def create_parser(general_defaults={}, constants={}):
"""Sets the accepted command options, variables, defaults and help
"""
now = constants.get('NOW',
datetime.datetime.now().strftime("%a%b%d%y_%H%M%S"))
max_models = constants.get('MAX_MODELS')
plurality = constants.get('PLURALITY')
last = constants.get('LAST_PREDICTION')
defaults = general_defaults['BigMLer']
version = pkg_resources.require("BigMLer")[0].version
version_text = """\
BigMLer %s - A Higher Level API to BigML's API
Copyright 2012 BigML
Licensed under the Apache License, Version 2.0 (the \"License\"); you may
not use this file except in compliance with the License. You may obtain
a copy of the License at http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an \"AS IS\" BASIS, WITHOUT
WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
License for the specific language governing permissions and limitations
under the License.""" % version
main_parser = argparse.ArgumentParser(
description="A higher level API to BigML's API.",
epilog="Happy predictive modeling!",
version=version_text,
formatter_class=argparse.RawTextHelpFormatter)
subparsers = main_parser.add_subparsers()
parser = subparsers.add_parser('main')
# list of common options
common_options_set = common_options(parser, defaults=defaults)
# Path to the test set.
parser.add_argument('--test',
action='store',
dest='test_set',
nargs='?',
default=defaults.get('test', None),
help="Test set path.")
# Name of the file to output predictions.
parser.add_argument('--output',
action='store',
dest='predictions',
default=defaults.get('output', None),
help="Path to the file to output predictions.")
# Category code.
parser.add_argument('--category',
action='store',
dest='category',
default=defaults.get('category', 12),
type=int,
help="Category code.")
# A file including a makdown description
parser.add_argument('--description',
action='store',
dest='description',
default=defaults.get('description', None),
help=("Path to a file with a description in plain"
" text or markdown."))
# The path to a file containing names if you want to alter BigML's
# default field names or the ones provided by the train file header.
# Kept for backwards compatibility
parser.add_argument('--field-names',
action='store',
dest='field_attributes',
default=defaults.get('field_names', None),
help=("Path to a csv file describing field names. One"
" definition per line (e.g., 0,'Last Name')."))
# The path to a file containing attributes if you want to alter BigML's
# default field attributes or the ones provided by the train file header.
parser.add_argument('--field-attributes',
action='store',
dest='field_attributes',
default=defaults.get('field_attributes', None),
help=("Path to a csv file describing field attributes."
" One definition per line"
" (e.g., 0,'Last Name')."))
# The path to a file containing types if you want to alter BigML's
# type auto-detect.
parser.add_argument('--types',
action='store',
dest='types',
default=defaults.get('types', None),
help=("Path to a file describing field types. One"
" definition per line (e.g., 0, 'numeric')."))
# Fields to include in the dataset.
parser.add_argument('--dataset-fields',
action='store',
dest='dataset_fields',
default=defaults.get('dataset_fields', None),
help=("Comma-separated list of field column numbers"
" to include in the dataset."))
# Path to a file that includes a JSON filter.
parser.add_argument('--json-filter',
action='store',
dest='json_filter',
default=defaults.get('json_filter', None),
help="File including a JSON filter.")
# Path to a file that includes a lisp filter.
parser.add_argument('--lisp-filter',
action='store',
dest='lisp_filter',
default=defaults.get('lisp_filter', None),
help="File including a Lisp filter.")
# Set when the training set file doesn't include a header on the first
# line.
parser.add_argument('--no-train-header',
action='store_false',
dest='train_header',
default=defaults.get('train_header', True),
help="The train set file hasn't a header.")
# Set when the test set file doesn't include a header on the first
# line.
parser.add_argument('--no-test-header',
action='store_false',
dest='test_header',
default=defaults.get('test_header', True),
help="The test set file hasn't a header.")
# Name to be used with the source and then with datasets, models and
# predictions.
parser.add_argument('--name',
action='store',
dest='name',
default=defaults.get('name', 'BigMLer_%s' % now),
help="Name for the resources in BigML.")
