Molecule MongoDB

class MoleculeMongoDb(mongo_client, database='stk', molecule_collection='molecules', position_matrix_collection='building_block_position_matrices', jsonizer=MoleculeJsonizer((InchiKey())), dejsonizer=MoleculeDejsonizer(), put_lru_cache_size=128, get_lru_cache_size=128, indices=('InChIKey'))[source]

Bases: stk.databases.molecule.MoleculeDatabase

Uses MongoDB to store and retrieve molecules.

See also

ConstructedMoleculeMongoDb

You can store ConstructedMolecule instances with MoleculeMongoDb, however you can only retrieve them as Molecule instances. If you want to store and retrieve ConstructedMolecule instances, you should use ConstructedMoleculeMongoDb.

Examples

Storing and Retrieving Molecules

You want to store and retrieve a molecule from the database

import stk
import pymongo

# Connect to a MongoDB. This example connects to a local
# MongoDB, but you can connect to a remote DB too with
# MongoClient() - read the documentation for pymongo to see how
# to do that.
client = pymongo.MongoClient()
db = stk.MoleculeMongoDb(client)

# Create a molecule.
molecule = stk.BuildingBlock('NCCN')

# Place it into the database.
db.put(molecule)

# Retrieve it from the database.
key_maker = stk.InchiKey()
retrieved = db.get({
    key_maker.get_key_name(): key_maker.get_key(molecule)
})

Note that the molecule retrieved from that database can have a different atom ordering than the one put into it. So while the molecule will have the same structure, the order of the atoms may be different to the molecule placed into the database. This is because the database gives the molecules a canonical atom ordering, which allows position matrices to be used across different atom id orderings.

Iterating over All Entries in the Database

All entries in a database can be iterated over very simply

for entry in db.get_all():
    # Do something to the entry.
    print(stk.Smiles().get_key(entry))
NCCN

Using Alternative Keys for Retrieving Molecules

By default, the only molecular key the database stores, is the InChIKey. However, additional keys can be added to the JSON stored in the database by using a different MoleculeJsonizer

import pymongo

db = stk.MoleculeMongoDb(
    mongo_client=pymongo.MongoClient(),
    # Store the InChI and the InChIKey of molecules in
    # the JSON representation.
    jsonizer=stk.MoleculeJsonizer(
        key_makers=(stk.Inchi(), stk.InchiKey()),
    )
)

# Create a molecule.
molecule = stk.BuildingBlock('NCCN')

# Places the JSON of the molecule into the database. In this
# case, the JSON includes both the InChI and the InChIKey.
db.put(molecule)

# You can now use the InChI or the InChIKey to retrieve the
# molecule from the database.
key_maker = stk.Inchi()
retrieved = db.get({
    key_maker.get_key_name(): key_maker.get_key(molecule),
})

Obviously, most of the time, you won’t have the molecule you are trying to retrieve from the database. Maybe you only have the SMILES of the molecule. You can still retrieve it.

import rdkit.Chem.AllChem as rdkit

retrieved2 = db.get({
    'InChI': rdkit.MolToInchi(rdkit.MolFromSmiles('NCCN')),
})

As long as you have the name of the key, and the expected value of the key, you can retrieve your molecule from the database.

Note that you can create your own keys and add them to the database

# Create your own key. This one is called "SMILES" and the
# value is the SMILES of the molecule.
class Smiles(stk.MoleculeKeyMaker):
    def __init__(self):
        return

    def get_key_name(self):
        return 'SMILES'

    def get_key(self, molecule):
        return rdkit.MolToSmiles(molecule.to_rdkit_mol())

db = stk.MoleculeMongoDb(
    mongo_client=pymongo.MongoClient(),
    jsonizer=stk.MoleculeJsonizer(
        # Include your own custom key maker in the JSON
        # representation.
        key_makers = (stk.Inchi(), stk.InchiKey(), Smiles()),
    ),
)

molecule2 = stk.BuildingBlock('BrBr')

# Place the JSON of your molecule into the database. In this
# case the JSON will include a key called "SMILES" and
# the value will be the SMILES of the molecule.
db.put(molecule2)

# You can now find your molecule by using SMILES as the key.
retrieved3 = db.get({'SMILES': 'BrBr'})

Often, it is unnecessary to create a whole subclass for a your custom key

smiles = stk.MoleculeKeyMaker(
    key_name='SMILES',
    get_key=lambda molecule:
        rdkit.MolToSmiles(molecule.to_rdkit_mol()),
)
db = stk.MoleculeMongoDb(
    mongo_client=pymongo.MongoClient(),
    jsonizer=stk.MoleculeJsonizer(
        key_makers=(stk.InchiKey(), smiles),
    ),
)

Note that the key you use to get the molecule back from the database should be unique. In other words, there should always just be one molecule which has that key in the database. Using a key that is matched by multiple molecules will likely cause your data to be jumbled. For example, you might return the atoms of one of the molecules matched by the key but holding the position matrix of the second molecule matched by the key.

Methods

get(key)

Get the molecule with key from the database.

get_all()

Get all molecules in the database.

put(molecule)

Put molecule into the database.

__init__(mongo_client, database='stk', molecule_collection='molecules', position_matrix_collection='building_block_position_matrices', jsonizer=MoleculeJsonizer((InchiKey())), dejsonizer=MoleculeDejsonizer(), put_lru_cache_size=128, get_lru_cache_size=128, indices=('InChIKey'))[source]

Initialize a MoleculeMongoDb instance.

Parameters
  • mongo_client (pymongo.MongoClient) – The database client.

  • database (str) – The name of the database to use.

  • molecule_collection (str) – The name of the collection which stores molecular information.

  • position_matrix_collection (str) – The name of the collection which stores the position matrices of the molecules put into and retrieved from the database.

  • jsonizer (MoleculeJsonizer) – Used to create the JSON representations of molecules stored in the database.

  • dejsonizer (MoleculeDejsonizer) – Used to create Molecule instances from their JSON representations.

  • put_lru_cache_size (int, optional) – A RAM-based least recently used cache is used to avoid writing to the database repeatedly. This sets the number of values which fit into the LRU cache. If None, the cache size will be unlimited.

  • get_lru_cache_size (int, optional) – A RAM-based least recently used cache is used to avoid reading from the database repeatedly. This sets the number of values which fit into the LRU cache. If None, the cache size will be unlimited.

  • indices (tuple of str, optional) – The names of molecule keys, on which an index should be created, in order to minimize lookup time.

get(key)[source]

Get the molecule with key from the database.

Parameters

key (object) – The key of a molecule, which is to be returned from the database.

Returns

The molecule held in the database under key.

Return type

Molecule

Raises

KeyError – If key is not found in the database.

get_all()[source]

Get all molecules in the database.

Yields

Molecule – A molecule in the database.

put(molecule)[source]

Put molecule into the database.

Parameters

molecule (Molecule) – The molecule to place into the database.

Returns

None

Return type

NoneType