User Manual: NoSQL¶

Basics¶

Data representation¶

Bingo NoSQL supports all formats which is supported by Indigo.

Storage¶

All databases data like chemical structures in binary format, fingerprints, hashes and other is stored in memory mapped files in folder that is defined by user at createDatabaseFile method parameter.

Queries¶

You can specify the query molecule as a string containing a Molfile (including various query features), a SMILES, or a SMARTS string. For reaction queries use Rxnfiles, reaction SMILES, or reaction SMARTS.

Molecules¶

Creating an Index¶

The following function creates the database for molecules:

bingo_db = Bingo.createDatabaseFile(indigo, 'db_dir', 'molecule', '')

bingo_db is the created database object
indigo is the indigo instance
'db_dir' is the path to the new database directory
'molecule' is the option which defines that database will contain molecules
'' is the string that can contain additional options
- min_mmf_size - first mmf file size in megabytes (not less than 32mb). Example: min_mmf_size:64
- min_mmf_size - largest mmf file size in megabytes. Example: max_mmf_size:1024
- read_only - if true no changes in database are accepted, only one thread can open database with read_only:false option. Example: read_only:true
- id_key - name of the records property that contain id. If that option is defined, then insert method will try to find records property with such name and set it as records id in database.``id_key:mol_id``

Updating and Closing Index¶

You can add or remove records in the database after it was created. No re-indexing is required after adding the records. To load and insert molecules to database you can do:

mol1 = indigo.loadMolecule('C1CCCCC1')
id1 = bingo_db.insert(mol1)
mol2 = indigo.loadMolecule('C1CCNCC1')
id2 = bingo_db.insert(mol2)

indigo is the indigo instance
bingo_db is the current database object
id1 and id2 are the indices of the added records. By these indices you can then access database records

Or you can insert record to database with your id:

bingo_db.insert(mol1, 100)

Exception will be thrown if id ‘100’ is already has been used.

The following function will remove record from database:

bingo_db.delete(id1)

To close database when you’ve done working with it call next function:

bingo_db.close()

Loading Index¶

The following function opens the database:

bingo_db = Bingo.loadDatabaseFile(indigo, 'db_dir', '')

bingo_db is the loaded database object
indigo is the indigo instance
'db_dir' is the path to the database directory
'' is the string that can contain additional options

Substructure Search¶

To run substructure search through your database you can do next couple of steps.

Load your query molecule:

query = indigo.loadQueryMolecule('CC')

Next function will create bingo substructure matcher object:

sub_matcher = bingo_db.searchSub(query, '')

'' is the matcher options
- part:<id of the part>/<number of parts> is the option for partial search. Example: part:1/5

And then iterate results with matchers next() method:

while sub_matcher.next():
    print(sub_matcher.getCurrentId())
exact_matcher.close()

getCurrentId() will return id of the matched record

Exact Search¶

Exact search can be done similarly as substructure search:

exact_matcher = bingo_db.searchExact(query, '')
while exact_matcher.next():
    print(exact_matcher.getCurrentId())
exact_matcher.close()

Similarity Search¶

The general form of similarity search query is as follows:

sim_matcher = bingo_db.searchSim(query, minSim, maxSim, sim_type)

Here is:

minSim and minSim are the real numbers, min and max bound of possible similarity value
bingo_db is the current database object
query is the query molecule
sim_type is the string that defines similarity measure type:
- 'tanimoto' - Tanimoto’s measure
- 'tversky' - Tversky’s measure
- 'euclid' - Euclid’s measure

The following loop will print all results of similarity matching:

cur_mol = sim_matcher.getIndigoObject()
while sim_matcher.next():
    print(sim_matcher.getCurrentId())
    print(sim_matcher.getCurrentSimilarityValue())
    print(cur_mol.smiles())
sim_matcher.close()

sim_matcher.getIndigoObject() will return object that will contain result target molecule at each iteration
sim_matcher.getCurrentId() will return current result record id
sim_matcher.getCurrentSimilarityValue() will return similarity value of current result and query
cur_mol.smiles() is the Indigo’s function that return smiles string for an object

