Enable downloads from AWS Open Data for all collections (#884)

* Ensure json schema fields are retained

* Add direct s3 querying for regular docs

* Fix timeout passing

* Fix s3 threadsafety

* Fix for latest data re-org

* Use smart open

* Fix deserialization performance

* Switch to s3 for all doc queryies

* Fix remaining type issues

* More type fixes

* Remove print

* Fix deprecated field in materials and s3

* Fix type issues

* Fix structure import

* Fix open data json sanitization

* Add projection to multithreaded s3 func

* Remove default s3 query

* Add comments for clarity in s3 func

* Update pbar for s3 download

* Switch grain boundary prefix

* Fix s3 suffix delim

* Fix task retrieval support

* Linting

* Docstring noqa

* Add smart_open to deps

* Linting

* More linting

* Add missing docstring

* Docstring arg rename

* Fix tests

* Fix bs s3 query

* Fix chgcar query

* Fix dos query

* More linting

* Fix s3 keys

* More s3 fixes

* Last s3 query fix

* Linting

* Skip alloys generic test

* Fix alloys skip

* Remove task chemsys query

* Linting

* Fix similarity search name

* More similarity fixes

* Remove fermi rester

* Remove fermi ref

* More fermi fixes

* Linting

* Fix s3 decoding

* Add deprecated filter for s3

* Fix object key pagination

---------

Co-authored-by: Jason Munro <jasonmunro@Jasons-MBP.dhcp.lbl.gov>

mp_api/client/core/utils.py

@@ -73,6 +73,7 @@ def api_sanitize(
         model_fields_to_leave = {f[1] for f in fields_tuples if model.__name__ == f[0]}
         for name in model.model_fields:
             field = model.model_fields[name]
+            field_json_extra = field.json_schema_extra
             field_type = field.annotation
 
             if field_type is not None and allow_dict_msonable:
@@ -87,6 +88,7 @@ def api_sanitize(
                 new_field = FieldInfo.from_annotated_attribute(
                     Optional[field_type], None
                 )
+                new_field.json_schema_extra = field_json_extra or {}
                 model.model_fields[name] = new_field
 
         model.model_rebuild(force=True)

mp_api/client/mprester.py

@@ -1,17 +1,19 @@
 from __future__ import annotations
 
 import itertools
+import json
+import os
 import warnings
 from functools import cache, lru_cache
 from json import loads
-from os import environ
 from typing import Literal
 
 from emmet.core.electronic_structure import BSPathType
 from emmet.core.mpid import MPID
 from emmet.core.settings import EmmetSettings
 from emmet.core.tasks import TaskDoc
 from emmet.core.vasp.calc_types import CalcType
+from monty.json import MontyDecoder
 from packaging import version
 from pymatgen.analysis.phase_diagram import PhaseDiagram
 from pymatgen.analysis.pourbaix_diagram import IonEntry
@@ -39,7 +41,6 @@
     ElectrodeRester,
     ElectronicStructureRester,
     EOSRester,
-    FermiRester,
     GrainBoundaryRester,
     MagnetismRester,
     OxidationStatesRester,
@@ -66,10 +67,12 @@
 )
 
 _EMMET_SETTINGS = EmmetSettings()  # type: ignore
-_MAPI_SETTINGS = MAPIClientSettings()  # type: ignore
+_MAPI_SETTINGS = MAPIClientSettings()  # typeL ignore # type: ignore
 
-DEFAULT_API_KEY = environ.get("MP_API_KEY", None)
-DEFAULT_ENDPOINT = environ.get("MP_API_ENDPOINT", "https://api.materialsproject.org/")
+DEFAULT_API_KEY = os.environ.get("MP_API_KEY", None)
+DEFAULT_ENDPOINT = os.environ.get(
+    "MP_API_ENDPOINT", "https://api.materialsproject.org/"
+)
 
 
 class MPRester:
@@ -86,8 +89,7 @@ class MPRester:
     similarity: SimilarityRester
     tasks: TaskRester
     xas: XASRester
-    fermi: FermiRester
-    grain_boundary: GrainBoundaryRester
+    grain_boundaries: GrainBoundaryRester
     substrates: SubstratesRester
     surface_properties: SurfacePropertiesRester
     phonon: PhononRester
@@ -195,7 +197,7 @@ def __init__(
             "tasks",
             "xas",
             "fermi",
-            "grain_boundary",
+            "grain_boundaries",
             "substrates",
             "surface_properties",
             "phonon",
@@ -340,7 +342,7 @@ def __molecules_getattr__(_self, attr):
             return rester
 
