The Spider execution plans are already present in the train and development splits in the dataset
folder. In order to work with them, do the following:
- Clone this repository.
- Create a new Python 3.9 virtual environment.
- Run
pip install -r requirements.txt
. - In order to read Spider instances, use the
get_train_dev_spider_instances()
function in theep_reader.py
module. TheSpiderInstance
dataclass is also defined there:
@dataclass(frozen=True)
class SpiderInstance:
db_id: str
query: str
question: str
ep: ExecutionPlan
The ExecutionPlan
type and its components are defined in ep_types.py
.