from __future__ import annotations
import pyarrow as pa
from emmet.core.arrow import arrowize
from emmet.core.synthesis import (
OperationTypeEnum,
SynthesisRecipe,
SynthesisSearchResultModel,
SynthesisTypeEnum,
)
from mp_api.client.core import BaseRester, MPRestError
[docs]
class SynthesisRester(BaseRester):
suffix = "materials/synthesis"
document_model = SynthesisSearchResultModel # type: ignore
def _download_schema(self) -> pa.Schema:
"""Full downloads hold recipes only: `search_score` and `highlights`
are added by the text search and aren't stored on S3 (they load as None).
"""
return pa.schema(arrowize(SynthesisRecipe))
[docs]
def search(
self,
keywords: list[str] | None = None,
synthesis_type: list[SynthesisTypeEnum] | None = None,
target_formula: str | None = None,
precursor_formula: str | None = None,
operations: list[OperationTypeEnum] | None = None,
condition_heating_temperature_min: float | None = None,
condition_heating_temperature_max: float | None = None,
condition_heating_time_min: float | None = None,
condition_heating_time_max: float | None = None,
condition_heating_atmosphere: list[str] | None = None,
condition_mixing_device: list[str] | None = None,
condition_mixing_media: list[str] | None = None,
num_chunks: int | None = None,
chunk_size: int | None = 10,
) -> list[SynthesisSearchResultModel] | list[dict]:
"""Search synthesis recipe text.
Arguments:
keywords (list[str] | None): List of string keywords to search synthesis paragraph text with
synthesis_type (list[SynthesisTypeEnum] | None): Type of synthesis to include
target_formula (str | None): Chemical formula of the target material
precursor_formula (str | None): Chemical formula of the precursor material
operations (list[OperationTypeEnum] | None): List of operations that syntheses must have
condition_heating_temperature_min (float | None): Minimal heating temperature
condition_heating_temperature_max (float | None): Maximal heating temperature
condition_heating_time_min (float | None): Minimal heating time
condition_heating_time_max (float | None): Maximal heating time
condition_heating_atmosphere (list[str] | None): Required heating atmosphere, such as "air", "argon"
condition_mixing_device (list[str] | None): Required mixing device, such as "zirconia", "Al2O3".
condition_mixing_media (list[str] | None): Required mixing media, such as "alcohol", "water"
num_chunks (int | None): Maximum number of chunks of data to yield. None will yield all possible.
chunk_size (int | None): Number of data entries per chunk.
With no search terms and `num_chunks=None`, every recipe is downloaded from
S3 as a local dataset instead; `search_score` and `highlights` are then None.
Returns:
([SynthesisSearchResultModel], [dict]): List of synthesis documents or dictionaries.
"""
# Turn None and empty list into None
keywords = keywords or None
synthesis_type = synthesis_type or None
operations = operations or None
condition_heating_atmosphere = condition_heating_atmosphere or None
condition_mixing_device = condition_mixing_device or None
condition_mixing_media = condition_mixing_media or None
if keywords:
keywords = ",".join([word.strip() for word in keywords]) # type: ignore
synthesis_docs = self._query_resource(
criteria={
"keywords": keywords,
"synthesis_type": synthesis_type,
"target_formula": target_formula,
"precursor_formula": precursor_formula,
"operations": operations,
"condition_heating_temperature_min": condition_heating_temperature_min,
"condition_heating_temperature_max": condition_heating_temperature_max,
"condition_heating_time_min": condition_heating_time_min,
"condition_heating_time_max": condition_heating_time_max,
"condition_heating_atmosphere": condition_heating_atmosphere,
"condition_mixing_device": condition_mixing_device,
"condition_mixing_media": condition_mixing_media,
"_limit": chunk_size,
},
chunk_size=chunk_size,
num_chunks=num_chunks,
).get("data", None)
if synthesis_docs is None:
raise MPRestError("Cannot find any matches.")
return synthesis_docs