         MaterialsRester.__getattr__ = __materials_getattr__  # type: ignore
-        MoleculeRester.__getattr__ = __molecules_getattr__  # type:  ignore
+        MoleculeRester.__getattr__ = __molecules_getattr__  # type: ignore
 
         for attr, rester in core_resters.items():
             setattr(
@@ -598,14 +600,15 @@ def get_structures(
             input_params = {"formula": chemsys_formula}
 
         if final:
-            return [
-                doc.structure if self.use_document_model else doc["structure"]  # type: ignore
-                for doc in self.materials.search(
-                    **input_params,  # type: ignore
-                    all_fields=False,
-                    fields=["structure"],
-                )
-            ]
+            docs = self.materials.search(
+                **input_params,  # type: ignore
+                all_fields=False,
+                fields=["structure"],
+            )
+            if not self.use_document_model:
+                return [doc["structure"] for doc in docs]  # type: ignore
+
+            return [doc.structure for doc in docs]  # type: ignore
         else:
             structures = []
 
@@ -614,11 +617,12 @@ def get_structures(
                 all_fields=False,
                 fields=["initial_structures"],
             ):
-                structures.extend(
+                initial_structures = (
                     doc.initial_structures  # type: ignore
                     if self.use_document_model
                     else doc["initial_structures"]  # type: ignore
                 )
+                structures.extend(initial_structures)
 
             return structures
 
@@ -1301,7 +1305,7 @@ def get_wulff_shape(self, material_id: str):
         )
         miller_energy_map = {}
         for surf in surfaces:
-            miller = tuple(surf.miller_index)
+            miller = tuple(surf.miller_index) if surf.miller_index else ()
             # Prefer reconstructed surfaces, which have lower surface energies.
             if (miller not in miller_energy_map) or surf.is_reconstructed:
                 miller_energy_map[miller] = surf.surface_energy
@@ -1339,20 +1343,22 @@ def get_charge_density_from_material_id(
             else x["last_updated"],  # type: ignore
         )
 
-        result = (
+        decoder = MontyDecoder().decode if self.monty_decode else json.loads
+        chgcar = (
             self.tasks._query_open_data(
                 bucket="materialsproject-parsed",
-                prefix="chgcars",
-                key=str(latest_doc.task_id),
-            )
+                key=f"chgcars/{str(latest_doc.task_id)}.json.gz",
+                decoder=decoder,
+                fields=["data"],
+            )[0]
             or {}
         )
 
-        chgcar = result.get("data", None)
-
-        if chgcar is None:
+        if not chgcar:
             raise MPRestError(f"No charge density fetched for {material_id}.")
 
+        chgcar = chgcar[0]["data"]  # type: ignore
+
         if inc_task_doc:
             task_doc = self.tasks.search(
                 task_ids=latest_doc.task_id
@@ -1384,7 +1390,7 @@ def get_download_info(self, material_ids, calc_types=None, file_patterns=None):
         )
 
         meta = {}
-        for doc in self.materials.search(
+        for doc in self.materials.search(  # type: ignore
             task_ids=material_ids,
             fields=["calc_types", "deprecated_tasks", "material_id"],
         ):

mp_api/client/routes/materials/__init__.py

@@ -13,8 +13,7 @@
     ElectronicStructureRester,
 )
 from .eos import EOSRester
-from .fermi import FermiRester
-from .grain_boundary import GrainBoundaryRester
+from .grain_boundaries import GrainBoundaryRester
 from .magnetism import MagnetismRester
 from .oxidation_states import OxidationStatesRester
 from .phonon import PhononRester

mp_api/client/routes/materials/electronic_structure.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import json
 import warnings
 from collections import defaultdict
 
@@ -8,6 +9,7 @@
     DOSProjectionType,
     ElectronicStructureDoc,
 )
+from monty.json import MontyDecoder
 from pymatgen.analysis.magnetism.analyzer import Ordering
 from pymatgen.core.periodic_table import Element
 from pymatgen.electronic_structure.core import OrbitalType, Spin
@@ -232,12 +234,16 @@ def get_bandstructure_from_task_id(self, task_id: str):
         Returns:
             bandstructure (BandStructure): BandStructure or BandStructureSymmLine object
         """
+        decoder = MontyDecoder().decode if self.monty_decode else json.loads
         result = self._query_open_data(
-            bucket="materialsproject-parsed", prefix="bandstructures", key=task_id
-        )
-
-        if result.get("data", None) is not None:
-            return result["data"]
+            bucket="materialsproject-parsed",
+            key=f"bandstructures/{task_id}.json.gz",
+            decoder=decoder,
+            fields=["data"],
+        )[0]
+
+        if result:
+            return result[0]["data"]
         else:
             raise MPRestError("No object found")
 