# If a BigML dataset is provided, the script won't create a new one
parser.add_argument('--dataset',
action='store',
dest='dataset',
default=defaults.get('dataset', None),
help="BigML dataset Id.")
# If a BigML model is provided, the script will use it to generate
# predictions.
parser.add_argument('--model',
action='store',
dest='model',
default=defaults.get('model', None),
help="BigML model Id.")
# Use it to compute predictions remotely.
parser.add_argument('--remote',
action='store_true',
dest='remote',
default=defaults.get('remote', False),
help="Compute predictions remotely.")
# The path to a file containing model ids.
parser.add_argument('--models',
action='store',
dest='models',
default=defaults.get('models', None),
help=("Path to a file containing model/ids. One model"
" per line (e.g., model/50a206a8035d0706dc000376"
")."))
# The path to a file containing dataset ids.
parser.add_argument('--datasets',
action='store',
dest='datasets',
default=defaults.get('datasets', None),
help=("Path to a file containing dataset/ids. Just"
" one dataset per line"
" (e.g., dataset/50a20697035d0706da0004a4)."))
# Sets pruning.
parser.add_argument('--pruning',
action='store',
default=defaults.get('pruning', "smart"),
choices=["smart", "statistical", "no-pruning"],
help=("Set pruning type: smart, statistical,"
" no-pruning."))
# Number of models to create when using ensembles.
parser.add_argument('--number-of-models',
action='store',
dest='number_of_models',
default=defaults.get('number_of_models', 1),
type=int,
help=("Number of models to create when using"
" ensembles."))
# Sampling to use when using bagging.
parser.add_argument('--sample-rate',
action='store',
dest='sample_rate',
default=defaults.get('sample_rate', 1.0),
type=float,
help="Sample rate to create models.")
# Replacement to use when using bagging.
parser.add_argument('--replacement',
action='store_true',
default=defaults.get('replacement', False),
help="Use replacement when sampling.")
# Max number of models to predict from in parallel.
parser.add_argument('--max-batch-models',
action='store',
dest='max_batch_models',
default=defaults.get('max_batch_models', max_models),
type=int,
help=("Max number of models to predict from"
" in parallel."))
# Randomize feature selection at each split.
parser.add_argument('--randomize',
action='store_true',
dest='randomize',
default=defaults.get('randomize', False),
help="Randomize feature selection at each split.")
# Use it to retrieve models that were tagged with tag.
parser.add_argument('--model-tag',
dest='model_tag',
default=defaults.get('model_tag', None),
help="Retrieve models that were tagged with tag.")
# Make dataset public.
parser.add_argument('--public-dataset',
action='store_true',
dest='public_dataset',
default=defaults.get('public_dataset', False),
help="Make generated dataset public.")
# Make model a public black-box model.
parser.add_argument('--black-box',
action='store_true',
dest='black_box',
default=defaults.get('black_box', False),
help="Make generated model black-box.")
# Make model a public white-box model.
parser.add_argument('--white-box',
action='store_true',
dest='white_box',
default=defaults.get('white_box', False),
help="Make generated model white-box.")
# Set a price tag to your white-box model.
parser.add_argument('--model-price',
action='store',
dest='model_price',
type=float,
default=defaults.get('model_price', 0.0),
help=("The price other users must pay to clone your"
" model."))
# Set a price tag to your dataset.
parser.add_argument('--dataset-price',
action='store',
dest='dataset_price',
type=float,
default=defaults.get('dataset_price', 0.0),
help="Price for the dataset.")
# Set credits per prediction to your white box or black box models.
parser.add_argument('--cpp',
action='store',
type=float,
default=defaults.get('cpp', 0.0),
help=("The number of credits that other users will"
" consume to make a prediction with your"
" model."))
# Shows progress information when uploading a file.
parser.add_argument('--progress-bar',
action='store_true',
dest='progress_bar',
default=defaults.get('progress_bar', False),
help="Show progress details when creating a source.")
