featurehub.admin package

Submodules

featurehub.admin.admin module

featurehub.admin.postprocessing module

exception featurehub.admin.postprocessing.TimeoutException[source]

Bases: Exception

featurehub.admin.postprocessing.append_feature_functions(features_df, inplace=False)[source]

Recover compiled functions and append column to DataFrame.

featurehub.admin.postprocessing.build_and_save_all_features(commands, session, suffix, splits=[], problem_names=[], features_on_disk=False)[source]

Build and save feature matrices.

>>> with orm.session_scope() as session:
        build_and_save_all_features(commands, session, suffix)
featurehub.admin.postprocessing.build_feature_matrix(features_df, dataset, group_id, group_feature_indices, feature_extraction_time_limit=40)[source]

Build feature matrix from human-generated features.

featurehub.admin.postprocessing.extract_and_save_all_tables(session, suffix)[source]
featurehub.admin.postprocessing.extract_table(session, mapper)[source]
featurehub.admin.postprocessing.load_dataset_from_dir(session, data_dir, problem_name)[source]
featurehub.admin.postprocessing.load_feature_matrix(problem_name, split, suffix)[source]
featurehub.admin.postprocessing.load_features_df(session, problem_name)[source]

Get all features for a specific problem as a DataFrame.

featurehub.admin.postprocessing.load_table(name)[source]
featurehub.admin.postprocessing.load_table1(name, suffix)[source]
featurehub.admin.postprocessing.null_feature(entities, name='null_feature', fill=0.0)[source]

Create null feature of an appropriate length.

featurehub.admin.postprocessing.prepare_automl_file_name(problem_name, split, suffix)[source]
featurehub.admin.postprocessing.recover_function(feature)[source]

Recover compiled function from Feature object.

featurehub.admin.postprocessing.save_feature_matrix(feature_matrix, problem_name, split, suffix)[source]
featurehub.admin.postprocessing.save_submission(df, problem_name, split_train, split_test, suffix)[source]
featurehub.admin.postprocessing.save_table(df, name)[source]
featurehub.admin.postprocessing.save_table1(df, name, suffix)[source]
featurehub.admin.postprocessing.time_limit(seconds)[source]

featurehub.admin.sqlalchemy_declarative module

class featurehub.admin.sqlalchemy_declarative.EvaluationAttempt(**kwargs)[source]

Bases: sqlalchemy.ext.declarative.api.Base

code
created_at
id
problem
problem_id
user
user_id
class featurehub.admin.sqlalchemy_declarative.Feature(**kwargs)[source]

Bases: sqlalchemy.ext.declarative.api.Base

code
created_at
description
feature_dill_quoted
id
md5
metrics
problem
problem_id
user
user_id
class featurehub.admin.sqlalchemy_declarative.Metric(**kwargs)[source]

Bases: sqlalchemy.ext.declarative.api.Base

created_at
feature
feature_id
id
name
scoring
value
class featurehub.admin.sqlalchemy_declarative.Problem(**kwargs)[source]

Bases: sqlalchemy.ext.declarative.api.Base

created_at
data_dir_test
data_dir_train
entities_featurized_table_name
entities_table_name
evaluationattempts
features
files
id
name
problem_type
problem_type_details
table_names
target_table_name
class featurehub.admin.sqlalchemy_declarative.User(**kwargs)[source]

Bases: sqlalchemy.ext.declarative.api.Base

created_at
evaluationattempts
features
id
name

featurehub.admin.sqlalchemy_main module

class featurehub.admin.sqlalchemy_main.ORMManager(database, admin=False)[source]

Bases: object

Initialize the sqlalchemy ORM engine and starts a database session.

database : string
Name of database to connect to
admin : bool, optional (default=False)
Whether to look up credentials in environment variables MYSQL_ROOT_USERNAME and MYSQL_ROOT_PASSWORD. If so, logs in as root user with admin permissions.
session_scope()[source]

Context manager to wrap a transaction.

>>> with orm.session_scope() as session:
        session.query(User.name).all()

Module contents