crfm-helm 0.5.6__py3-none-any.whl → 0.5.8__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of crfm-helm might be problematic. Click here for more details.
- {crfm_helm-0.5.6.dist-info → crfm_helm-0.5.8.dist-info}/METADATA +60 -125
- {crfm_helm-0.5.6.dist-info → crfm_helm-0.5.8.dist-info}/RECORD +293 -229
- helm/benchmark/adaptation/adapter_spec.py +5 -0
- helm/benchmark/adaptation/adapters/multimodal/multiple_choice_joint_multimodal_adapter.py +11 -3
- helm/benchmark/adaptation/adapters/multiple_choice_joint_adapter.py +11 -8
- helm/benchmark/annotation/aci_bench_annotator.py +11 -22
- helm/benchmark/annotation/air_bench_annotator.py +1 -1
- helm/benchmark/annotation/alrage_annotator.py +90 -0
- helm/benchmark/annotation/chw_care_plan_annotator.py +10 -21
- helm/benchmark/annotation/dischargeme_annotator.py +11 -22
- helm/benchmark/annotation/live_qa_annotator.py +1 -1
- helm/benchmark/annotation/med_dialog_annotator.py +11 -22
- helm/benchmark/annotation/medalign_annotator.py +11 -22
- helm/benchmark/annotation/medi_qa_annotator.py +11 -22
- helm/benchmark/annotation/medication_qa_annotator.py +11 -22
- helm/benchmark/annotation/mental_health_annotator.py +11 -22
- helm/benchmark/annotation/mimic_bhc_annotator.py +11 -22
- helm/benchmark/annotation/mimic_rrs_annotator.py +11 -22
- helm/benchmark/annotation/model_as_judge.py +23 -18
- helm/benchmark/annotation/mtsamples_procedures_annotator.py +11 -22
- helm/benchmark/annotation/mtsamples_replicate_annotator.py +11 -22
- helm/benchmark/annotation/starr_patient_instructions_annotator.py +11 -22
- helm/benchmark/metrics/air_bench_metrics.py +3157 -1
- helm/benchmark/metrics/alrage_metric.py +35 -0
- helm/benchmark/metrics/basic_metrics.py +267 -2
- helm/benchmark/metrics/classification_metrics.py +19 -1
- helm/benchmark/metrics/codeinsights_code_efficiency_metrics.py +186 -0
- helm/benchmark/metrics/codeinsights_code_evaluation_metrics.py +477 -0
- helm/benchmark/metrics/codeinsights_correct_code_metrics.py +366 -0
- helm/benchmark/metrics/codeinsights_edge_case_metrics.py +92 -0
- helm/benchmark/metrics/codeinsights_metric_specs.py +51 -0
- helm/benchmark/metrics/comet_metric.py +1 -1
- helm/benchmark/metrics/conv_fin_qa_calc_metrics.py +12 -1
- helm/benchmark/metrics/copyright_metrics.py +1 -1
- helm/benchmark/metrics/decodingtrust_stereotype_bias_metrics.py +1 -1
- helm/benchmark/metrics/dry_run_metrics.py +30 -1
- helm/benchmark/metrics/efficiency_metrics.py +74 -0
- helm/benchmark/metrics/ehr_sql_metrics.py +57 -1
- helm/benchmark/metrics/evaluate_reference_metrics.py +300 -1
- helm/benchmark/metrics/gpqa_chain_of_thought_metric.py +13 -1
- helm/benchmark/metrics/helpdesk_call_summarization_metrics.py +13 -1
- helm/benchmark/metrics/ifeval_metrics.py +13 -1
- helm/benchmark/metrics/image_generation/clip_score_metrics.py +13 -2
- helm/benchmark/metrics/image_generation/fractal_dimension/fractal_dimension_util.py +1 -1
- helm/benchmark/metrics/instruction_following_critique_metrics.py +41 -1
- helm/benchmark/metrics/kpi_edgar_metrics.py +21 -0
- helm/benchmark/metrics/language_modeling_metrics.py +13 -1