# Does not create a dataset.
parser.add_argument('--no-dataset',
action='store_true',
dest='no_dataset',
default=defaults.get('no_dataset', False),
help="Do not create a dataset.")
# Does not create a model just a dataset.
parser.add_argument('--no-model',
action='store_true',
dest='no_model',
default=defaults.get('no_model', False),
help="Do not create a model.")
# Changes to delete mode.
parser.add_argument('--delete',
action='store_true',
help="Delete command.")
# Resources to be deleted.
parser.add_argument('--ids',
action='store',
dest='delete_list',
help=("Select comma-separated list of"
" resources to be deleted."))
# Resources to be deleted are taken from file.
parser.add_argument('--from-file',
action='store',
dest='delete_file',
default=defaults.get('from_file', None),
help=("Path to a file containing resources ids."
" One resource per line"
" (e.g., model/50a206a8035d0706dc000376)."))
# Sources selected by tag to be deleted.
parser.add_argument('--source-tag',
dest='source_tag',
default=defaults.get('source_tag', None),
help=("Select sources tagged with tag to"
" be deleted"))
# Datasets selected by tag to be deleted.
parser.add_argument('--dataset-tag',
dest='dataset_tag',
default=defaults.get('dataset_tag', None),
help=("Select datasets tagged with tag to"
" be deleted"))
# Predictions selected by tag to be deleted.
parser.add_argument('--prediction-tag',
dest='prediction_tag',
default=defaults.get('prediction_tag', None),
help=("Select prediction tagged with tag to"
" be deleted."))
# Evaluations selected by tag to be deleted.
parser.add_argument('--evaluation-tag',
dest='evaluation_tag',
default=defaults.get('evaluation_tag', None),
help=("Select evaluation tagged with tag to"
" be deleted."))
# Ensembles selected by tag to be deleted.
parser.add_argument('--ensemble-tag',
dest='ensemble_tag',
default=defaults.get('ensemble_tag', None),
help=("Select ensemble tagged with tag to"
" be deleted."))
# Batch predictions selected by tag to be deleted.
parser.add_argument('--batch-prediction-tag',
dest='batch_prediction_tag',
default=defaults.get('batch_prediction_tag', None),
help=("Select batch prediction tagged with tag to"
" be deleted."))
# Resources selected by tag to be deleted.
parser.add_argument('--all-tag',
dest='all_tag',
default=defaults.get('all_tag', None),
help=("Select resources tagged with tag to"
" be deleted."))
# Locale settings.
parser.add_argument('--locale',
action='store',
dest='user_locale',
default=defaults.get('locale', None),
help="Chosen locale code string.")
# Prediction directories to be combined.
parser.add_argument('--combine-votes',
action='store',
dest='votes_dirs',
default=defaults.get('combine_votes', None),
help=("Comma separated list of"
" directories that contain models' votes"
" for the same test set."))
# Method to combine votes in multiple models predictions
parser.add_argument('--method',
action='store',
dest='method',
default=defaults.get('method', plurality),
choices=["plurality", "confidence weighted",
"probability weighted", "threshold",
"combined"],
help="Method to combine votes from ensemble"
" predictions. Allowed methods: plurality"
", \"confidence weighted\", "
" \"probability weighted\", threshold. Also"
" \"combined\" for datasets with subsets of"
" categories")
# Evaluate a model
parser.add_argument('--evaluate',
action='store_true',
help="Evaluate command.")
# The path to a file containing the mapping of fields' ids from
# the test dataset fields to the model fields.
parser.add_argument('--fields-map',
action='store',
dest='fields_map',
default=defaults.get('fields_map', None),
help=("Path to a csv file describing fields mapping. "
"One definition per line (e.g., 00000,"
"00000a)."))
# Set the part of training data to be held out for cross-validation
parser.add_argument('--cross-validation-rate',
action='store',
dest='cross_validation_rate',
type=float,
default=defaults.get('cross_validation_rate', 0.0),
help=("Part of training data to be held out for "
"cross-validation."))
