refactor: move evals/ into src/openjarvis/evals/ as proper subpackage

Move the standalone evals framework from the project root into the
openjarvis package. Rewrite all ~50+ import statements from 'from evals.'
to 'from openjarvis.evals.' across the package, CLI, and tests. Remove
the evals-specific pyproject.toml (no longer a standalone package).
Update ruff per-file-ignores paths and fix line-length violations
introduced by the longer import paths.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

evals/__main__.py

@@ -1,6 +0,0 @@
-"""Allow running as ``python -m evals``."""
-
-from evals.cli import main
-
-if __name__ == "__main__":
-    main()

evals/pyproject.toml

@@ -1,24 +0,0 @@
-[project]
-name = "openjarvis-evals"
-version = "0.1.0"
-description = "Evaluation framework for OpenJarvis"
-requires-python = ">=3.10"
-dependencies = [
-    "openjarvis>=1.0.0",
-    "click>=8",
-    "datasets>=2.14",
-    "huggingface-hub>=0.20",
-    "tqdm>=4.65",
-    "rich>=13",
-    "tomli>=2.0; python_version < '3.11'",
-]
-
-[project.optional-dependencies]
-dev = ["pytest>=8", "pytest-cov>=5"]
-
-[project.scripts]
-openjarvis-eval = "evals.cli:main"
-
-[build-system]
-requires = ["hatchling"]
-build-backend = "hatchling.build"

pyproject.toml

@@ -132,5 +132,5 @@ src = ["src", "tests"]
 select = ["E", "F", "I", "W"]
 
 [tool.ruff.lint.per-file-ignores]
-"evals/datasets/*.py" = ["E501"]
-"evals/scorers/*.py" = ["E501"]
+"src/openjarvis/evals/datasets/*.py" = ["E501"]
+"src/openjarvis/evals/scorers/*.py" = ["E501"]

src/openjarvis/cli/eval_cmd.py

@@ -135,7 +135,7 @@ def eval_run(
     # Config-driven mode: load TOML suite, expand, run all
     if config_path is not None:
         try:
-            from evals.core.config import expand_suite, load_eval_config
+            from openjarvis.evals.core.config import expand_suite, load_eval_config
         except ImportError:
             console.print(
                 "[red]Eval framework not available. "
@@ -159,7 +159,7 @@ def eval_run(
         )
 
         try:
-            from evals.cli import _run_single
+            from openjarvis.evals.cli import _run_single
         except ImportError:
             console.print(
                 "[red]Eval CLI module not available.[/red]"
@@ -195,7 +195,7 @@ def eval_run(
         )
 
     try:
-        from evals.core.types import RunConfig
+        from openjarvis.evals.core.types import RunConfig
     except ImportError:
         console.print(
             "[red]Eval framework not available. "
@@ -219,7 +219,7 @@ def eval_run(
     )
 
     try:
-        from evals.cli import _run_single
+        from openjarvis.evals.cli import _run_single
 
         console.print(
             f"[cyan]Benchmark:[/cyan] {benchmark}\n"

src/openjarvis/evals/__main__.py

@@ -0,0 +1,6 @@
+"""Allow running as ``python -m openjarvis.evals``."""
+
+from openjarvis.evals.cli import main
+
+if __name__ == "__main__":
+    main()

src/openjarvis/evals/backends/jarvis_agent.py

@@ -5,7 +5,7 @@ from __future__ import annotations
 import time
 from typing import Any, Dict, List, Optional
 
-from evals.core.backend import InferenceBackend
+from openjarvis.evals.core.backend import InferenceBackend
 
 
 class JarvisAgentBackend(InferenceBackend):

src/openjarvis/evals/backends/jarvis_direct.py

@@ -5,7 +5,7 @@ from __future__ import annotations
 import time
 from typing import Any, Dict, Optional
 
-from evals.core.backend import InferenceBackend
+from openjarvis.evals.core.backend import InferenceBackend
 
 
 class JarvisDirectBackend(InferenceBackend):

src/openjarvis/evals/cli.py

@@ -17,7 +17,7 @@ from rich.progress import (
     TimeRemainingColumn,
 )
 