- helm/benchmark/metrics/live_qa_metrics.py +13 -1
- helm/benchmark/metrics/llm_jury_metrics.py +13 -1
- helm/benchmark/metrics/lmkt_metric_specs.py +12 -0
- helm/benchmark/metrics/lmkt_metrics.py +47 -0
- helm/benchmark/metrics/medcalc_bench_metrics.py +14 -1
- helm/benchmark/metrics/medec_metrics.py +25 -2
- helm/benchmark/metrics/melt_toxicity_metric.py +1 -1
- helm/benchmark/metrics/metric.py +25 -0
- helm/benchmark/metrics/mimiciv_billing_code_metrics.py +32 -1
- helm/benchmark/metrics/omni_math_metrics.py +13 -1
- helm/benchmark/metrics/seahelm_metrics.py +14 -1
- helm/benchmark/metrics/summac/model_summac.py +3 -3
- helm/benchmark/metrics/summarization_metrics.py +129 -1
- helm/benchmark/metrics/toxicity_metrics.py +31 -1
- helm/benchmark/metrics/wildbench_metrics.py +21 -1
- helm/benchmark/model_deployment_registry.py +11 -19
- helm/benchmark/presentation/create_plots.py +11 -2
- helm/benchmark/presentation/schema.py +10 -22
- helm/benchmark/presentation/summarize.py +189 -14
- helm/benchmark/presentation/taxonomy_info.py +20 -0
- helm/benchmark/presentation/test_create_plots.py +4 -1
- helm/benchmark/run.py +7 -1
- helm/benchmark/run_expander.py +4 -0
- helm/benchmark/run_specs/arabic_run_specs.py +191 -0
- helm/benchmark/run_specs/bluex_run_specs.py +40 -0
- helm/benchmark/run_specs/classic_run_specs.py +2 -55
- helm/benchmark/run_specs/codeinsights_run_specs.py +192 -0
- helm/benchmark/run_specs/healthqa_br_run_specs.py +40 -0
- helm/benchmark/run_specs/heim_run_specs.py +3 -1
- helm/benchmark/run_specs/lmkt_run_specs.py +144 -0
- helm/benchmark/run_specs/long_context_run_specs.py +48 -1
- helm/benchmark/run_specs/medhelm/__init__.py +0 -0
- helm/benchmark/run_specs/medhelm/benchmark_config.py +219 -0
- helm/benchmark/run_specs/medhelm_run_specs.py +360 -50
- helm/benchmark/run_specs/multilingual_run_specs.py +50 -0
- helm/benchmark/run_specs/speech_disorder_audio_run_specs.py +5 -11
- helm/benchmark/scenarios/aci_bench_scenario.py +23 -0
- helm/benchmark/scenarios/air_bench_scenario.py +21 -0
- helm/benchmark/scenarios/alghafa_scenario.py +126 -0
- helm/benchmark/scenarios/alrage_scenario.py +54 -0
- helm/benchmark/scenarios/anthropic_hh_rlhf_scenario.py +23 -1
- helm/benchmark/scenarios/arabic_exams_scenario.py +114 -0
- helm/benchmark/scenarios/arabic_mmlu_scenario.py +82 -0
- helm/benchmark/scenarios/aratrust_scenario.py +95 -0
- helm/benchmark/scenarios/audio_language/casual_conversations2_scenario.py +1 -1
- helm/benchmark/scenarios/audio_language/mustard_scenario.py +1 -1
- helm/benchmark/scenarios/audio_language/{ultra_suite_asr_classification.py → ultra_suite_asr_classification_scenario.py} +9 -8
- helm/benchmark/scenarios/audio_language/ultra_suite_asr_transcription_scenario.py +99 -0
- helm/benchmark/scenarios/audio_language/ultra_suite_classification_scenario.py +13 -5
- helm/benchmark/scenarios/audio_language/ultra_suite_disorder_breakdown_scenario.py +13 -5
- helm/benchmark/scenarios/audio_language/ultra_suite_disorder_symptoms_scenario.py +13 -5
- helm/benchmark/scenarios/babi_qa_scenario.py +15 -0
- helm/benchmark/scenarios/bbq_scenario.py +15 -0