# Number of evaluations used in cross-validation
parser.add_argument('--number-of-evaluations',
action='store',
dest='number_of_evaluations',
type=int,
default=defaults.get('number_of_evaluations', 0),
help=("Number of evaluations used for"
" cross-validation."))
# Set the part of training data to be held out for testing
parser.add_argument('--test-split',
action='store',
dest='test_split',
type=float,
default=defaults.get('test_split', 0.0),
help=("Part of training data to be held out for "
"testing (e.g. --test-split 0.2)."))
# If a BigML ensemble is provided, the script will use it to generate
# predictions.
parser.add_argument('--ensemble',
action='store',
dest='ensemble',
default=defaults.get('ensemble', None),
help="BigML ensemble Id.")
# If a BigML ensemble is created, creation will use this task-level
# parallelism
parser.add_argument('--tlp',
action='store',
dest='tlp',
default=defaults.get('tlp', 1),
type=int,
help=("BigML ensemble's creation task-level"
" parallelism."))
# Prediction log format: `short` will only log predictions, `long` will
# log also confidence information
parser.add_argument('--prediction-info',
action='store',
dest='prediction_info',
default=defaults.get('prediction_info', 'normal'),
choices=["brief", "normal", "full", "full data"],
help=("Prediction log format: 'brief' will only "
"log predictions, 'normal' will write confidence"
" too, 'full' will write in a row the"
" input data that generates the prediction"
" followed by the latter."))
# Test set field separator. Defaults to the locale csv
# separator.
parser.add_argument('--test-separator',
action='store',
dest='test_separator',
default=defaults.get('test_separator', None),
help="Test set field separator.")
# Multi-label. The objective field has multiple labels.
parser.add_argument('--multi-label',
action='store_true',
dest='multi_label',
default=defaults.get('multi_label', False),
help=("The objective field has multiple labels that"
" should be treated independently."))
# Multi-label labels. If set, only the given labels are expanded
parser.add_argument('--labels',
action='store',
dest='labels',
default=defaults.get('labels', None),
help=("Comma-separated list of the labels"
" to be expanded from a multi-label field."))
# Multi-label label separator. Separator used when splitting labels in the
# objective field.
parser.add_argument('--label-separator',
action='store',
dest='label_separator',
default=defaults.get('label_separator', None),
help=("Separator used when splitting labels in the"
" objective field."))
# Training set field separator. Defaults to the locale csv
# separator.
parser.add_argument('--training-separator',
action='store',
dest='training_separator',
default=defaults.get('training_separator', None),
help=("Training set field separator."))
# Prediction header. If set, headers are added to the prediction file.
parser.add_argument('--prediction-header',
action='store_true',
dest='prediction_header',
default=defaults.get('prediction_header', False),
help="Headers are added to the prediction file.")
# Prediction fields. A comma-separated list of the fields that should
# be included in the prediction file.
parser.add_argument('--prediction-fields',
action='store',
dest='prediction_fields',
default=defaults.get('prediction_fields', None),
help="Fields added to the prediction file.")
# Seed. The value used in dataset's splits as seed
parser.add_argument('--seed',
action='store',
dest='seed',
default=defaults.get('seed', None),
help="Value used as seed in dataset splits.")
# Max number of ensembles to create in parallel.
parser.add_argument('--max-parallel-ensembles',
action='store',
dest='max_parallel_ensembles',
default=defaults.get('max_parallel_ensembles', 1),
type=int,
help="Max number of ensembles to create in parallel.")
# The path to a file containing ensemble ids.
parser.add_argument('--ensembles',
action='store',
dest='ensembles',
default=defaults.get('ensembles', None),
help=("Path to a file containing ensemble/ids. One "
"ensemble per line (e.g., "
"ensemble/50a206a8035d0706dc000376)."))
# Threshold. Minimum necessary number of votes to issue a prediction.
parser.add_argument('--threshold',
action='store',
dest='threshold',
default=defaults.get('threshold', 1),
type=int,
help=("Minimum number of votes to issue a prediction"
" for the threshold combiner."))
