 1)[0]
+            if self.eval_set_path
+            else "eval"
+        )
+        truncated_name = eval_set_name[:8]
+        time_str = self.start_time.strftime("%H:%M:%S")
+        duration_str = self.duration[:6]
+
+        text = Text()
+        text.append(f"{status_icon:<2} ", style=status_colors.get(self.status, "white"))
+        text.append(f"{truncated_name:<8} ")
+        text.append(f"({time_str:<8}) ")
+        text.append(f"[{duration_str:<6}]")
+
+        return text
+
+    @property
+    def total_evaluations(self) -> int:
+        """Get total number of evaluations."""
+        return len(self.evaluation_results)
+
+    @property
+    def evaluator_scores(self) -> dict[str, float]:
+        """Get average score per evaluator across all evaluations."""
+        scores: dict[str, list[float]] = {}
+        for eval_result in self.evaluation_results:
+            for ev_result in eval_result.evaluator_results:
+                if ev_result.evaluator_id not in scores:
+                    scores[ev_result.evaluator_id] = []
+                scores[ev_result.evaluator_id].append(ev_result.score)
+
+        return {
+            ev_id: sum(s) / len(s) if s else 0.0
+            for ev_id, s in scores.items()
+        }
+
+    @property
+    def overall_score(self) -> float:
+        """Get overall average score."""
+        all_scores = []
+        for eval_result in self.evaluation_results:
+            for ev_result in eval_result.evaluator_results:
+                all_scores.append(ev_result.score)
+        return sum(all_scores) / len(all_scores) if all_scores else 0.0
+
+    def to_dict(self) -> dict[str, Any]:
+        """Convert to dictionary for serialization."""
+        return {
+            "id": self.id,
+            "name": self.name,
+            "eval_set_path": self.eval_set_path,
+            "entrypoint": self.entrypoint,
+            "status": self.status,
+            "start_time": self.start_time.isoformat(),
+            "end_time": self.end_time.isoformat() if self.end_time else None,
+            "evaluator_refs": self.evaluator_refs,
+            "evaluation_results": [
+                {
+                    "eval_id": er.eval_id,
+                    "eval_name": er.eval_name,
+                    "evaluator_results": [
+                        {
+                            "evaluator_id": evr.evaluator_id,
+                            "evaluator_name": evr.evaluator_name,
+                            "score": evr.score,
+                            "details": evr.details,
+                            "evaluation_time": evr.evaluation_time,
+                            "justification": evr.justification,
+                        }
+                        for evr in er.evaluator_results
+                    ],
+                }
+                for er in self.evaluation_results
+            ],
+            "error": self.error.to_dict() if self.error else None,
+            # Execution options
+            "no_report": self.no_report,
+            "workers": self.workers,
+            "eval_set_run_id": self.eval_set_run_id,
+            "enable_mocker_cache": self.enable_mocker_cache,
+            "eval_ids": self.eval_ids,
+            "report_coverage": self.report_coverage,
+            "output_file": self.output_file,
+        }
+
+    @classmethod
+    def from_dict(cls, data: dict[str, Any]) -> "EvalRun":
+        """Create from dictionary."""
+        error_data = data.get("error")
+        error = UiPathErrorContract.from_dict(error_data) if error_data else None
+
+        eval_run = cls(
+            id=data["id"],
+            name=data.get("name", ""),
+            eval_set_path=data["eval_set_path"],
+            entrypoint=data["entrypoint"],
+            status=data.get("status", "pending"),
+            start_time=datetime.fromisoformat(data["start_time"]) if data.get("start_time") else None,
+            end_time=datetime.fromisoformat(data["end_time"]) if data.get("end_time") else None,
+            evaluator_refs=data.get("evaluator_refs", []),
+            error=error,
+            # Execution options
+            no_report=data.get("no_report", False),
+            workers=data.get("workers", 1),
+            eval_set_run_id=data.get("eval_set_run_id"),
+            enable_mocker_cache=data.get("enable_mocker_cache", False),
+            eval_ids=data.get("eval_ids", []),
+            report_coverage=data.get("report_coverage", False),
+            output_file=data.get("output_file"),
+        )
+
+        # Parse evaluation results
+        for er_data in data.get("evaluation_results", []):
+            eval_result = EvaluationResult(
+                eval_id=er_data["eval_id"],
+                eval_name=er_data.get("eval_name", er_data["eval_id"]),
+            )
+            for evr_data in er_data.get("evaluator_results", []):
+                eval_result.evaluator_results.append(
+                    EvaluatorResult(
+                        evaluator_id=evr_data["evaluator_id"],
+                        evaluator_name=evr_data.get("evaluator_name", evr_data["evaluator_id"]),
+                        score=evr_data.get("score", 0.0),
+                        details=evr_data.get("details", ""),
+                        evaluation_time=evr_data.get("evaluation_time", 0.0),
+                        justification=evr_data.get("justification", ""),
+                    )
+                )
+            eval_run.evaluation_results.append(eval_result)
+
+        return eval_run