-from evals.core.display import (
+from openjarvis.evals.core.display import (
     print_banner,
     print_completion,
     print_full_results,
@@ -72,7 +72,7 @@ def _build_backend(backend_name: str, engine_key: Optional[str],
                     telemetry: bool = False, gpu_metrics: bool = False):
     """Construct the appropriate backend."""
     if backend_name == "jarvis-agent":
-        from evals.backends.jarvis_agent import JarvisAgentBackend
+        from openjarvis.evals.backends.jarvis_agent import JarvisAgentBackend
         return JarvisAgentBackend(
             engine_key=engine_key,
             agent_name=agent_name,
@@ -81,7 +81,7 @@ def _build_backend(backend_name: str, engine_key: Optional[str],
             gpu_metrics=gpu_metrics,
         )
     else:
-        from evals.backends.jarvis_direct import JarvisDirectBackend
+        from openjarvis.evals.backends.jarvis_direct import JarvisDirectBackend
         return JarvisDirectBackend(
             engine_key=engine_key,
             telemetry=telemetry,
@@ -92,49 +92,51 @@ def _build_backend(backend_name: str, engine_key: Optional[str],
 def _build_dataset(benchmark: str):
     """Construct the dataset provider for a benchmark."""
     if benchmark == "supergpqa":
-        from evals.datasets.supergpqa import SuperGPQADataset
+        from openjarvis.evals.datasets.supergpqa import SuperGPQADataset
         return SuperGPQADataset()
     elif benchmark == "gpqa":
-        from evals.datasets.gpqa import GPQADataset
+        from openjarvis.evals.datasets.gpqa import GPQADataset
         return GPQADataset()
     elif benchmark == "mmlu-pro":
-        from evals.datasets.mmlu_pro import MMLUProDataset
+        from openjarvis.evals.datasets.mmlu_pro import MMLUProDataset
         return MMLUProDataset()
     elif benchmark == "math500":
-        from evals.datasets.math500 import MATH500Dataset
+        from openjarvis.evals.datasets.math500 import MATH500Dataset
         return MATH500Dataset()
     elif benchmark == "natural-reasoning":
-        from evals.datasets.natural_reasoning import NaturalReasoningDataset
+        from openjarvis.evals.datasets.natural_reasoning import NaturalReasoningDataset
         return NaturalReasoningDataset()
     elif benchmark == "hle":
-        from evals.datasets.hle import HLEDataset
+        from openjarvis.evals.datasets.hle import HLEDataset
         return HLEDataset()
     elif benchmark == "simpleqa":
-        from evals.datasets.simpleqa import SimpleQADataset
+        from openjarvis.evals.datasets.simpleqa import SimpleQADataset
         return SimpleQADataset()
     elif benchmark == "wildchat":
-        from evals.datasets.wildchat import WildChatDataset
+        from openjarvis.evals.datasets.wildchat import WildChatDataset
         return WildChatDataset()
     elif benchmark == "ipw":
-        from evals.datasets.ipw_mixed import IPWDataset
+        from openjarvis.evals.datasets.ipw_mixed import IPWDataset
         return IPWDataset()
     elif benchmark == "gaia":
-        from evals.datasets.gaia import GAIADataset
+        from openjarvis.evals.datasets.gaia import GAIADataset
         return GAIADataset()
     elif benchmark == "frames":
-        from evals.datasets.frames import FRAMESDataset
+        from openjarvis.evals.datasets.frames import FRAMESDataset
         return FRAMESDataset()
     elif benchmark == "swebench":
-        from evals.datasets.swebench import SWEBenchDataset
+        from openjarvis.evals.datasets.swebench import SWEBenchDataset
         return SWEBenchDataset()
     elif benchmark == "swefficiency":
-        from evals.datasets.swefficiency import SWEfficiencyDataset
+        from openjarvis.evals.datasets.swefficiency import SWEfficiencyDataset
         return SWEfficiencyDataset()
     elif benchmark == "terminalbench":
-        from evals.datasets.terminalbench import TerminalBenchDataset
+        from openjarvis.evals.datasets.terminalbench import TerminalBenchDataset
         return TerminalBenchDataset()
     elif benchmark == "terminalbench-native":
-        from evals.datasets.terminalbench_native import TerminalBenchNativeDataset
+        from openjarvis.evals.datasets.terminalbench_native import (
+            TerminalBenchNativeDataset,
+        )
         return TerminalBenchNativeDataset()
     else:
         raise click.ClickException(f"Unknown benchmark: {benchmark}")
@@ -143,46 +145,46 @@ def _build_dataset(benchmark: str):
 def _build_scorer(benchmark: str, judge_backend, judge_model: str):
     """Construct the scorer for a benchmark."""
     if benchmark == "supergpqa":
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
         return SuperGPQAScorer(judge_backend, judge_model)
     elif benchmark == "gpqa":
-        from evals.scorers.gpqa_mcq import GPQAScorer
+        from openjarvis.evals.scorers.gpqa_mcq import GPQAScorer
         return GPQAScorer(judge_backend, judge_model)
     elif benchmark == "mmlu-pro":
-        from evals.scorers.mmlu_pro_mcq import MMLUProScorer
+        from openjarvis.evals.scorers.mmlu_pro_mcq import MMLUProScorer
         return MMLUProScorer(judge_backend, judge_model)
     elif benchmark == "math500" or benchmark == "natural-reasoning":
-        from evals.scorers.reasoning_judge import ReasoningJudgeScorer
+        from openjarvis.evals.scorers.reasoning_judge import ReasoningJudgeScorer
         return ReasoningJudgeScorer(judge_backend, judge_model)
     elif benchmark == "hle":
-        from evals.scorers.hle_judge import HLEScorer
+        from openjarvis.evals.scorers.hle_judge import HLEScorer
         return HLEScorer(judge_backend, judge_model)
     elif benchmark == "simpleqa":
-        from evals.scorers.simpleqa_judge import SimpleQAScorer
+        from openjarvis.evals.scorers.simpleqa_judge import SimpleQAScorer
         return SimpleQAScorer(judge_backend, judge_model)
     elif benchmark == "wildchat":
-        from evals.scorers.wildchat_judge import WildChatScorer
+        from openjarvis.evals.scorers.wildchat_judge import WildChatScorer
         return WildChatScorer(judge_backend, judge_model)
     elif benchmark == "ipw":
-        from evals.scorers.ipw_mixed import IPWMixedScorer
+        from openjarvis.evals.scorers.ipw_mixed import IPWMixedScorer
         return IPWMixedScorer(judge_backend, judge_model)
     elif benchmark == "gaia":
-        from evals.scorers.gaia_exact import GAIAScorer
+        from openjarvis.evals.scorers.gaia_exact import GAIAScorer
         return GAIAScorer(judge_backend, judge_model)
     elif benchmark == "frames":
-        from evals.scorers.frames_judge import FRAMESScorer
+        from openjarvis.evals.scorers.frames_judge import FRAMESScorer
         return FRAMESScorer(judge_backend, judge_model)
     elif benchmark == "swebench":
-        from evals.scorers.swebench_structural import SWEBenchScorer
+        from openjarvis.evals.scorers.swebench_structural import SWEBenchScorer
         return SWEBenchScorer(judge_backend, judge_model)
     elif benchmark == "swefficiency":
-        from evals.scorers.swefficiency_structural import SWEfficiencyScorer
+        from openjarvis.evals.scorers.swefficiency_structural import SWEfficiencyScorer
         return SWEfficiencyScorer(judge_backend, judge_model)
     elif benchmark == "terminalbench":
-        from evals.scorers.terminalbench_judge import TerminalBenchScorer
+        from openjarvis.evals.scorers.terminalbench_judge import TerminalBenchScorer
         return TerminalBenchScorer(judge_backend, judge_model)
     elif benchmark == "terminalbench-native":
-        from evals.scorers.terminalbench_native_structural import (
+        from openjarvis.evals.scorers.terminalbench_native_structural import (
             TerminalBenchNativeScorer,
         )
         return TerminalBenchNativeScorer(judge_backend, judge_model)
@@ -192,7 +194,7 @@ def _build_scorer(benchmark: str, judge_backend, judge_model: str):
 