# Class. Label for the category used in threshold voting predictions.
parser.add_argument('--class',
action='store',
dest='threshold_class',
default=defaults.get('threshold_class', None),
help="Category used in threshold combiner method.")
# Max number of categories to be included in a model
parser.add_argument('--max-categories',
action='store',
dest='max_categories',
default=defaults.get('max_categories', 0),
type=int,
help=("Max number of categories to be included in"
" a model."))
# The path to a file containing attributes if you want to alter BigML's
# default field attributes or the ones provided by the test file header.
parser.add_argument('--test-field-attributes',
action='store',
dest='test_field_attributes',
default=defaults.get('test_field_attributes', None),
help=("Path to a csv file describing field attributes."
" One definition per line"
" (e.g., 0,'Last Name')."))
# The path to a file containing types if you want to alter BigML's
# type auto-detect.
parser.add_argument('--test-types',
action='store',
dest='test_types',
default=defaults.get('test_types', None),
help=("Path to a file describing field types. One"
" definition per line (e.g., 0, 'numeric')."))
# If a BigML test source is provided, the script won't create a new one
parser.add_argument('--test-source',
action='store',
dest='test_source',
default=defaults.get('test_source', None),
help="BigML test source Id.")
# If a BigML test dataset is provided, the script won't create a new one
parser.add_argument('--test-dataset',
action='store',
dest='test_dataset',
default=defaults.get('test_dataset', None),
help="BigML test dataset Id.")
# No batch predictions. Remote predictions are created individually.
parser.add_argument('--no-batch',
action='store_true',
dest='no_batch',
default=defaults.get('no_batch', False),
help="Create remote predictions individually.")
# The path to a file containing dataset attributes.
parser.add_argument('--dataset-attributes',
action='store',
dest='dataset_attributes',
default=defaults.get('dataset_attributes', None),
help=("Path to a json file describing dataset"
" attributes."))
# Name of the directory where session files will be stored. If --output
# is set, this setting will be overriden by it.
parser.add_argument('--output-dir',
action='store',
dest='output_dir',
default=defaults.get('output_dir', None),
help=("Directory where session files will be stored."
" --output file path will override it if both"
" are set."))
# The path to a file containing dataset ids.
parser.add_argument('--test-datasets',
action='store',
dest='test_datasets',
default=defaults.get('test_datasets', None),
help=("Path to a file containing dataset/ids. Just"
" one dataset per line"
" (e.g., dataset/50a20697035d0706da0004a4)."))
# Evaluations flag: excluding one dataset from the datasets list to test
parser.add_argument('--dataset-off',
action='store_true',
dest='dataset_off',
default=defaults.get('dataset_off', False),
help=("Excluding one dataset at a time from the"
" datasets list to test."))
# Evaluations flag: excluding one dataset from the datasets list to test
parser.add_argument('--version',
action='version',
version=version_text)
# Path to the file containing fields generators for the new dataset.
# Used when generating a dataset from another by adding new fields
# combining or setting its contents.
parser.add_argument('--new-fields',
action='store',
dest='new_fields',
default=defaults.get('new_fields', None),
help=("Path to the file containing fields generators."
" Used to create a new dataset from an existing"
" one by adding new fields combining or"
" setting its contents."))
# The path to a file containing model attributes.
parser.add_argument('--model-attributes',
action='store',
dest='model_attributes',
default=defaults.get('model_attributes', None),
help=("Path to a json file describing model"
" attributes."))
# Node threshold. Maximum number of nodes in the tree.
parser.add_argument('--node-threshold',
action='store',
dest='node_threshold',
default=defaults.get('node_threshold', 0),
type=int,
help="Maximum number of nodes in the model.")
# Multi-label fields. Comma-separated list of fields that should be treated
# as being multi-label fields. Either its name or column number.
parser.add_argument('--multi-label-fields',
action='store',
dest='multi_label_fields',
default=defaults.get('multi_label_fields', None),
help=("Comma-separated list of the fields"
" to be expanded as being multi-label. Name"
" or column number."))