- helm/benchmark/scenarios/best_chatgpt_prompts.yaml +473 -0
- helm/benchmark/scenarios/bluex_scenario.py +70 -0
- helm/benchmark/scenarios/bold_scenario.py +15 -0
- helm/benchmark/scenarios/boolq_scenario.py +20 -0
- helm/benchmark/scenarios/chw_care_plan_scenario.py +23 -0
- helm/benchmark/scenarios/civil_comments_scenario.py +13 -0
- helm/benchmark/scenarios/clear_scenario.py +23 -0
- helm/benchmark/scenarios/cleva_scenario.py +480 -1
- helm/benchmark/scenarios/code_scenario.py +28 -0
- helm/benchmark/scenarios/codeinsights_code_efficiency_scenario.py +197 -0
- helm/benchmark/scenarios/codeinsights_correct_code_scenario.py +78 -0
- helm/benchmark/scenarios/codeinsights_edge_case_scenario.py +192 -0
- helm/benchmark/scenarios/codeinsights_student_coding_scenario.py +162 -0
- helm/benchmark/scenarios/codeinsights_student_mistake_scenario.py +188 -0
- helm/benchmark/scenarios/commonsense_scenario.py +26 -0
- helm/benchmark/scenarios/compositional_instructions.yaml +70 -0
- helm/benchmark/scenarios/conv_fin_qa_calc_scenario.py +21 -0
- helm/benchmark/scenarios/copyright_scenario.py +35 -1
- helm/benchmark/scenarios/cti_to_mitre_scenario.py +21 -0
- helm/benchmark/scenarios/decodingtrust_adv_demonstration_scenario.py +22 -1
- helm/benchmark/scenarios/decodingtrust_adv_robustness_scenario.py +23 -1
- helm/benchmark/scenarios/decodingtrust_fairness_scenario.py +22 -1
- helm/benchmark/scenarios/decodingtrust_machine_ethics_scenario.py +21 -1
- helm/benchmark/scenarios/decodingtrust_ood_robustness_scenario.py +13 -0
- helm/benchmark/scenarios/decodingtrust_privacy_scenario.py +13 -1
- helm/benchmark/scenarios/decodingtrust_stereotype_bias_scenario.py +13 -1
- helm/benchmark/scenarios/decodingtrust_toxicity_prompts_scenario.py +13 -1
- helm/benchmark/scenarios/dischargeme_scenario.py +24 -0
- helm/benchmark/scenarios/disinformation_scenario.py +22 -0
- helm/benchmark/scenarios/dyck_language_scenario.py +15 -0
- helm/benchmark/scenarios/ehrshot_scenario.py +22 -0
- helm/benchmark/scenarios/enem_challenge_scenario.py +19 -0
- helm/benchmark/scenarios/entity_data_imputation_scenario.py +14 -0
- helm/benchmark/scenarios/entity_matching_scenario.py +14 -0
- helm/benchmark/scenarios/exams_multilingual_scenario.py +115 -0
- helm/benchmark/scenarios/financial_phrasebank_scenario.py +21 -0
- helm/benchmark/scenarios/gold_commodity_news_scenario.py +21 -0
- helm/benchmark/scenarios/gpqa_scenario.py +18 -0
- helm/benchmark/scenarios/grammar_scenario.py +20 -1
- helm/benchmark/scenarios/gsm_scenario.py +15 -0
- helm/benchmark/scenarios/headqa_scenario.py +22 -0
- helm/benchmark/scenarios/healthqa_br_scenario.py +80 -0
- helm/benchmark/scenarios/helpdesk_call_summarization_scenario.py +13 -0
- helm/benchmark/scenarios/ice_scenario.py +21 -1
- helm/benchmark/scenarios/ifeval_scenario.py +18 -0
- helm/benchmark/scenarios/imdb_scenario.py +15 -0
- helm/benchmark/scenarios/infinite_bench_en_mc_scenario.py +90 -0
- helm/benchmark/scenarios/infinite_bench_en_qa_scenario.py +1 -1
- helm/benchmark/scenarios/koala_scenario.py +21 -1
- helm/benchmark/scenarios/kpi_edgar_scenario.py +21 -0