 def _build_judge_backend(judge_model: str):
     """Build the judge backend (always cloud for LLM-as-judge)."""
-    from evals.backends.jarvis_direct import JarvisDirectBackend
+    from openjarvis.evals.backends.jarvis_direct import JarvisDirectBackend
     return JarvisDirectBackend(engine_key="cloud")
 
 
@@ -217,7 +219,7 @@ def _print_summary(
 
 def _run_single(config, console: Optional[Console] = None) -> object:
     """Run a single eval from a RunConfig and return the summary."""
-    from evals.core.runner import EvalRunner
+    from openjarvis.evals.core.runner import EvalRunner
 
     if console is None:
         console = Console()
@@ -264,7 +266,7 @@ def _run_single(config, console: Optional[Console] = None) -> object:
 
 def _run_from_config(config_path: str, verbose: bool) -> None:
     """Load a TOML config and run the full models x benchmarks matrix."""
-    from evals.core.config import expand_suite, load_eval_config
+    from openjarvis.evals.core.config import expand_suite, load_eval_config
 
     console = Console()
 
@@ -381,7 +383,7 @@ def run(ctx, config_path, benchmark, backend, model, engine_key, agent_name,
             "(required when --config is not provided)"
         )
 
-    from evals.core.types import RunConfig
+    from openjarvis.evals.core.types import RunConfig
 
     tool_list = [t.strip() for t in tools.split(",") if t.strip()] if tools else []
 
@@ -451,8 +453,8 @@ def run_all(model, engine_key, max_samples, max_workers, judge_model,
     """Run all benchmarks."""
     _setup_logging(verbose)
 
-    from evals.core.runner import EvalRunner
-    from evals.core.types import RunConfig
+    from openjarvis.evals.core.runner import EvalRunner
+    from openjarvis.evals.core.types import RunConfig
 
     console = Console()
 

src/openjarvis/evals/core/config.py

@@ -7,7 +7,7 @@ import sys
 from pathlib import Path
 from typing import List
 
-from evals.core.types import (
+from openjarvis.evals.core.types import (
     BenchmarkConfig,
     DefaultsConfig,
     EvalSuiteConfig,

src/openjarvis/evals/core/dataset.py

@@ -5,7 +5,7 @@ from __future__ import annotations
 from abc import ABC, abstractmethod
 from typing import Iterable, Optional
 
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.types import EvalRecord
 
 
 class DatasetProvider(ABC):

src/openjarvis/evals/core/display.py

@@ -12,7 +12,7 @@ from rich.table import Table
 if TYPE_CHECKING:
     from pathlib import Path
 
-    from evals.core.types import MetricStats, RunSummary
+    from openjarvis.evals.core.types import MetricStats, RunSummary
 
 OPENJARVIS_BANNER = r"""
   ___                       _                  _

src/openjarvis/evals/core/runner.py

@@ -11,10 +11,10 @@ from concurrent.futures import ThreadPoolExecutor, as_completed
 from pathlib import Path
 from typing import Any, Callable, Dict, List, Optional
 
-from evals.core.backend import InferenceBackend
-from evals.core.dataset import DatasetProvider
-from evals.core.scorer import Scorer
-from evals.core.types import EvalRecord, EvalResult, MetricStats, RunConfig, RunSummary
+from openjarvis.evals.core.backend import InferenceBackend
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.scorer import Scorer
+from openjarvis.evals.core.types import EvalRecord, EvalResult, MetricStats, RunConfig, RunSummary
 
 try:
     from openjarvis.telemetry.efficiency import compute_efficiency

src/openjarvis/evals/core/scorer.py

@@ -5,8 +5,8 @@ from __future__ import annotations
 from abc import ABC, abstractmethod
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.backend import InferenceBackend
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.backend import InferenceBackend
+from openjarvis.evals.core.types import EvalRecord
 
 
 class Scorer(ABC):

src/openjarvis/evals/datasets/frames.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 _DEFAULT_INPUT_PROMPT = """Please answer the question below. You should:
 

src/openjarvis/evals/datasets/gaia.py

@@ -11,8 +11,8 @@ import shutil
 from pathlib import Path
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 _DEFAULT_CACHE_DIR = Path.home() / ".cache" / "gaia_benchmark"
 

src/openjarvis/evals/datasets/gpqa.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 
 def _format_options(options: Iterable[str]) -> str:

src/openjarvis/evals/datasets/hle.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 # Fields whose presence signals a multimodal row.
 _MULTIMODAL_FIELDS = frozenset(

src/openjarvis/evals/datasets/ipw_mixed.py

@@ -13,8 +13,8 @@ import random
 from pathlib import Path
 from typing import Any, Dict, Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/datasets/math500.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 _PROMPT_TEMPLATE = (
     "Solve the following math problem step by step. "

src/openjarvis/evals/datasets/mmlu_pro.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 
 def _format_options(options: Iterable[str]) -> str:

src/openjarvis/evals/datasets/natural_reasoning.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 _PROMPT_TEMPLATE = (
     "Please solve the following reasoning problem. "

src/openjarvis/evals/datasets/simpleqa.py

@@ -9,8 +9,8 @@ import ast
 import random
 from typing import Any, Dict, Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 _PROMPT_TEMPLATE = """Please answer the following question with a short, factual response.
 Your answer should be a word, phrase, name, number, or date.