# The path to a file containing ensemble attributes.
parser.add_argument('--ensemble-attributes',
action='store',
dest='ensemble_attributes',
default=defaults.get('ensemble_attributes', None),
help=("Path to a json file describing ensemble"
" attributes."))
# The path to a file containing source attributes.
parser.add_argument('--source-attributes',
action='store',
dest='source_attributes',
default=defaults.get('source_attributes', None),
help=("Path to a json file describing source"
" attributes."))
# The path to a file containing evaluation attributes.
parser.add_argument('--evaluation-attributes',
action='store',
dest='evaluation_attributes',
default=defaults.get('evaluation_attributes', None),
help=("Path to a json file describing evaluation"
" attributes."))
# The path to a file containing batch prediction attributes.
parser.add_argument('--batch-prediction-attributes',
action='store',
dest='batch_prediction_attributes',
default=defaults.get('batch_prediction_attributes',
None),
help=("Path to a json file describing batch prediction"
" attributes."))
# Balance. Automatically balance all the classes evenly.
parser.add_argument('--balance',
action='store_true',
dest='balance',
default=defaults.get('balance', False),
help=("Automatically balance all the classes"
" evenly."))
# Weight-field. Use the contents of the given field as weights.
parser.add_argument('--weight-field',
action='store',
dest='weight_field',
default=defaults.get('weight_field', None),
help=("Sets the name (or column) of the field"
" that contains the weights for the instances."))
# Objective-weights. Path a to a CSV file of class, weight pairs.
parser.add_argument('--objective-weights',
action='store',
dest='objective_weights',
default=defaults.get('objective_weights', None),
help="Path to a CSV file of class, weight pairs.")
# Label-aggregates. Comma-separated list of aggregation functions
# for the multi-label fields.
parser.add_argument('--label-aggregates',
action='store',
dest='label_aggregates',
default=defaults.get('label_aggregates', None),
help=("Comma-separated list of aggregation functions "
"for the multi-label field labels."
" Allowed aggregates: count, first and last"))
# Strategy used in predictions when a missing value is found for the
# field used to split the node.
parser.add_argument('--missing-strategy',
action='store',
dest='missing_strategy',
default=defaults.get('missing_strategy', last),
choices=["last", "proportional"],
help="Strategy used when the field used in the rules"
" to next nodes is missing in the input data."
" Allowed values: last or proportional")
# Condition to select resources for deletion: olther than.
parser.add_argument('--older-than',
action='store',
dest='older_than',
default=defaults.get('older_than', None),
help=("Upper limit to select the resources older than"
" the given number of days, date, or resource."))
# Condition to select resources for deletion: newer than.
parser.add_argument('--newer-than',
action='store',
dest='newer_than',
default=defaults.get('newer_than', None),
help=("Lower limit to select the resources newer than"
" the given number of days, date, or resource."))
# Multi-dataset. Generate new dataset from a list of existing datasets.
parser.add_argument('--multi-dataset',
action='store_true',
dest='multi_dataset',
default=defaults.get('multi_dataset', False),
help=("Generate a new dataset by adding existing"
" datasets."))
# The path to a file containing multi-dataset attributes.
parser.add_argument('--multi-dataset-attributes',
action='store',
dest='multi_dataset_attributes',
default=defaults.get('multi_dataset_attributes', None),
help=("Path to a json file describing multi-dataset"
" attributes."))
# Shared. Shares all shareable resources and uses its shared links in
# reports
parser.add_argument('--shared',
action='store_true',
dest='shared',
default=defaults.get('shared', False),
help=("Share resources and use its shared urls "
" in reports."))
# Report. Additional output report formats
parser.add_argument('--reports',
action='store',
dest='reports',
nargs='*',
default=defaults.get('reports', []),
choices=["gazibit"],
help="Output report formats.")
# Disables reports upload.
parser.add_argument('--no-upload',
action='store_false',
dest='upload',
default=defaults.get('upload', True),
help="Disables upload for reports")
# The following options are only useful to deactivate the corresponding
# oposed default values
#
# Set when the training set file does include a header on the first
# line.
parser.add_argument('--train-header',
action='store_true',
dest='train_header',
default=defaults.get('train_header', True),
help="The train set file has a header.")
