Creating a local MP database for fast querying.

create_local_mp.py

import pymongo
from pymatgen.ext.matproj import MPRester

# This connects to your local MongoDB without any credentials.
client = pymongo.MongoClient()
db = client.matproj
mpr = MPRester()
entries = mpr.get_entries(
    {},
    inc_structure=True,
    property_data=[  # These are entirely optional, but you can include any useful properties you want
        "band_gap",
        "e_above_hull",
        "icsd_ids"
    ],
)
docs = []
for e in entries:
    comp = e.composition
    elements = sorted(e.composition.keys())
    elements_str = sorted([el.symbol for el in elements])
    d = e.as_dict()
    d["pretty_formula"] = comp.reduced_formula
    d["elements"] = elements_str
    d["nelements"] = len(comp)
    d["chemsys"] = "-".join(elements_str)
    docs.append(d)

coll = db.entries
coll.insert_many(docs)

# These create useful indexes to speed up querying.
coll.create_index("entry_id")
coll.create_index("pretty_formula")
coll.create_index("chemsys")
coll.create_index("nelements")
coll.create_index("elements")