src/openjarvis/evals/datasets/supergpqa.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 
 def _format_options(options: Iterable[str]) -> str:

src/openjarvis/evals/datasets/swebench.py

@@ -9,8 +9,8 @@ import json
 import random
 from typing import Any, Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 _HF_PATHS = {
     "verified": "princeton-nlp/SWE-bench_Verified",

src/openjarvis/evals/datasets/swefficiency.py

@@ -9,8 +9,8 @@ import json
 import random
 from typing import Any, Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 _HF_PATH = "swefficiency/swefficiency"
 

src/openjarvis/evals/datasets/terminalbench.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 try:
     from datasets import load_dataset as _load_dataset  # noqa: F401

src/openjarvis/evals/datasets/terminalbench_native.py

@@ -10,8 +10,8 @@ import random
 from pathlib import Path
 from typing import Any, Dict, Iterable, List, Optional
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 try:
     from terminal_bench import Task, TaskPaths

src/openjarvis/evals/datasets/wildchat.py

@@ -8,8 +8,8 @@ from __future__ import annotations
 import random
 from typing import Iterable, List, MutableMapping, Optional, Sequence
 
-from evals.core.dataset import DatasetProvider
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.types import EvalRecord
 
 
 class WildChatDataset(DatasetProvider):

src/openjarvis/evals/scorers/frames_judge.py

@@ -9,8 +9,8 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/gaia_exact.py

@@ -10,8 +10,8 @@ import re
 import string
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/gpqa_mcq.py

@@ -9,8 +9,8 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/hle_judge.py

@@ -11,9 +11,9 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
-from evals.scorers.reasoning_judge import reasoning_exact_match
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
+from openjarvis.evals.scorers.reasoning_judge import reasoning_exact_match
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/ipw_mixed.py

@@ -11,8 +11,8 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/mmlu_pro_mcq.py

@@ -9,8 +9,8 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/reasoning_judge.py

@@ -11,8 +11,8 @@ import re
 import string
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/simpleqa_judge.py

@@ -11,8 +11,8 @@ import re
 import string
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/supergpqa_mcq.py

@@ -9,8 +9,8 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/swebench_structural.py

@@ -11,8 +11,8 @@ from __future__ import annotations
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import Scorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import Scorer
+from openjarvis.evals.core.types import EvalRecord
 
 _DIFF_MARKERS = [
     r"^---\s",

src/openjarvis/evals/scorers/swefficiency_structural.py

@@ -12,8 +12,8 @@ from __future__ import annotations
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import Scorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import Scorer
+from openjarvis.evals.core.types import EvalRecord
 
 _DIFF_MARKERS = [
     r"^---\s",

src/openjarvis/evals/scorers/terminalbench_judge.py

@@ -10,8 +10,8 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/scorers/terminalbench_native_structural.py

@@ -9,8 +9,8 @@ from __future__ import annotations
 
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import Scorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import Scorer
+from openjarvis.evals.core.types import EvalRecord
 
 
 class TerminalBenchNativeScorer(Scorer):

src/openjarvis/evals/scorers/wildchat_judge.py

@@ -9,8 +9,8 @@ import logging
 import re
 from typing import Any, Dict, Optional, Tuple
 
-from evals.core.scorer import LLMJudgeScorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.scorer import LLMJudgeScorer
+from openjarvis.evals.core.types import EvalRecord
 
 LOGGER = logging.getLogger(__name__)
 

src/openjarvis/evals/tests/conftest.py

@@ -11,10 +11,10 @@ import pytest
 # Ensure evals package is importable from the repo root
 sys.path.insert(0, str(Path(__file__).resolve().parent.parent.parent))
 
-from evals.core.backend import InferenceBackend
-from evals.core.dataset import DatasetProvider
-from evals.core.scorer import Scorer
-from evals.core.types import EvalRecord
+from openjarvis.evals.core.backend import InferenceBackend
+from openjarvis.evals.core.dataset import DatasetProvider
+from openjarvis.evals.core.scorer import Scorer
+from openjarvis.evals.core.types import EvalRecord
 
 # ---------------------------------------------------------------------------
 # Mock backend

src/openjarvis/evals/tests/test_backends.py

@@ -16,7 +16,7 @@ class TestJarvisDirectBackend:
         mock_builder.build.return_value = mock_system
         mock_builder_cls.return_value = mock_builder
 
-        from evals.backends.jarvis_direct import JarvisDirectBackend
+        from openjarvis.evals.backends.jarvis_direct import JarvisDirectBackend
 
         backend = JarvisDirectBackend()
         assert backend.backend_id == "jarvis-direct"
@@ -33,7 +33,7 @@ class TestJarvisDirectBackend:
         mock_builder.build.return_value = MagicMock()
         mock_builder_cls.return_value = mock_builder
 
-        from evals.backends.jarvis_direct import JarvisDirectBackend
+        from openjarvis.evals.backends.jarvis_direct import JarvisDirectBackend
 
         JarvisDirectBackend(engine_key="cloud")
         mock_builder.engine.assert_called_with("cloud")
@@ -54,7 +54,7 @@ class TestJarvisDirectBackend:
         mock_builder.build.return_value = mock_system
         mock_builder_cls.return_value = mock_builder
 