# Set when the test set file does include a header on the first
# line.
parser.add_argument('--test-header',
action='store_true',
dest='test_header',
default=defaults.get('test_header', True),
help="The test set file has a header.")
# Use it to compute predictions locally.
parser.add_argument('--local',
action='store_false',
dest='remote',
default=defaults.get('remote', False),
help="Compute predictions locally")
# Deactivate replacement to use when using bagging.
parser.add_argument('--no-replacement',
action='store_false',
dest='replacement',
default=defaults.get('replacement', False),
help="Don't use replacement when sampling.")
# Doesn't randomize feature selection at each split.
parser.add_argument('--no-randomize',
action='store_false',
dest='randomize',
default=defaults.get('randomize', False),
help=("Doesn't randomize feature selection at each"
" split."))
# Set default tagging of resources.
parser.add_argument('--no-no-tag',
action='store_true',
dest='no_tag',
default=defaults.get('no_tag', True),
help="No tag resources with default BigMLer tags.")
# Doesn't make dataset public.
parser.add_argument('--no-public-dataset',
action='store_false',
dest='public_dataset',
default=defaults.get('public_dataset', False),
help="Doesn't make generated dataset public.")
# Doesn't make model a public black-box model.
parser.add_argument('--no-black-box',
action='store_false',
dest='black_box',
default=defaults.get('black_box', False),
help="Doesn't make generated model black-box.")
# Doesn't make model a public white-box model.
parser.add_argument('--no-white-box',
action='store_false',
dest='white_box',
default=defaults.get('white_box', False),
help="Doesn't make generated model white-box.")
# Hides progress information when uploading a file.
parser.add_argument('--no-progress-bar',
action='store_false',
dest='progress_bar',
default=defaults.get('progress_bar', False),
help="Show progress details when creating a source.")
# Create a dataset.
parser.add_argument('--no-no-dataset',
action='store_false',
dest='no_dataset',
default=defaults.get('no_dataset', False),
help="Create a dataset.")
# Create a model just a dataset.
parser.add_argument('--no-no-model',
action='store_false',
dest='no_model',
default=defaults.get('no_model', False),
help="Create a model.")
# Don't clear global bigmler log files
parser.add_argument('--no-clear-logs',
action='store_false',
dest='clear_logs',
default=defaults.get('clear_logs', False),
help="Don't clear global bigmler log files.")
# Don't store the retrieved resources in the output directory
parser.add_argument('--no-store',
action='store_false',
dest='store',
default=defaults.get('store', False),
help=("Don't store the retrieved resources in the"
" output directory."))
# Multi-label. The objective field hasn't multiple labels.
parser.add_argument('--no-multi-label',
action='store_false',
dest='multi_label',
default=defaults.get('multi_label', False),
help=("The objective field has not multiple labels."))
# Prediction-header.
parser.add_argument('--no-prediction-header',
action='store_false',
dest='prediction_header',
default=defaults.get('prediction_header', False),
help="Headers are not added to the prediction file.")
# Batch predictions. Remote predictions are created in batch mode.
parser.add_argument('--batch',
action='store_false',
dest='no_batch',
default=defaults.get('no_batch', False),
help="Create remote predictions in batch.")
# Balance. Automatically balance all the classes evenly.
parser.add_argument('--no-balance',
action='store_false',
dest='balance',
default=defaults.get('balance', False),
help=("Automatically balance all the classes"
" evenly."))
# Multi-dataset. Generating a new dataset from a list of existing datasets.
parser.add_argument('--no-multi-dataset',
action='store_false',
dest='multi_dataset',
default=defaults.get('multi_dataset', False),
help="Do not generate a new dataset.")
# Shared. Shares all shareable resources and uses its shared links in
# reports
parser.add_argument('--unshared',
action='store_false',
dest='shared',
default=defaults.get('shared', False),
help=("Share resources and use its shared urls "
" in reports."))