- helm/benchmark/scenarios/legal_contract_summarization_scenario.py +20 -0
- helm/benchmark/scenarios/legal_summarization_scenario.py +50 -0
- helm/benchmark/scenarios/legal_support_scenario.py +13 -0
- helm/benchmark/scenarios/legalbench_scenario.py +20 -0
- helm/benchmark/scenarios/lex_glue_scenario.py +11 -0
- helm/benchmark/scenarios/lextreme_scenario.py +11 -0
- helm/benchmark/scenarios/lmkt_scenarios.py +288 -0
- helm/benchmark/scenarios/lsat_qa_scenario.py +14 -0
- helm/benchmark/scenarios/madinah_qa_scenario.py +73 -0
- helm/benchmark/scenarios/math_scenario.py +47 -20
- helm/benchmark/scenarios/mbzuai_human_translated_arabic_mmlu.py +68 -0
- helm/benchmark/scenarios/med_dialog_scenario.py +32 -1
- helm/benchmark/scenarios/med_mcqa_scenario.py +14 -0
- helm/benchmark/scenarios/med_qa_scenario.py +14 -0
- helm/benchmark/scenarios/medalign_scenario.py +23 -0
- helm/benchmark/scenarios/medalign_scenario_helper.py +19 -125
- helm/benchmark/scenarios/medbullets_scenario.py +22 -0
- helm/benchmark/scenarios/medcalc_bench_scenario.py +22 -0
- helm/benchmark/scenarios/medec_scenario.py +23 -0
- helm/benchmark/scenarios/medhallu_scenario.py +23 -0
- helm/benchmark/scenarios/medhelm/__init__.py +0 -0
- helm/benchmark/scenarios/medhelm/judges.yaml +14 -0
- helm/benchmark/scenarios/medhelm_configurable_scenario.py +101 -0
- helm/benchmark/scenarios/medi_qa_scenario.py +23 -0
- helm/benchmark/scenarios/medication_qa_scenario.py +31 -1
- helm/benchmark/scenarios/melt_scenarios.py +2 -2
- helm/benchmark/scenarios/mental_health_scenario.py +23 -0
- helm/benchmark/scenarios/mimic_bhc_scenario.py +25 -1
- helm/benchmark/scenarios/mimic_rrs_scenario.py +23 -0
- helm/benchmark/scenarios/mimiciv_billing_code_scenario.py +22 -0
- helm/benchmark/scenarios/mmlu_pro_scenario.py +18 -0
- helm/benchmark/scenarios/mmlu_scenario.py +15 -0
- helm/benchmark/scenarios/mmmlu_scenario.py +85 -0
- helm/benchmark/scenarios/msmarco_scenario.py +30 -0
- helm/benchmark/scenarios/mtsamples_procedures_scenario.py +22 -0
- helm/benchmark/scenarios/mtsamples_replicate_scenario.py +22 -0
- helm/benchmark/scenarios/n2c2_ct_matching_scenario.py +20 -0
- helm/benchmark/scenarios/narrativeqa_scenario.py +20 -0
- helm/benchmark/scenarios/natural_qa_scenario.py +32 -0
- helm/benchmark/scenarios/omni_math_scenario.py +18 -0
- helm/benchmark/scenarios/open_assistant_scenario.py +22 -0
- helm/benchmark/scenarios/pubmed_qa_scenario.py +22 -0
- helm/benchmark/scenarios/quac_scenario.py +14 -0
- helm/benchmark/scenarios/race_based_med_scenario.py +23 -0
- helm/benchmark/scenarios/raft_scenario.py +15 -0
- helm/benchmark/scenarios/real_toxicity_prompts_scenario.py +14 -1
- helm/benchmark/scenarios/scenario.py +31 -0
- helm/benchmark/scenarios/seahelm_scenario.py +350 -2
- helm/benchmark/scenarios/self_instruct_scenario.py +29 -1
- helm/benchmark/scenarios/shc_bmt_scenario.py +22 -0
- helm/benchmark/scenarios/shc_cdi_scenario.py +20 -0
- helm/benchmark/scenarios/shc_conf_scenario.py +23 -0
- helm/benchmark/scenarios/shc_ent_scenario.py +21 -0
- helm/benchmark/scenarios/shc_gip_scenario.py +20 -0