-        from evals.backends.jarvis_direct import JarvisDirectBackend
+        from openjarvis.evals.backends.jarvis_direct import JarvisDirectBackend
 
         backend = JarvisDirectBackend()
         result = backend.generate_full("What is 2+2?", model="test-model")
@@ -77,7 +77,7 @@ class TestJarvisDirectBackend:
         mock_builder.build.return_value = mock_system
         mock_builder_cls.return_value = mock_builder
 
-        from evals.backends.jarvis_direct import JarvisDirectBackend
+        from openjarvis.evals.backends.jarvis_direct import JarvisDirectBackend
 
         backend = JarvisDirectBackend()
         text = backend.generate("Capital of France?", model="m")
@@ -96,7 +96,7 @@ class TestJarvisAgentBackend:
         mock_builder.build.return_value = MagicMock()
         mock_builder_cls.return_value = mock_builder
 
-        from evals.backends.jarvis_agent import JarvisAgentBackend
+        from openjarvis.evals.backends.jarvis_agent import JarvisAgentBackend
 
         backend = JarvisAgentBackend(
             engine_key="cloud", agent_name="orchestrator",
@@ -128,7 +128,7 @@ class TestJarvisAgentBackend:
         mock_builder.build.return_value = mock_system
         mock_builder_cls.return_value = mock_builder
 
-        from evals.backends.jarvis_agent import JarvisAgentBackend
+        from openjarvis.evals.backends.jarvis_agent import JarvisAgentBackend
 
         backend = JarvisAgentBackend(agent_name="orchestrator")
         result = backend.generate_full("What is 2+2?", model="gpt-4o")

src/openjarvis/evals/tests/test_cli_flags.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 from click.testing import CliRunner
 
-from evals.cli import main
+from openjarvis.evals.cli import main
 
 
 class TestCompactFlag:

src/openjarvis/evals/tests/test_config.py

@@ -7,8 +7,8 @@ from pathlib import Path
 
 import pytest
 
-from evals.core.config import EvalConfigError, expand_suite, load_eval_config
-from evals.core.types import (
+from openjarvis.evals.core.config import EvalConfigError, expand_suite, load_eval_config
+from openjarvis.evals.core.types import (
     BenchmarkConfig,
     DefaultsConfig,
     EvalSuiteConfig,
@@ -597,7 +597,7 @@ class TestExpandSuite:
 class TestCLIConfig:
     def test_run_missing_benchmark_and_config(self):
         from click.testing import CliRunner
-        from evals.cli import main
+        from openjarvis.evals.cli import main
 
         runner = CliRunner()
         result = runner.invoke(main, ["run", "-m", "qwen3:8b"])
@@ -606,7 +606,7 @@ class TestCLIConfig:
 
     def test_run_missing_model_and_config(self):
         from click.testing import CliRunner
-        from evals.cli import main
+        from openjarvis.evals.cli import main
 
         runner = CliRunner()
         result = runner.invoke(main, ["run", "-b", "supergpqa"])
@@ -615,7 +615,7 @@ class TestCLIConfig:
 
     def test_run_config_file_not_found(self):
         from click.testing import CliRunner
-        from evals.cli import main
+        from openjarvis.evals.cli import main
 
         runner = CliRunner()
         result = runner.invoke(main, ["run", "--config", "/nonexistent.toml"])
@@ -630,7 +630,7 @@ class TestCLIConfig:
         from unittest.mock import patch
 
         from click.testing import CliRunner
-        from evals.cli import main
+        from openjarvis.evals.cli import main
 
         p = _write_toml(tmp_path, """\
             [meta]

src/openjarvis/evals/tests/test_display.py

@@ -6,7 +6,7 @@ from io import StringIO
 
 from rich.console import Console
 
-from evals.core.display import (
+from openjarvis.evals.core.display import (
     print_accuracy_panel,
     print_energy_table,
     print_latency_table,
@@ -14,7 +14,7 @@ from evals.core.display import (
     print_compact_table,
     print_full_results,
 )
-from evals.core.types import MetricStats, RunSummary
+from openjarvis.evals.core.types import MetricStats, RunSummary
 
 
 def _make_summary(**overrides) -> RunSummary:

src/openjarvis/evals/tests/test_frames.py

@@ -2,9 +2,9 @@
 
 from __future__ import annotations
 
-from evals.core.types import EvalRecord
-from evals.scorers.frames_judge import _GRADER_TEMPLATE, FRAMESScorer
-from evals.tests.conftest import MockBackend
+from openjarvis.evals.core.types import EvalRecord
+from openjarvis.evals.scorers.frames_judge import _GRADER_TEMPLATE, FRAMESScorer
+from openjarvis.evals.tests.conftest import MockBackend
 
 
 class TestGraderTemplate:

src/openjarvis/evals/tests/test_gaia.py

@@ -2,8 +2,8 @@
 
 from __future__ import annotations
 
-from evals.core.types import EvalRecord
-from evals.scorers.gaia_exact import (
+from openjarvis.evals.core.types import EvalRecord
+from openjarvis.evals.scorers.gaia_exact import (
     GAIAScorer,
     _is_float,
     _normalize_number_str,
@@ -11,7 +11,7 @@ from evals.scorers.gaia_exact import (
     _split_string,
     exact_match,
 )
-from evals.tests.conftest import MockBackend
+from openjarvis.evals.tests.conftest import MockBackend
 
 
 class TestNormalization:

