verifiers 0.2.2.dev33__py3-none-any.whl → 0.2.2.dev34__py3-none-any.whl

verifiers/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from importlib.metadata import PackageNotFoundError
 from importlib.metadata import version as _version
 

verifiers/cli/commands/build.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Environment build command module for external hosts."""
 
 from verifiers.scripts.build import main

verifiers/cli/commands/eval.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Evaluation command module for external hosts."""
 
 from verifiers.scripts.eval import (

verifiers/cli/commands/gepa.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """GEPA command module for external hosts."""
 
 from verifiers.scripts.gepa import main

verifiers/cli/commands/init.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Environment initialization command module for external hosts."""
 
 from verifiers.scripts.init import main

verifiers/cli/commands/install.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Environment install command module for external hosts."""
 
 from verifiers.scripts.install import main

verifiers/cli/plugins/prime.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Prime-hosted command plugin contract."""
 
 from dataclasses import dataclass

verifiers/clients/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from verifiers.clients.anthropic_messages_client import AnthropicMessagesClient
 from verifiers.clients.client import Client
 from verifiers.clients.nemorl_chat_completions_client import (

verifiers/clients/anthropic_messages_client.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import functools
 import json
 import time

verifiers/clients/client.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import logging
 from abc import ABC, abstractmethod
 from typing import TYPE_CHECKING, Generic, TypeVar

verifiers/clients/openai_chat_completions_client.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import functools
 from collections.abc import Iterable, Mapping
 from typing import Any, TypeAlias, cast

verifiers/clients/openai_chat_completions_token_client.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from collections.abc import Mapping
 from typing import Any, Optional, cast
 

verifiers/clients/openai_completions_client.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from openai import (
     AsyncOpenAI,
 )

verifiers/clients/openai_responses_client.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import time
 from collections.abc import Iterable, Mapping
 from typing import Any, TypeAlias

verifiers/clients/renderer_client.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Renderer-based client.
 
 All tokenization happens client-side via a Renderer from the renderers package.

verifiers/decorators.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import inspect
 from typing import Any, Callable, Literal, TypeVar, overload
 

verifiers/envs/env_group.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import json
 from collections.abc import Mapping
 from typing import TYPE_CHECKING, Any, cast, final

verifiers/envs/environment.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import atexit
 import json

verifiers/envs/experimental/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from verifiers.envs.experimental.sandbox_mixin import SandboxMixin
 
 __all__ = [

verifiers/envs/experimental/cli_agent_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import logging
 import os

verifiers/envs/experimental/composable/README.md

@@ -30,13 +30,13 @@ from verifiers.envs.experimental.composable.harnesses.opencode import opencode_h
 from verifiers.envs.experimental.composable import ComposableEnv
 
 # Create a taskset
-taskset = R2EGymTaskSet()                 # 4578 SWE instances
+taskset = R2EGymTaskSet()  # 4578 SWE instances
 
 # Explore instances
-task = taskset[0]                          # one Task
-task.prompt                                # the problem statement
-task.sandbox_spec                          # SandboxSpec(image=..., cpu=4, ...)
-task.sandbox_spec.image                    # per-instance docker image
+task = taskset[0]  # one Task
+task.prompt  # the problem statement
+task.sandbox_spec  # SandboxSpec(image=..., cpu=4, ...)
+task.sandbox_spec.image  # per-instance docker image
 
 # Slice and filter
 small = taskset.take(100)
@@ -46,9 +46,9 @@ filtered = taskset.filter(lambda ex: ...)
 results = await taskset.validate(
     concurrency=50,
     out_path="outputs/validate.jsonl",
-    max_retries=2,       # retry on vf.InfraError
+    max_retries=2,  # retry on vf.InfraError
     sandbox_client_max_workers=100,  # optional sandbox client worker cap override
-    resume=True,         # skip indices already in out_path
+    resume=True,  # skip indices already in out_path
 )
 
 # Run with an agent
@@ -135,7 +135,9 @@ class MySWERubric(vf.Rubric):
         sandbox_client = state["sandbox_client"]
         sandbox_id = state["sandbox_id"]
         # Run tests in the sandbox
-        test_output = await self.taskset._run_tests(sandbox_client, sandbox_id, state, ...)
+        test_output = await self.taskset._run_tests(
+            sandbox_client, sandbox_id, state, ...
+        )
         return float(self.taskset._calculate_reward(test_output, info))
 
     @vf.cleanup

verifiers/envs/experimental/composable/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from verifiers.envs.experimental.composable.task import (
     SandboxSpec,
     Task,

verifiers/envs/experimental/composable/_filter.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Helper for the base ``TaskSet.filter_fn`` kwarg.
 