# Enables reports upload.
parser.add_argument('--upload',
action='store_true',
dest='upload',
default=defaults.get('upload', True),
help="Enables upload for reports")
# Dataset-off. Turning off the dataset-off flag.
parser.add_argument('--no-dataset-off',
action='store_false',
dest='dataset_off',
default=defaults.get('dataset_off', False),
help="Turning off the dataset-off flag.")
# Subcommands
analyze.subparser_options(
subparsers, defaults=general_defaults['BigMLer analyze'],
constants=constants)
return main_parser, common_options_set
def subparser_options(subparsers, defaults={}, constants={}):
"""Adding arguments for the analyze subcommand
"""
parser_analysis = subparsers.add_parser('analyze')
common_options(parser_analysis, defaults=defaults, constants=constants);
# K-fold cross-validation flag
parser_analysis.add_argument('--cross-validation',
action='store_true',
dest='cv',
default=defaults.get('cv', False),
help="K-fold cross-validation flag.")
# K-fold cross-validation index
parser_analysis.add_argument('--k-folds',
action='store',
dest='k_folds',
default=defaults.get('k_folds', None),
type=int,
help="K-fold cross-validation index.")
# Name of the directory where session files will be stored. If --output
# is set, this setting will be overriden by it.
parser_analysis.add_argument('--output-dir',
action='store',
dest='output_dir',
default=defaults.get('output_dir', None),
help=("Directory where session files will be"
" stored. --output file path will "
" override it if both"
" are set."))
# If a BigML dataset is provided, the script won't create a new one
parser_analysis.add_argument('--dataset',
action='store',
dest='dataset',
default=defaults.get('dataset', None),
help="BigML dataset Id.")
# Staleness to stop --features analysis
parser_analysis.add_argument('--staleness',
action='store',
dest='staleness',
type=float,
default=defaults.get('staleness', None),
help=("Limit staleness when using --features"
" analysis."))
# Penalty per feature in --features analysis
parser_analysis.add_argument('--penalty',
action='store',
dest='penalty',
type=float,
default=defaults.get('penalty', None),
help=("Penalty for each feature in --features"
" analysis."))
# Features: Set on the feature exclusion analysis
parser_analysis.add_argument('--features',
action='store_true',
dest='features',
default=defaults.get('features', False),
help="Features analysis.")
# Mazimize: evaluation measure to be maximized
parser_analysis.add_argument('--maximize',
action='store',
dest='maximize',
choices=MAXIMIZE_OPTIONS,
default=defaults.get('maximize', ACCURACY),
help="Evaluation measure to be maximized.")
# Min_nodes: Minimum number of nodes to start the node threshold analysis
parser_analysis.add_argument('--min-nodes',
action='store',
dest='min_nodes',
type=int,
default=defaults.get('min_nodes', None),
help=("Minimum number of nodes to start the"
"node threshold analysis."))
# Max_nodes: Maximum number of nodes to end the node threshold analysis
parser_analysis.add_argument('--max-nodes',
action='store',
dest='max_nodes',
type=int,
default=defaults.get('max_nodes',
BIGML_MAX_NODES),
help=("Maximum number of nodes to end the"
"node threshold analysis."))
# Nodes: Set on node threshold analysis.
parser_analysis.add_argument('--nodes',
action='store_true',
dest='nodes',
default=defaults.get('nodes', False),
help="Node threshold analysis.")
# Nodes step: Step to increase the nodes threshold number
parser_analysis.add_argument('--nodes-step',
action='store',
dest='nodes_step',
type=int,
default=defaults.get('nodes_step', None),
help=("Step to increase the nodes"
" threshold number in nodes "
" threshold analysis. If not set,"
" an increase of 100 is used"))
# Exclude some features from the features analyze
parser_analysis.add_argument('--exclude-features',
action='store',
dest='exclude_features',
default=defaults.get('exclude_features',
None),
help=("List of comma-separated field names to"
" be excluded from the features"
" analysis."))