src/openjarvis/evals/tests/test_runner.py

@@ -6,9 +6,13 @@ import json
 
 import pytest
 
-from evals.core.runner import EvalRunner, _metric_stats, _metric_stats_to_dict
-from evals.core.types import EvalRecord, MetricStats, RunConfig
-from evals.tests.conftest import MockBackend, MockDataset, MockScorer
+from openjarvis.evals.core.runner import (
+    EvalRunner,
+    _metric_stats,
+    _metric_stats_to_dict,
+)
+from openjarvis.evals.core.types import EvalRecord, MetricStats, RunConfig
+from openjarvis.evals.tests.conftest import MockBackend, MockDataset, MockScorer
 
 
 class TestEvalRunner:

src/openjarvis/evals/tests/test_supergpqa.py

@@ -2,8 +2,8 @@
 
 from __future__ import annotations
 
-from evals.core.types import EvalRecord
-from evals.tests.conftest import MockBackend
+from openjarvis.evals.core.types import EvalRecord
+from openjarvis.evals.tests.conftest import MockBackend
 
 
 class TestSuperGPQAScorer:
@@ -25,7 +25,7 @@ class TestSuperGPQAScorer:
         )
 
     def test_correct_extraction(self):
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
 
         backend = MockBackend(responses={})
         backend._default_response = "B"
@@ -39,7 +39,7 @@ class TestSuperGPQAScorer:
         assert meta["candidate_letter"] == "B"
 
     def test_incorrect_extraction(self):
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
 
         backend = MockBackend()
         backend._default_response = "A"
@@ -52,7 +52,7 @@ class TestSuperGPQAScorer:
         assert meta["candidate_letter"] == "A"
 
     def test_missing_reference(self):
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
 
         backend = MockBackend()
         scorer = SuperGPQAScorer(backend, "gpt-4o")
@@ -64,7 +64,7 @@ class TestSuperGPQAScorer:
         assert meta["reason"] == "missing_reference_letter"
 
     def test_no_extraction(self):
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
 
         backend = MockBackend()
         backend._default_response = "NONE"
@@ -77,7 +77,7 @@ class TestSuperGPQAScorer:
         assert meta["reason"] == "no_choice_letter_extracted"
 
     def test_valid_letters_from_options(self):
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
 
         backend = MockBackend()
         scorer = SuperGPQAScorer(backend, "gpt-4o")
@@ -92,7 +92,7 @@ class TestSuperGPQAScorer:
         assert letters == "ABCD"
 
     def test_extraction_with_verbose_response(self):
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
 
         backend = MockBackend()
         backend._default_response = "THE ANSWER IS: C"

src/openjarvis/evals/tests/test_types.py

@@ -2,7 +2,7 @@
 
 from __future__ import annotations
 
-from evals.core.types import (
+from openjarvis.evals.core.types import (
     BenchmarkConfig,
     DefaultsConfig,
     EvalRecord,

src/openjarvis/evals/tests/test_wildchat.py

@@ -2,9 +2,9 @@
 
 from __future__ import annotations
 
-from evals.core.types import EvalRecord
-from evals.scorers.wildchat_judge import WildChatScorer
-from evals.tests.conftest import MockBackend
+from openjarvis.evals.core.types import EvalRecord
+from openjarvis.evals.scorers.wildchat_judge import WildChatScorer
+from openjarvis.evals.tests.conftest import MockBackend
 
 
 class TestVerdictParsing:

tests/evals/test_benchmark_datasets.py

@@ -23,91 +23,91 @@ class TestDatasetInstantiation:
     """Verify each dataset class can be instantiated with correct attributes."""
 
     def test_supergpqa(self) -> None:
-        from evals.datasets.supergpqa import SuperGPQADataset
+        from openjarvis.evals.datasets.supergpqa import SuperGPQADataset
         ds = SuperGPQADataset()
         assert ds.dataset_id == "supergpqa"
         assert ds.dataset_name == "SuperGPQA"
 
     def test_gpqa(self) -> None:
-        from evals.datasets.gpqa import GPQADataset
+        from openjarvis.evals.datasets.gpqa import GPQADataset
         ds = GPQADataset()
         assert ds.dataset_id == "gpqa"
         assert ds.dataset_name == "GPQA"
 
     def test_mmlu_pro(self) -> None:
-        from evals.datasets.mmlu_pro import MMLUProDataset
+        from openjarvis.evals.datasets.mmlu_pro import MMLUProDataset
         ds = MMLUProDataset()
         assert ds.dataset_id == "mmlu-pro"
         assert ds.dataset_name == "MMLU-Pro"
 
     def test_math500(self) -> None:
-        from evals.datasets.math500 import MATH500Dataset
+        from openjarvis.evals.datasets.math500 import MATH500Dataset
         ds = MATH500Dataset()
         assert ds.dataset_id == "math500"
         assert ds.dataset_name == "MATH-500"
 
     def test_natural_reasoning(self) -> None:
-        from evals.datasets.natural_reasoning import NaturalReasoningDataset
+        from openjarvis.evals.datasets.natural_reasoning import NaturalReasoningDataset
         ds = NaturalReasoningDataset()
         assert ds.dataset_id == "natural-reasoning"
         assert ds.dataset_name == "Natural Reasoning"
 
     def test_hle(self) -> None:
-        from evals.datasets.hle import HLEDataset
+        from openjarvis.evals.datasets.hle import HLEDataset
         ds = HLEDataset()
         assert ds.dataset_id == "hle"
         assert ds.dataset_name == "HLE"
 
     def test_simpleqa(self) -> None:
-        from evals.datasets.simpleqa import SimpleQADataset
+        from openjarvis.evals.datasets.simpleqa import SimpleQADataset
         ds = SimpleQADataset()
         assert ds.dataset_id == "simpleqa"
         assert ds.dataset_name == "SimpleQA"
 
