planbench.py

# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.


import json
import datasets


_CITATION = """\
@misc{valmeekam2023planbenchextensiblebenchmarkevaluating,
      title={PlanBench: An Extensible Benchmark for Evaluating Large Language Models on Planning and Reasoning about Change}, 
      author={Karthik Valmeekam and Matthew Marquez and Alberto Olmo and Sarath Sreedharan and Subbarao Kambhampati},
      year={2023},
      eprint={2206.10498},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2206.10498}, 
}
"""

_DESCRIPTION = """\
PlanBench is a benchmark for evaluating models' capabilities of planning and reasoning by evaluating them on IPC problems"""

_HOMEPAGE = "https://github.com/karthikv792/LLMs-Planning/tree/main/plan-bench"

_LICENSE = "MIT"


_URLS_prefix = {
    "blocksworld" : "https://raw.githubusercontent.com/karthikv792/LLMs-Planning/main/plan-bench/prompts/blocksworld",
    "blocksworld_3": "https://raw.githubusercontent.com/karthikv792/LLMs-Planning/main/plan-bench/prompts/blocksworld_3",
    "mystery_blocksworld": "https://raw.githubusercontent.com/karthikv792/LLMs-Planning/main/plan-bench/prompts/mystery_blocksworld",
    "mystery_blocksworld_3": "https://raw.githubusercontent.com/karthikv792/LLMs-Planning/main/plan-bench/prompts/mystery_blocksworld_3",
    "logistics": "https://raw.githubusercontent.com/karthikv792/LLMs-Planning/main/plan-bench/prompts/logistics",
}
_URLS = {
    "blocksworld_plan_generation": {
        "test": _URLS_prefix["blocksworld"] + "/task_1_plan_generation.json"
    },
    "blocksworld_plan_optimality": {
        "test": _URLS_prefix["blocksworld"] + "/task_2_plan_optimality.json"
    },
    "blocksworld_plan_verification": {
        "test": _URLS_prefix["blocksworld"] + "/task_3_plan_verification.json"
    },
    "blocksworld_plan_reuse": {
        "test": _URLS_prefix["blocksworld"] + "/task_4_plan_reuse.json"
    },
    "blocksworld_plan_generalization": {
        "test": _URLS_prefix["blocksworld"] + "/task_5_plan_generalization.json"
    },
    "blocksworld_replanning": {
        "test": _URLS_prefix["blocksworld"] + "/task_6_replanning.json"
    },
    "blocksworld_plan_execution": {
        "test": _URLS_prefix["blocksworld"] + "/task_7_plan_execution.json"
    },
    "blocksworld_goal_shuffling": {
        "test": _URLS_prefix["blocksworld"] + "/task_8_1_goal_shuffling.json"
    },
    "blocksworld_full_to_partial": {
        "test": _URLS_prefix["blocksworld"] + "/task_8_2_full_to_partial.json"
    },
    "blocksworld_partial_to_full": {
        "test": _URLS_prefix["blocksworld"] + "/task_8_3_partial_to_full.json"
    },
    "blocksworld_3_plan_generation": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_1_plan_generation.json"
    },
    "blocksworld_3_plan_optimality": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_2_plan_optimality.json"
    },
    "blocksworld_3_plan_verification": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_3_plan_verification.json"
    },
    "blocksworld_3_plan_reuse": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_4_plan_reuse.json"
    },
    "blocksworld_3_plan_generalization": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_5_plan_generalization.json"
    },
    "blocksworld_3_replanning": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_6_replanning.json"
    },
    "blocksworld_3_plan_execution": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_7_plan_execution.json"
    },
    "blocksworld_3_goal_shuffling": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_8_1_goal_shuffling.json"
    },
    "blocksworld_3_full_to_partial": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_8_2_full_to_partial.json"
    },
    "blocksworld_3_partial_to_full": {
        "test": _URLS_prefix["blocksworld_3"] + "/task_8_3_partial_to_full.json"
    },
    "mystery_blocksworld_plan_generation": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_1_plan_generation.json"
    },
    "mystery_blocksworld_plan_optimality": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_2_plan_optimality.json"
    },
    "mystery_blocksworld_plan_verification": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_3_plan_verification.json"
    },
    "mystery_blocksworld_plan_reuse": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_4_plan_reuse.json"
    },
    "mystery_blocksworld_plan_generalization": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_5_plan_generalization.json"
    },
    "mystery_blocksworld_replanning": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_6_replanning.json"
    },
    "mystery_blocksworld_plan_execution": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_7_plan_execution.json"
    },
    "mystery_blocksworld_goal_shuffling": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_8_1_goal_shuffling.json"
    },
    "mystery_blocksworld_full_to_partial": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_8_2_full_to_partial.json"
    },
    "mystery_blocksworld_partial_to_full": {
        "test": _URLS_prefix["mystery_blocksworld"] + "/task_8_3_partial_to_full.json"
    },
    "mystery_blocksworld_3_plan_generation": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_1_plan_generation.json"
    },
    "mystery_blocksworld_3_plan_optimality": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_2_plan_optimality.json"
    },
    "mystery_blocksworld_3_plan_verification": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_3_plan_verification.json"
    },
    "mystery_blocksworld_3_plan_reuse": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_4_plan_reuse.json"
    },
    "mystery_blocksworld_3_plan_generalization": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_5_plan_generalization.json"
    },
    "mystery_blocksworld_3_replanning": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_6_replanning.json"
    },
    "mystery_blocksworld_3_plan_execution": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_7_plan_execution.json"
    },
    "mystery_blocksworld_3_goal_shuffling": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_8_1_goal_shuffling.json"
    },
    "mystery_blocksworld_3_full_to_partial": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_8_2_full_to_partial.json"
    },
    "mystery_blocksworld_3_partial_to_full": {
        "test": _URLS_prefix["mystery_blocksworld_3"] + "/task_8_3_partial_to_full.json"
    },
    "logistics_plan_generation": {
        "test": _URLS_prefix["logistics"] + "/task_1_plan_generation.json"
    },
    "logistics_plan_optimality": {
        "test": _URLS_prefix["logistics"] + "/task_2_plan_optimality.json"
    },
    "logistics_plan_verification": {
        "test": _URLS_prefix["logistics"] + "/task_3_plan_verification.json"
    },
    "logistics_plan_reuse": {
        "test": _URLS_prefix["logistics"] + "/task_4_plan_reuse.json"
    },
    "logistics_plan_generalization": {
        "test": _URLS_prefix["logistics"] + "/task_5_plan_generalization.json"
    },
    "logistics_replanning": {
        "test": _URLS_prefix["logistics"] + "/task_6_replanning.json"
    },
    "logistics_plan_execution": {
        "test": _URLS_prefix["logistics"] + "/task_7_plan_execution.json"
    },
    "logistics_goal_shuffling": {
        "test": _URLS_prefix["logistics"] + "/task_8_1_goal_shuffling.json"
    },
    "logistics_full_to_partial": {
        "test": _URLS_prefix["logistics"] + "/task_8_2_full_to_partial.json"
    },
    "logistics_partial_to_full": {
        "test": _URLS_prefix["logistics"] + "/task_8_3_partial_to_full.json"
    }
}