- helm/benchmark/scenarios/shc_privacy_scenario.py +22 -0
- helm/benchmark/scenarios/shc_proxy_scenario.py +22 -0
- helm/benchmark/scenarios/shc_ptbm_scenario.py +23 -0
- helm/benchmark/scenarios/shc_sequoia_scenario.py +21 -0
- helm/benchmark/scenarios/situation_prompts.yaml +49 -0
- helm/benchmark/scenarios/starr_patient_instructions_scenario.py +22 -0
- helm/benchmark/scenarios/summarization_scenario.py +37 -0
- helm/benchmark/scenarios/synthetic_efficiency_scenario.py +22 -1
- helm/benchmark/scenarios/synthetic_reasoning_natural_scenario.py +13 -0
- helm/benchmark/scenarios/test_alghafa_scenario.py +29 -0
- helm/benchmark/scenarios/test_alrage_scenario.py +23 -0
- helm/benchmark/scenarios/test_arabic_exams_scenario.py +21 -0
- helm/benchmark/scenarios/test_aratrust_scenario.py +21 -0
- helm/benchmark/scenarios/test_bluex_scenario.py +59 -0
- helm/benchmark/scenarios/test_exams_multilingual_scenario.py +29 -0
- helm/benchmark/scenarios/test_healtha_br_scenario.py +57 -0
- helm/benchmark/scenarios/the_pile_scenario.py +13 -1
- helm/benchmark/scenarios/truthful_qa_scenario.py +14 -0
- helm/benchmark/scenarios/twitter_aae_scenario.py +20 -1
- helm/benchmark/scenarios/vicuna_scenario.py +21 -1
- helm/benchmark/scenarios/wikifact_scenario.py +20 -0
- helm/benchmark/scenarios/wildbench_scenario.py +18 -0
- helm/benchmark/scenarios/wmt_14_scenario.py +12 -0
- helm/benchmark/slurm_jobs.py +1 -2
- helm/benchmark/slurm_runner.py +8 -1
- helm/benchmark/static/schema_arabic.yaml +271 -0
- helm/benchmark/static/schema_classic.yaml +0 -17
- helm/benchmark/static/schema_long_context.yaml +24 -6
- helm/benchmark/static/schema_medhelm.yaml +36 -0
- helm/benchmark/static/schema_slp.yaml +219 -0
- helm/benchmark/static_build/assets/index-671a5e06.js +10 -0
- helm/benchmark/static_build/assets/index-9352595e.css +1 -0
- helm/benchmark/static_build/index.html +2 -2
- helm/benchmark/window_services/image_generation/clip_window_service.py +1 -3
- helm/clients/audio_language/llama_omni/arguments.py +61 -0
- helm/clients/audio_language/llama_omni/constants.py +9 -0
- helm/clients/audio_language/llama_omni/conversation.py +213 -0
- helm/clients/audio_language/llama_omni/model/__init__.py +0 -0
- helm/clients/audio_language/llama_omni/model/builder.py +88 -0
- helm/clients/audio_language/llama_omni/model/language_model/omni_speech2s_llama.py +190 -0
- helm/clients/audio_language/llama_omni/model/language_model/omni_speech_llama.py +118 -0
- helm/clients/audio_language/llama_omni/model/omni_speech_arch.py +249 -0
- helm/clients/audio_language/llama_omni/model/speech_encoder/builder.py +9 -0
- helm/clients/audio_language/llama_omni/model/speech_encoder/speech_encoder.py +27 -0
- helm/clients/audio_language/llama_omni/model/speech_generator/builder.py +9 -0
- helm/clients/audio_language/llama_omni/model/speech_generator/generation.py +622 -0
- helm/clients/audio_language/llama_omni/model/speech_generator/speech_generator.py +104 -0
- helm/clients/audio_language/llama_omni/model/speech_projector/builder.py +9 -0
- helm/clients/audio_language/llama_omni/model/speech_projector/speech_projector.py +27 -0