@@ -418,12 +424,16 @@ def get_dos_from_task_id(self, task_id: str):
         Returns:
             bandstructure (CompleteDos): CompleteDos object
         """
+        decoder = MontyDecoder().decode if self.monty_decode else json.loads
         result = self._query_open_data(
-            bucket="materialsproject-parsed", prefix="dos", key=task_id
-        )
-
-        if result.get("data", None) is not None:
-            return result["data"]
+            bucket="materialsproject-parsed",
+            key=f"dos/{task_id}.json.gz",
+            decoder=decoder,
+            fields=["data"],
+        )[0]
+
+        if result:
+            return result[0]["data"]  # type: ignore
         else:
             raise MPRestError("No object found")
 

mp_api/client/routes/materials/materials.py

@@ -19,7 +19,6 @@
     ElectrodeRester,
     ElectronicStructureRester,
     EOSRester,
-    FermiRester,
     GrainBoundaryRester,
     MagnetismRester,
     OxidationStatesRester,
@@ -50,8 +49,7 @@ class MaterialsRester(BaseRester[MaterialsDoc]):
         "similarity",
         "tasks",
         "xas",
-        "fermi",
-        "grain_boundary",
+        "grain_boundaries",
         "substrates",
         "surface_properties",
         "phonon",
@@ -81,7 +79,6 @@ class MaterialsRester(BaseRester[MaterialsDoc]):
     similarity: SimilarityRester
     tasks: TaskRester
     xas: XASRester
-    fermi: FermiRester
     grain_boundary: GrainBoundaryRester
     substrates: SubstratesRester
     surface_properties: SurfacePropertiesRester

mp_api/client/routes/materials/similarity.py

@@ -11,10 +11,9 @@ class SimilarityRester(BaseRester[SimilarityDoc]):
     document_model = SimilarityDoc  # type: ignore
     primary_key = "material_id"
 
-    def search_docs(
+    def search(
         self,
         material_ids: str | list[str] | None = None,
-        deprecated: bool | None = False,
         num_chunks: int | None = None,
         chunk_size: int = 1000,
         all_fields: bool = True,
@@ -25,7 +24,6 @@ def search_docs(
         Arguments:
             material_ids (str, List[str]): A single Material ID string or list of strings
                 (e.g., mp-149, [mp-149, mp-13]).
-            deprecated (bool): Whether the material is tagged as deprecated.
             num_chunks (int): Maximum number of chunks of data to yield. None will yield all possible.
             chunk_size (int): Number of data entries per chunk.
             all_fields (bool): Whether to return all fields in the document. Defaults to True.
@@ -35,7 +33,7 @@ def search_docs(
         Returns:
             ([SimilarityDoc], [dict]) List of similarity documents or dictionaries.
         """
-        query_params = {"deprecated": deprecated}  # type: dict
+        query_params = {}  # type: dict
 
         if material_ids:
             if isinstance(material_ids, str):

mp_api/client/routes/materials/tasks.py

@@ -36,7 +36,6 @@ def get_trajectory(self, task_id):
     def search(
         self,
         task_ids: str | list[str] | None = None,
-        chemsys: str | list[str] | None = None,
         elements: list[str] | None = None,
         exclude_elements: list[str] | None = None,
         formula: str | list[str] | None = None,
@@ -50,8 +49,6 @@ def search(
 
         Arguments:
             task_ids (str, List[str]): List of Materials Project IDs to return data for.
-            chemsys (str, List[str]): A chemical system or list of chemical systems
-                (e.g., Li-Fe-O, Si-*, [Si-O, Li-Fe-P]).
             elements (List[str]): A list of elements.
             exclude_elements (List[str]): A list of elements to exclude.
             formula (str, List[str]): A formula including anonymized formula
@@ -84,12 +81,6 @@ def search(
         if exclude_elements:
             query_params.update({"exclude_elements": ",".join(exclude_elements)})
 
-        if chemsys:
-            if isinstance(chemsys, str):
-                chemsys = [chemsys]
-
-            query_params.update({"chemsys": ",".join(chemsys)})
-
         if last_updated:
             query_params.update(
                 {

pyproject.toml

@@ -28,6 +28,7 @@ dependencies = [
   "requests>=2.23.0",
   "monty>=2023.9.25",
   "emmet-core>=0.78.0rc3",
+  "smart_open",
 ]
 dynamic = ["version"]