qmlearn.io package
Submodules
qmlearn.io.core module
qmlearn.io.hdf5 module
- class qmlearn.io.hdf5.DBHDF5(filename, mode='a', qmmol=None)[source]
Bases:
objectClass to create and manipulate database in HDF5 format …
- Attributes
- filenamestr
Filename of database.
Methods
get_all_names([fh, attr])Return a list of all databases in the database file
get_names([name])Return a list of database name based on pattern
read_images([name])Read atomic structure from the database
read_properties(name, **kwargs)read properties from database file
read_qmmol(name, **kwargs)read qmmol object from the database
write_images([images, prefix, name])Write atomic structure in the database
write_properties([properties, prefix, name])Write properties (i.e.
write_qmmol([qmmol, name])Write qmmol object in the database
close
- get_all_names(fh=None, attr=None)[source]
Return a list of all databases in the database file
- Returns
- nameslist
return list of all database from the database file
- Attributes
- fhobj
Filehandler object for database file
- mode{‘a’,’r’,’w’} str
mode of operation (i.e. append, read, or write) for the database
- get_names(name='*')[source]
Return a list of database name based on pattern
- Returns
- nameslist
return list of all database from the database file
- Attributes
- namestr
pattern of database names. Only support ‘*’, ‘?’, ‘[seq]’ and ‘[!seq]’ with fnmatch
- property group
- property names
- property qmmol
Get qmmol object from database
- read_images(name=None, **kwargs)[source]
Read atomic structure from the database
- Returns
- imageslist
List of ASE Atoms objects
- Attributes
- namestr
name of the database file
- read_properties(name, **kwargs)[source]
read properties from database file
- Returns
- propertiesdict
return properties as dictionary
- Attributes
- namestr
name of the database file
- read_qmmol(name, **kwargs)[source]
read qmmol object from the database
- Returns
- qmmolobj
return qmmol object from the database
- Attributes
- namestr
name of the database file
- write_images(images=None, prefix='train', name=None, **kwargs)[source]
Write atomic structure in the database
- Attributes
- imageslist
List of ASE Atoms objects
- prefix{‘train’, ‘test’} str
determine wheter to write properties in the training or testing database
- namestr
name of the database file
- write_properties(properties=None, prefix='train', name=None, **kwargs)[source]
Write properties (i.e. gamma, energy, forces, Vext) to the database
- Attributes
- propertiesdict
Dictionary of properties calulate with Pyscf, Psi4numpy or any other external engine
- prefix{‘train’, ‘test’} str
determine wheter to write properties in the training or testing database
- namestr
name of the database file
qmlearn.io.model module
- qmlearn.io.model.db2qmmodel(filename, names='*', mmodels=None)[source]
Train QMModel to learn \({\gamma}\) in terms of \(V_{ext}\) from training data then an additional layer of training learn \({\delta}E\) and \({\delta}{\gamma}\) based on previously learned \({\gamma}\).
- Parameters
- filenamestr
Name of database file
- namesstr, optional
name of database, by default ‘*’
- mmodelsdict, optional
set of machine learning models used for training , If not provided by default KKR will be used to learn gamma and linear regression for delta learning
- Returns
- modelobj
trained model