 ``filter_fn`` accepts a Python expression string (typically a lambda) that

verifiers/envs/experimental/composable/composable_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """ComposableEnv — a CliAgentEnv that delegates to a TaskSet + Harness.
 
 Subclasses ``CliAgentEnv`` and overrides its hooks to delegate to the

verifiers/envs/experimental/composable/harness.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Harness — agent-side configuration for ComposableEnv.
 
 A Harness declares how to install and run an agent binary, and where it

verifiers/envs/experimental/composable/harnesses/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from verifiers.envs.experimental.composable.harnesses.rlm import (
     DEFAULT_RLM_EXEC_TIMEOUT,
     DEFAULT_RLM_MAX_TURNS,

verifiers/envs/experimental/composable/harnesses/mini_swe_agent.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """mini-SWE-agent harness configuration."""
 
 from pathlib import PurePosixPath

verifiers/envs/experimental/composable/harnesses/rlm.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """RLM agent harness: install script, run command, and harness factory."""
 
 import hashlib

verifiers/envs/experimental/composable/sandbox_debug_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """No-agent debugger for sandbox-backed TaskSet instances."""
 
 import shlex

verifiers/envs/experimental/composable/swe_debug_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Deprecated SWE-specific debug environment wrappers."""
 
 from typing import Any

verifiers/envs/experimental/composable/task.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Task, TaskSet, SandboxTaskSet, and SandboxSpec — WHAT to solve.
 
 A **Task** is a single, fully-bound problem instance.

verifiers/envs/experimental/composable/tasksets/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from verifiers.envs.experimental.composable.tasksets.swe.swe_tasksets import (
     make_multiswe_taskset,
     make_openswe_taskset,

verifiers/envs/experimental/composable/tasksets/cp/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from .cp_task import CPRubric, CPTaskSet
 
 __all__ = ["CPTaskSet", "CPRubric"]

verifiers/envs/experimental/composable/tasksets/cp/cp_task.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Competitive Programming TaskSet.
 
 Usage::

verifiers/envs/experimental/composable/tasksets/cp/test_utils.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 # modifed from https://github.com/hendrycks/apps/blob/main/eval/testing_util.py
 # https://github.com/NovaSky-AI/SkyThought/blob/main/skythought/skythought_evals/tasks/taco/taco_util.py
 import asyncio

verifiers/envs/experimental/composable/tasksets/harbor/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from .harbor import (
     HarborDatasetRubric,
     HarborDatasetTaskSet,

verifiers/envs/experimental/composable/tasksets/harbor/harbor.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import json
 import logging
 import tarfile

verifiers/envs/experimental/composable/tasksets/harbor/terminal_lego.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Terminal-Lego taskset built on the composable Harbor task layout."""
 
 import contextlib

verifiers/envs/experimental/composable/tasksets/math/math_task.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Math TaskSet — QA-style math problems with sandbox-based scoring.
 
 Usage::

verifiers/envs/experimental/composable/tasksets/swe/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from .swe_tasksets import (
     make_multiswe_taskset,
     make_openswe_taskset,

verifiers/envs/experimental/composable/tasksets/swe/multi_swe/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import logging
 import shlex
 import tempfile

verifiers/envs/experimental/composable/tasksets/swe/openswe/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import logging
 import re
 import tempfile

verifiers/envs/experimental/composable/tasksets/swe/r2e_gym/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import json
 import logging
 import tempfile

verifiers/envs/experimental/composable/tasksets/swe/scale_swe/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Scale-SWE TaskSet.
 
 Scale-SWE (https://huggingface.co/datasets/AweAI-Team/Scale-SWE) is a

verifiers/envs/experimental/composable/tasksets/swe/shared/test_patch.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Revert-agent-test-edits + re-apply-test_patch helper.
 
 Mirrors the canonical SWE-bench pattern used by upstream's harness before

verifiers/envs/experimental/composable/tasksets/swe/swe_bench/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import json
 import logging
 import re

verifiers/envs/experimental/composable/tasksets/swe/swe_lego/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import json
 import logging
 import re

verifiers/envs/experimental/composable/tasksets/swe/swe_rebench_v2/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """SWE-rebench-V2 TaskSet.
 