Gross Formula Search¶

Gross formula search can be done as described below:

formula_matcher = bingo_db.searchMolFormula('C1CCNCC1', '')
while formula_matcher.next():
    print(formula_matcher.getCurrentId())
formula_matcher.close()

Where:

'C1CCNCC1' is the query formula

Reactions¶

Creating an Index¶

The following function creates the database for reactions:

bingo_db = Bingo.createDatabaseFile(indigo, 'db_dir', 'reaction', '')

bingo_db is the created database object
indigo is the indigo instance
'db_dir' is the path to the new database directory
'reaction' is the option which defines that database will contain reactions.
'' is the string that can contain additional options

Updating and Closing Index¶

You can add or remove records in the database after it was created. No re-indexing is required after adding the records. To load and insert reactions to database you can do:

rxn1 = indigo.loadReaction('C1CCCCC1>>CCC')
id1 = bingo_db.insert(rxn1)
rxn2 = indigo.loadReaction('C1CCNCC1>>CCN')
id2 = bingo_db.insert(rxn2)

indigo is the indigo instance.
bingo_db is the current database object
id1 and id2 are the indices of the added records. By these indices you can then access to database records.

Or you can insert record to database with your id:

bingo_db.insert(rxn1, 100)

Exception will be thrown if id ‘100’ is already has been used.

The following function will remove record from database:

bingo_db.delete(id1)

To close database when you’ve done working with it call next function:

bingo_db.close()

Loading Index¶

Follow the steps described in ‘Molecules’ section Loading Index

Reaction Substructure Search¶

To run substructure search through your database you can do next couple of steps.

Load your query reaction:

query = indigo.loadQueryMolecule('CC>>CN')

And then follow the steps described in ‘Molecules’ section Substructure Search

Reaction Exact Search¶

Follow the steps described in ‘Molecules’ section Exact Search

Reaction Similarity Search¶

Follow the steps described in ‘Molecules’ section Similarity Search

Reaction Gross Formula Search¶

Follow the steps described in ‘Molecules’ section Gross Formula Search

Remaining results estimation¶

Next functions will return remaining time and results count estimations:

sim_matcher.estimateRemainingResultsCount()
sim_matcher.estimateRemainingResultsCountError()
sim_matcher.estimateRemainingTime()))

Python example:¶

#Bingo database creating
bingo = Bingo.createDatabaseFile(indigo, 'db_dir', 'molecule', '')

# Molecules loading and inserting them to database
m = indigo.loadMolecule('C1CCCCC1')
bingo.insert(m)
m = indigo.loadMolecule('C1CCNCC1')
bingo.insert(m)

# Query molecule loading
qm = indigo.loadQueryMolecule('C')

# Bingo substructure matcher creation
matcher = bingo.searchSub(qm, '')

# Results iterating (robj will contain result target molecule on each iteration)
robj = matcher.getIndigoObject()
while matcher.next():
    print(matcher.getCurrentId())
    print(robj.smiles())

# Substructure matcher closing
matcher.close()

# Bingo database closing
bingo.close()

C Interface¶

Creation/loading/closing database.

int bingoCreateDatabaseFile (const char *location, const char *type, const char *options);

int bingoLoadDatabaseFile (const char *location, const char *options);

int bingoCloseDatabase (int db);

Insertion/Removing chemical structure

int bingoInsertRecordObj (int db, int obj);

int bingoInsertRecordObjWithId (int db, int obj, int id);

int bingoDeleteRecord (int db, int id);

Searching

int bingoSearchSub (int db, int query_obj, const char *options);

int bingoSearchExact (int db, int query_obj, const char *options);

int bingoSearchMolFormula (int db, const char *query, const char *options);

int bingoSearchSim (int db, int query_obj, float min, float max, const char *options);