database Module

database Module

Defines a database object to store model results.

class pyanno.database.PyannoDatabase(filename)

Bases: object

Database to store model results.

The database is based on shelve. Keys are strings that uniquely identify data sets. Values are lists of PyannoResult objects, which contain a copy of the annotations, the pyanno model that has been applied on them, and the value of the log likelihood of the annotations given the model.

close()

Close database.

get_available_id()

Return an data ID that has is not present in the database.

The returned IDs have the form “<new_data_N>”, where N is an integer number.

remove(data_id, idx)

Remove entry from database.

Parameters:

• data_id (string) – Readable unique identifier for the annotations set
• idx (int) – Index in the list of entries with id data_id

retrieve_id(data_id)

Return all entries with given data ID.

Parameters:data_id (string) – Readable unique identifier for the annotations set

store_result(data_id, anno_container, model, value)

Store a pyAnno result in the database.

The data_id must be a unique identifier for an annotations set.

Parameters:

• data_id (string) – Readable unique identifier for the annotations set (e.g., the file name where the annotations are stored).
• anno_container (AnnotationsContainer) – An annotations container (see AnnotationsContainer).
• model (object) – pyAnno model object instance (subclass of AbstractModel
• value (float) – Value of the objective function for the model-annotations pair, typically the log likelihood of the annotations given the model

closed = None

True if the database is closed

database = None

shelve database storing the models

class pyanno.database.PyannoResult

Bases: traits.has_traits.HasStrictTraits

Class for database entries

anno_container = None

AnnotationsContainer object

model = None

pyAnno model (subclass of AbstractModel)

value = None

value of the model performance (usually the log likelihood)