     def test_wildchat(self) -> None:
-        from evals.datasets.wildchat import WildChatDataset
+        from openjarvis.evals.datasets.wildchat import WildChatDataset
         ds = WildChatDataset()
         assert ds.dataset_id == "wildchat"
         assert ds.dataset_name == "WildChat"
 
     def test_ipw(self) -> None:
-        from evals.datasets.ipw_mixed import IPWDataset
+        from openjarvis.evals.datasets.ipw_mixed import IPWDataset
         ds = IPWDataset()
         assert ds.dataset_id == "ipw"
         assert ds.dataset_name == "IPW"
 
     def test_gaia(self) -> None:
-        from evals.datasets.gaia import GAIADataset
+        from openjarvis.evals.datasets.gaia import GAIADataset
         ds = GAIADataset()
         assert ds.dataset_id == "gaia"
         assert ds.dataset_name == "GAIA"
 
     def test_frames(self) -> None:
-        from evals.datasets.frames import FRAMESDataset
+        from openjarvis.evals.datasets.frames import FRAMESDataset
         ds = FRAMESDataset()
         assert ds.dataset_id == "frames"
         assert ds.dataset_name == "FRAMES"
 
     def test_swebench(self) -> None:
-        from evals.datasets.swebench import SWEBenchDataset
+        from openjarvis.evals.datasets.swebench import SWEBenchDataset
         ds = SWEBenchDataset()
         assert ds.dataset_id == "swebench"
         assert ds.dataset_name == "SWE-bench"
 
     def test_swefficiency(self) -> None:
-        from evals.datasets.swefficiency import SWEfficiencyDataset
+        from openjarvis.evals.datasets.swefficiency import SWEfficiencyDataset
         ds = SWEfficiencyDataset()
         assert ds.dataset_id == "swefficiency"
         assert ds.dataset_name == "SWEfficiency"
 
     def test_terminalbench(self) -> None:
-        from evals.datasets.terminalbench import TerminalBenchDataset
+        from openjarvis.evals.datasets.terminalbench import TerminalBenchDataset
         ds = TerminalBenchDataset()
         assert ds.dataset_id == "terminalbench"
         assert ds.dataset_name == "TerminalBench"
 
     def test_terminalbench_native(self) -> None:
-        from evals.datasets.terminalbench_native import (
+        from openjarvis.evals.datasets.terminalbench_native import (
             TerminalBenchNativeDataset,
         )
         ds = TerminalBenchNativeDataset()
@@ -131,74 +131,74 @@ class TestScorerInstantiation:
     """Verify each scorer class can be constructed."""
 
     def test_supergpqa_scorer(self) -> None:
-        from evals.scorers.supergpqa_mcq import SuperGPQAScorer
+        from openjarvis.evals.scorers.supergpqa_mcq import SuperGPQAScorer
         s = SuperGPQAScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "supergpqa"
 
     def test_gpqa_scorer(self) -> None:
-        from evals.scorers.gpqa_mcq import GPQAScorer
+        from openjarvis.evals.scorers.gpqa_mcq import GPQAScorer
         s = GPQAScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "gpqa"
 
     def test_mmlu_pro_scorer(self) -> None:
-        from evals.scorers.mmlu_pro_mcq import MMLUProScorer
+        from openjarvis.evals.scorers.mmlu_pro_mcq import MMLUProScorer
         s = MMLUProScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "mmlu-pro"
 
     def test_reasoning_judge_scorer(self) -> None:
-        from evals.scorers.reasoning_judge import ReasoningJudgeScorer
+        from openjarvis.evals.scorers.reasoning_judge import ReasoningJudgeScorer
         s = ReasoningJudgeScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "reasoning_judge"
 
     def test_hle_scorer(self) -> None:
-        from evals.scorers.hle_judge import HLEScorer
+        from openjarvis.evals.scorers.hle_judge import HLEScorer
         s = HLEScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "hle"
 
     def test_simpleqa_scorer(self) -> None:
-        from evals.scorers.simpleqa_judge import SimpleQAScorer
+        from openjarvis.evals.scorers.simpleqa_judge import SimpleQAScorer
         s = SimpleQAScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "simpleqa"
 
     def test_wildchat_scorer(self) -> None:
-        from evals.scorers.wildchat_judge import WildChatScorer
+        from openjarvis.evals.scorers.wildchat_judge import WildChatScorer
         s = WildChatScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "wildchat"
 
     def test_ipw_mixed_scorer(self) -> None:
-        from evals.scorers.ipw_mixed import IPWMixedScorer
+        from openjarvis.evals.scorers.ipw_mixed import IPWMixedScorer
         s = IPWMixedScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "ipw"
 
     def test_gaia_scorer(self) -> None:
-        from evals.scorers.gaia_exact import GAIAScorer
+        from openjarvis.evals.scorers.gaia_exact import GAIAScorer
         s = GAIAScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "gaia"
 
     def test_frames_scorer(self) -> None:
-        from evals.scorers.frames_judge import FRAMESScorer
+        from openjarvis.evals.scorers.frames_judge import FRAMESScorer
         s = FRAMESScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "frames"
 
     def test_swebench_scorer(self) -> None:
-        from evals.scorers.swebench_structural import SWEBenchScorer
+        from openjarvis.evals.scorers.swebench_structural import SWEBenchScorer
         s = SWEBenchScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "swebench"
 
     def test_swefficiency_scorer(self) -> None:
-        from evals.scorers.swefficiency_structural import (
+        from openjarvis.evals.scorers.swefficiency_structural import (
             SWEfficiencyScorer,
         )
         s = SWEfficiencyScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "swefficiency"
 
     def test_terminalbench_scorer(self) -> None:
-        from evals.scorers.terminalbench_judge import TerminalBenchScorer
+        from openjarvis.evals.scorers.terminalbench_judge import TerminalBenchScorer
         s = TerminalBenchScorer(_mock_backend(), "test-model")
         assert s.scorer_id == "terminalbench"
 
     def test_terminalbench_native_scorer(self) -> None:
-        from evals.scorers.terminalbench_native_structural import (
+        from openjarvis.evals.scorers.terminalbench_native_structural import (
             TerminalBenchNativeScorer,
         )
         s = TerminalBenchNativeScorer(_mock_backend(), "test-model")
@@ -222,7 +222,7 @@ class TestCLIFactories:
 