class PlanBench(datasets.GeneratorBasedBuilder):
    """ LMentry is a benchmark for measuring language model performance on tasks that are trivial to humans.  LMentry consists of 25 tasks which humans are generally expected to perform perfectly, e.g. writing a sentence containing a specific word, identifying which words in a list belong to a specific category, choosing which of two words is longer, or identifying which of two words rhymes with a third word.
"""

    BUILDER_CONFIGS = [
        datasets.BuilderConfig(
            name=config_name,
            version=datasets.Version("0.0.1"),
            description=f"{config_name} task from PlanBench"
        )
        for config_name in _URLS.keys()
    ]
    def _info(self):
        features = {
            "instance_id": datasets.Value("int32"),
            "query": datasets.Value("string"),
            "ground_truth_plan": datasets.Value("string"),
        }
        if ("plan_generation" in self.config.name or
                "plan_optimality" in self.config.name or
                "plan_generalization" in self.config.name or
                "replanning" in self.config.name or
                "plan_execution" in self.config.name):
            features.update({"example_instance_ids": datasets.Sequence(datasets.Value("string"))})
        if "plan_reuse" in self.config.name or "replanning" in self.config.name:
            features.update({"new_instance": datasets.Value("string")})
        if "goal_shuffling" in self.config.name:
            features.update({"single_goal_instances": datasets.Value("int32")})
        features = datasets.Features(features)
        return datasets.DatasetInfo(
                description=_DESCRIPTION,
                features=features,
                homepage=_HOMEPAGE,
                citation=_CITATION,
                license=_LICENSE,
            )


    def _split_generators(self, dl_manager):
        urls = _URLS[self.config.name]
        data_dir = dl_manager.download_and_extract(urls)
        return [
            datasets.SplitGenerator(
                name = datasets.Split.TEST,
                gen_kwargs = {
                    "filepath" : data_dir["test"],
                    "split" : "test",
                }
            )
        ]


    def _generate_examples(self, filepath, split):
        with open(filepath, encoding = "utf-8") as fin :
            data = json.load(fin)
            for instance in data["instances"]:
                yield instance["instance_id"], instance