- helm/clients/audio_language/llama_omni/preprocess.py +295 -0
- helm/clients/audio_language/llama_omni/utils.py +202 -0
- helm/clients/audio_language/qwen2_5_omni_client.py +19 -7
- helm/clients/audio_language/qwen_omni/configuration_qwen2_5_omni.py +519 -0
- helm/clients/audio_language/qwen_omni/modeling_qwen2_5_omni.py +4308 -0
- helm/clients/audio_language/qwen_omni/processing_qwen2_5_omni.py +270 -0
- helm/clients/audio_language/qwen_omni/qwen2_5_omni_utils/__init__.py +0 -0
- helm/clients/audio_language/qwen_omni/qwen2_5_omni_utils/v2_5/__init__.py +8 -0
- helm/clients/audio_language/qwen_omni/qwen2_5_omni_utils/v2_5/audio_process.py +56 -0
- helm/clients/audio_language/qwen_omni/qwen2_5_omni_utils/v2_5/vision_process.py +380 -0
- helm/clients/huggingface_client.py +2 -2
- helm/clients/image_generation/cogview2/sr_pipeline/dsr_model.py +1 -1
- helm/clients/image_generation/mindalle/models/stage1/layers.py +2 -2
- helm/clients/openai_client.py +33 -20
- helm/clients/openai_responses_client.py +34 -8
- helm/clients/openrouter_client.py +31 -0
- helm/clients/test_huggingface_client.py +3 -3
- helm/clients/test_openrouter_client.py +69 -0
- helm/clients/together_client.py +48 -13
- helm/clients/vertexai_client.py +19 -11
- helm/clients/vllm_client.py +43 -7
- helm/clients/vllm_granite_thinking_client.py +56 -0
- helm/common/critique_request.py +0 -1
- helm/common/hierarchical_logger.py +83 -34
- helm/common/object_spec.py +23 -8
- helm/common/test_logging.py +94 -0
- helm/config/model_deployments.yaml +525 -172
- helm/config/model_metadata.yaml +185 -10
- helm/config/tokenizer_configs.yaml +100 -2
- helm/proxy/cli.py +1 -1
- helm/proxy/example_queries.py +8 -8
- helm/proxy/retry.py +5 -0
- helm/proxy/server.py +2 -1
- helm/proxy/static/index.css +4 -0
- helm/proxy/static/index.js +7 -1
- helm/tokenizers/grok_tokenizer.py +2 -0
- helm/benchmark/metrics/aci_bench_metrics.py +0 -14
- helm/benchmark/metrics/chw_care_plan_metrics.py +0 -14
- helm/benchmark/metrics/dischargeme_metrics.py +0 -14
- helm/benchmark/metrics/med_dialog_metrics.py +0 -14
- helm/benchmark/metrics/medalign_metrics.py +0 -14
- helm/benchmark/metrics/medi_qa_metrics.py +0 -14
- helm/benchmark/metrics/medication_qa_metrics.py +0 -14
- helm/benchmark/metrics/mental_health_metrics.py +0 -14
- helm/benchmark/metrics/mimic_bhc_metrics.py +0 -14
- helm/benchmark/metrics/mimic_rrs_metrics.py +0 -14
- helm/benchmark/metrics/mtsamples_procedures_metrics.py +0 -14
- helm/benchmark/metrics/mtsamples_replicate_metrics.py +0 -14
- helm/benchmark/metrics/numeracy_metrics.py +0 -72
- helm/benchmark/metrics/starr_patient_instructions_metrics.py +0 -14
- helm/benchmark/metrics/test_numeracy_metrics.py +0 -95
- helm/benchmark/scenarios/numeracy_scenario.py +0 -794
- helm/benchmark/static_build/assets/index-94295e78.js +0 -10
- helm/benchmark/static_build/assets/index-b9779128.css +0 -1
- {crfm_helm-0.5.6.dist-info → crfm_helm-0.5.8.dist-info}/WHEEL +0 -0
- {crfm_helm-0.5.6.dist-info → crfm_helm-0.5.8.dist-info}/entry_points.txt +0 -0
- {crfm_helm-0.5.6.dist-info → crfm_helm-0.5.8.dist-info}/licenses/LICENSE +0 -0