     @pytest.mark.parametrize("benchmark", ALL_BENCHMARKS)
     def test_build_dataset(self, benchmark: str) -> None:
-        from evals.cli import _build_dataset
+        from openjarvis.evals.cli import _build_dataset
         ds = _build_dataset(benchmark)
         assert ds is not None
         assert hasattr(ds, "load")
@@ -231,20 +231,20 @@ class TestCLIFactories:
 
     @pytest.mark.parametrize("benchmark", ALL_BENCHMARKS)
     def test_build_scorer(self, benchmark: str) -> None:
-        from evals.cli import _build_scorer
+        from openjarvis.evals.cli import _build_scorer
         scorer = _build_scorer(benchmark, _mock_backend(), "test-model")
         assert scorer is not None
         assert hasattr(scorer, "score")
 
     def test_build_dataset_unknown(self) -> None:
         import click
-        from evals.cli import _build_dataset
+        from openjarvis.evals.cli import _build_dataset
         with pytest.raises(click.ClickException, match="Unknown benchmark"):
             _build_dataset("nonexistent")
 
     def test_build_scorer_unknown(self) -> None:
         import click
-        from evals.cli import _build_scorer
+        from openjarvis.evals.cli import _build_scorer
         with pytest.raises(click.ClickException, match="Unknown benchmark"):
             _build_scorer("nonexistent", _mock_backend(), "test-model")
 
@@ -258,12 +258,12 @@ class TestConfigBenchmarks:
     """Verify KNOWN_BENCHMARKS includes all 15 benchmarks."""
 
     def test_all_benchmarks_known(self) -> None:
-        from evals.core.config import KNOWN_BENCHMARKS
+        from openjarvis.evals.core.config import KNOWN_BENCHMARKS
         for b in ALL_BENCHMARKS:
             assert b in KNOWN_BENCHMARKS, f"{b} missing from KNOWN_BENCHMARKS"
 
     def test_benchmarks_count(self) -> None:
-        from evals.core.config import KNOWN_BENCHMARKS
+        from openjarvis.evals.core.config import KNOWN_BENCHMARKS
         assert len(KNOWN_BENCHMARKS) == 15
 
 
@@ -276,8 +276,8 @@ class TestStructuralScorers:
     """Test structural scorers that don't need LLM calls."""
 
     def test_swebench_empty_response(self) -> None:
-        from evals.core.types import EvalRecord
-        from evals.scorers.swebench_structural import SWEBenchScorer
+        from openjarvis.evals.core.types import EvalRecord
+        from openjarvis.evals.scorers.swebench_structural import SWEBenchScorer
         scorer = SWEBenchScorer(_mock_backend(), "test-model")
         record = EvalRecord(
             record_id="swe-1", problem="Fix bug", reference="patch",
@@ -288,8 +288,8 @@ class TestStructuralScorers:
         assert meta["reason"] == "empty_response"
 
     def test_swebench_with_diff(self) -> None:
-        from evals.core.types import EvalRecord
-        from evals.scorers.swebench_structural import SWEBenchScorer
+        from openjarvis.evals.core.types import EvalRecord
+        from openjarvis.evals.scorers.swebench_structural import SWEBenchScorer
         scorer = SWEBenchScorer(_mock_backend(), "test-model")
         record = EvalRecord(
             record_id="swe-2", problem="Fix bug", reference="patch",
@@ -302,8 +302,8 @@ class TestStructuralScorers:
         assert meta["has_diff_markers"] is True
 
     def test_terminalbench_native_no_results(self) -> None:
-        from evals.core.types import EvalRecord
-        from evals.scorers.terminalbench_native_structural import (
+        from openjarvis.evals.core.types import EvalRecord
+        from openjarvis.evals.scorers.terminalbench_native_structural import (
             TerminalBenchNativeScorer,
         )
         scorer = TerminalBenchNativeScorer(_mock_backend(), "test-model")
@@ -316,8 +316,8 @@ class TestStructuralScorers:
         assert meta["reason"] == "no_test_results"
 
     def test_terminalbench_native_resolved(self) -> None:
-        from evals.core.types import EvalRecord
-        from evals.scorers.terminalbench_native_structural import (
+        from openjarvis.evals.core.types import EvalRecord
+        from openjarvis.evals.scorers.terminalbench_native_structural import (
             TerminalBenchNativeScorer,
         )
         scorer = TerminalBenchNativeScorer(_mock_backend(), "test-model")

tests/evals/test_display.py

@@ -1,11 +1,11 @@
-"""Tests for the Rich display helpers in evals.core.display."""
+"""Tests for the Rich display helpers in openjarvis.evals.core.display."""
 
 from __future__ import annotations
 
 from io import StringIO
 from pathlib import Path
 
-from evals.core.display import (
+from openjarvis.evals.core.display import (
     print_banner,
     print_completion,
     print_metrics_table,
@@ -14,7 +14,7 @@ from evals.core.display import (
     print_subject_table,
     print_suite_summary,
 )
-from evals.core.types import MetricStats, RunSummary
+from openjarvis.evals.core.types import MetricStats, RunSummary
 from rich.console import Console