 SWE-rebench-V2 (https://huggingface.co/datasets/nebius/SWE-rebench-V2) is a

verifiers/envs/experimental/composable/tasksets/swe/swe_smith/taskset.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """SWE-Smith multilingual TaskSet.
 
 SWE-Smith (https://github.com/SWE-bench/SWE-smith) is a synthetic bug-fix

verifiers/envs/experimental/gym_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from collections.abc import Callable
 from typing import Any, Protocol, TypeAlias, cast
 

verifiers/envs/experimental/harbor_env/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from .env import HarborEnv
 from .mcp import (
     NETWORK_TRANSPORTS,

verifiers/envs/experimental/harbor_env/env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import json
 import logging
 import tarfile

verifiers/envs/experimental/mcp_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import logging
 import threading

verifiers/envs/experimental/opencode_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import json
 import logging

verifiers/envs/experimental/opencode_rlm_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """
 OpenCode RLM Environment.
 

verifiers/envs/experimental/sandbox_mixin.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import io
 import logging

verifiers/envs/experimental/utils/file_locks.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from collections.abc import Iterator
 from contextlib import contextmanager
 import fcntl

verifiers/envs/experimental/utils/git_checkout_cache.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import fcntl
 import hashlib
 import logging

verifiers/envs/integrations/README.md

@@ -32,9 +32,18 @@ from datasets import Dataset
 import verifiers as vf
 
 # Create your dataset
-dataset = Dataset.from_list([
-    {"prompt": [{"role": "user", "content": "Navigate to example.com and find the main heading"}]},
-])
+dataset = Dataset.from_list(
+    [
+        {
+            "prompt": [
+                {
+                    "role": "user",
+                    "content": "Navigate to example.com and find the main heading",
+                }
+            ]
+        },
+    ]
+)
 
 # Create a rubric
 rubric = vf.Rubric(funcs=[my_reward_func])

verifiers/envs/integrations/browser_env/modes/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Browser mode implementations."""
 
 import warnings

verifiers/envs/integrations/browser_env/modes/base.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """Protocol defining the interface for browser mode implementations."""
 
 from typing import Protocol, Any, TYPE_CHECKING

verifiers/envs/integrations/browser_env/modes/cua_mode.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """CUA-based browser mode supporting both local HTTP and sandbox execution."""
 
 import asyncio

verifiers/envs/integrations/browser_env/modes/dom_mode.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """DOM-based browser mode using Stagehand SDK."""
 
 import asyncio

verifiers/envs/integrations/openenv_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import inspect
 import json

verifiers/envs/integrations/reasoninggym_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from typing import List, Tuple
 
 from datasets import Dataset

verifiers/envs/integrations/textarena_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import logging
 import random

verifiers/envs/multiturn_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import logging
 import time

verifiers/envs/sandbox_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import logging
 import time
 from typing import Any, TypedDict, cast

verifiers/envs/stateful_tool_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import inspect
 import json
 from abc import abstractmethod

verifiers/envs/tool_env.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import json
 from typing import Callable, cast
 

verifiers/errors.py

@@ -1,3 +1,6 @@
+# ruff: noqa
+
+
 class Error(Exception):
     """Base class for all errors."""
 

verifiers/gepa/__init__.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from verifiers.gepa.adapter import VerifiersGEPAAdapter, make_reflection_lm
 from verifiers.gepa.gepa_utils import save_gepa_results
 from verifiers.gepa.config import GEPAConfig

verifiers/gepa/adapter.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import logging
 from dataclasses import dataclass, field

verifiers/gepa/gepa_utils.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 """
 Simple artifact saving for GEPA optimization.
 

verifiers/parsers/maybe_think_parser.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from typing import Callable
 
 from verifiers.parsers.parser import Parser

verifiers/parsers/parser.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import logging
 from typing import Any, Callable
 

verifiers/parsers/think_parser.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from typing import Callable
 
 from verifiers.parsers.parser import Parser

verifiers/parsers/xml_parser.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import re
 from types import SimpleNamespace
 from typing import Any, Callable

verifiers/rubrics/experimental/hybrid_math_rubric.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 
 from openai import AsyncOpenAI

verifiers/rubrics/judge_rubric.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 from typing import Any
 
 from openai import APIError, APITimeoutError, AsyncOpenAI, RateLimitError

verifiers/rubrics/math_rubric.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import os
 import sys

verifiers/rubrics/rubric.py

@@ -1,3 +1,5 @@
+# ruff: noqa
+
 import asyncio
 import inspect
 import logging