- {crfm_helm-0.5.6.dist-info → crfm_helm-0.5.8.dist-info}/top_level.txt +0 -0
|
@@ -0,0 +1,94 @@
|
|
|
1
|
+
import sys
|
|
2
|
+
import tempfile
|
|
3
|
+
import textwrap
|
|
4
|
+
import pathlib
|
|
5
|
+
from helm.benchmark import run
|
|
6
|
+
from typing import List, Optional
|
|
7
|
+
|
|
8
|
+
|
|
9
|
+
class ArgvContext:
|
|
10
|
+
"""
|
|
11
|
+
Helper to assign a temporary value to sys.argv and then restore it
|
|
12
|
+
"""
|
|
13
|
+
|
|
14
|
+
def __init__(self, argv: Optional[List[str]]):
|
|
15
|
+
self.argv = argv
|
|
16
|
+
self._original_argv: Optional[List[str]] = None
|
|
17
|
+
|
|
18
|
+
def __enter__(self):
|
|
19
|
+
self._original_argv = sys.argv[:]
|
|
20
|
+
sys.argv = self.argv or []
|
|
21
|
+
|
|
22
|
+
def __exit__(self, exc_type, exc_val, exc_tb):
|
|
23
|
+
assert self._original_argv is not None # Satisfies mypy
|
|
24
|
+
sys.argv = self._original_argv
|
|
25
|
+
|
|
26
|
+
|
|
27
|
+
def test_run_with_custom_logging_config():
|
|
28
|
+
# Setup temporary directory
|
|
29
|
+
with tempfile.TemporaryDirectory(prefix="helm_test_") as tmp_dir_str:
|
|
30
|
+
tmp_dir = pathlib.Path(tmp_dir_str)
|
|
31
|
+
log_path = tmp_dir / "test.log"
|
|
32
|
+
log_config_path = tmp_dir / "test_config.yaml"
|
|
33
|
+
|
|
34
|
+
# Write custom YAML log config to file
|
|
35
|
+
log_config_text = textwrap.dedent(
|
|
36
|
+
f"""
|
|
37
|
+
version: 1
|
|
38
|
+
disable_existing_loggers: false
|
|
39
|
+
formatters:
|
|
40
|
+
simple:
|
|
41
|
+
datefmt: '%Y-%m-%dT%H:%M:%S'
|
|
42
|
+
format: '%(asctime)s %(levelname)s %(name)s %(message)s'
|
|
43
|
+
handlers:
|
|
44
|
+
file:
|
|
45
|
+
class: logging.FileHandler
|
|
46
|
+
filename: {log_path}
|
|
47
|
+
formatter: simple
|
|
48
|
+
level: DEBUG
|
|
49
|
+
mode: w
|
|
50
|
+
loggers:
|
|
51
|
+
helm:
|
|
52
|
+
handlers:
|
|
53
|
+
- file
|
|
54
|
+
level: DEBUG
|
|
55
|
+
propagate: false
|
|
56
|
+
"""
|
|
57
|
+
).strip()
|
|
58
|
+
|
|
59
|
+
log_config_path.write_text(log_config_text)
|
|
60
|
+
|
|
61
|
+
# Simulate command-line arguments
|
|
62
|
+
argv = [
|
|
63
|
+
"run.py", # Fake script name
|
|
64
|
+
"--run-entries",
|
|
65
|
+
"mmlu:subject=philosophy,model=openai/gpt2",
|
|
66
|
+
"-m",
|
|
67
|
+
"1",
|
|
68
|
+
"--suite",
|
|
69
|
+
"my-suite",
|
|
70
|
+
"--dry-run",
|
|
71
|
+
"--log-config",
|
|
72
|
+
str(log_config_path),
|
|
73
|
+
]
|
|
74
|
+
|
|
75
|
+
# Call main
|
|
76
|
+
with ArgvContext(argv):
|
|
77
|
+
run.main()
|
|
78
|
+
|
|
79
|
+
# Check log file contents
|
|
80
|
+
assert log_path.exists(), "Log file was not created"
|
|
81
|
+
log_contents = log_path.read_text()
|
|
82
|
+
|
|
83
|
+
# Test that log file was written to disk as requested
|
|
84
|
+
print("Log Contents")
|
|
85
|
+
print("------------")
|
|
86
|
+
print(log_contents)
|
|
87
|
+
|
|
88
|
+
assert (
|
|
89
|
+
"mscoco" in log_contents or "huggingface" in log_contents or "dry-run" in log_contents
|
|
90
|
+
), "Expected log content not found in log file:\n"
|
|
91
|
+
|
|
92
|
+
|
|
93
|
+
if __name__ == "__main__":
|
|
94
|
+
test_run_with_custom_logging_config()
|