BigML Python Bindings
=====================
`BigML <https://bigml.com>`_ makes machine learning easy by taking care
of the details required to add data-driven decisions and predictive
power to your company. Unlike other machine learning services, BigML
creates
`beautiful predictive models <https://bigml.com/gallery/models>`_ that
can be easily understood and interacted with.
These BigML Python bindings allow you to interact with
`BigML.io <https://bigml.io/>`_, the API
for BigML. You can use it to easily create, retrieve, list, update, and
delete BigML resources (i.e., sources, datasets, models and,
predictions). For additional information, see
the `full documentation for the Python
bindings on Read the Docs <http://bigml.readthedocs.org>`_.
This module is licensed under the `Apache License, Version
2.0 <http://www.apache.org/licenses/LICENSE-2.0.html>`_.
Support
-------
Please report problems and bugs to our `BigML.io issue
tracker <https://github.com/bigmlcom/io/issues>`_.
Discussions about the different bindings take place in the general
`BigML mailing list <http://groups.google.com/group/bigml>`_. Or join us
in our `Campfire chatroom <https://bigmlinc.campfirenow.com/f20a0>`_.
Requirements
------------
Only ``Python 3`` versions are currently supported by these bindings.
Support for Python 2.7.X ended in version ``4.32.3``.
The basic third-party dependencies are the
`requests <https://github.com/kennethreitz/requests>`_,
`unidecode <http://pypi.python.org/pypi/Unidecode/#downloads>`_,
`requests-toolbelt <https://pypi.python.org/pypi/requests-toolbelt>`_,
`bigml-chronos <https://pypi.org/project/bigml-chronos>`_,
`msgpack <https://pypi.org/project/msgpack>`_,
`numpy <http://www.numpy.org/>`_ and
`scipy <http://www.scipy.org/>`_ libraries. These
libraries are automatically installed during the basic setup.
Support for Google App Engine has been added as of version 3.0.0,
using the `urlfetch` package instead of `requests`.
The bindings will also use ``simplejson`` if you happen to have it
installed, but that is optional: we fall back to Python's built-in JSON
libraries is ``simplejson`` is not found.
The bindings provide support to use the ``BigML`` platform to create, update,
get and delete resources, but also to produce local predictions using the
models created in ``BigML``. Most of them will be actionable with the basic
installation, but some additional dependencies are needed to use local
``Topic Models`` and Image Processing models. Please, refer to the
`Installation <#installation>`_ section for details.
OS Requirements
~~~~~~~~~~~~~~~
The basic installation of the bindings is compatible and can be used
on Linux and Windows based Operating Systems.
However, the extra options that allow working with
image processing models (``[images]`` and ``[full]``) are only supported
and tested on Linux-based Operating Systems.
For image models, Windows OS is not recommended and cannot be supported out of
the box, because the specific compiler versions or dlls required are
unavailable in general.
Installation
------------
To install the basic latest stable release with
`pip <http://www.pip-installer.org/>`_, please use:
.. code-block:: bash
$ pip install bigml
Support for local Topic Distributions (Topic Models' predictions)
and local predictions for datasets that include Images will only be
available as extras, because the libraries used for that are not
usually available in all Operative Systems. If you need to support those,
please check the `Installation Extras <#installation-extras>`_ section.
Installation Extras
-------------------
Local Topic Distributions support can be installed using:
.. code-block:: bash
pip install bigml[topics]
Images local predictions support can be installed using:
.. code-block:: bash
pip install bigml[images]
The full set of features can be installed using:
.. code-block:: bash
pip install bigml[full]
WARNING: Mind that installing these extras can require some extra work, as
explained in the `Requirements <#requirements>`_ section.
You can also install the development version of the bindings directly
from the Git repository
.. code-block:: bash
$ pip install -e git://github.com/bigmlcom/python.git#egg=bigml_python
Running the Tests
-----------------
The tests will be run using `pytest <https://docs.pytest.org/en/7.2.x/>`_.
You'll need to set up your authentication
via environment variables, as explained
in the authentication section. Also some of the tests need other environment
variables like ``BIGML_ORGANIZATION`` to test calls when used by Organization
members and ``BIGML_EXTERNAL_CONN_HOST``, ``BIGML_EXTERNAL_CONN_PORT``,
``BIGML_EXTERNAL_CONN_DB``, ``BIGML_EXTERNAL_CONN_USER``,
``BIGML_EXTERNAL_CONN_PWD`` and ``BIGML_EXTERNAL_CONN_SOURCE``
in order to test external data connectors.
With that in place, you can run the test suite simply by issuing
.. code-block:: bash
$ pytest
Additionally, `Tox <http://tox.testrun.org/>`_ can be used to
automatically run the test suite in virtual environments for all
supported Python versions. To install Tox:
.. code-block:: bash
$ pip install tox
Then run the tests from the top-level project directory:
.. code-block:: bash
$ tox
Importing the module
--------------------
To import the module:
.. code-block:: python
import bigml.api
Alternatively you can just import the BigML class:
.. code-block:: python
from bigml.api import BigML
Authentication
--------------
All the requests to BigML.io must be authenticated using your username
and `API key <https://bigml.com/account/apikey>`_ and are always
transmitted over HTTPS.
This module will look for your username and API key in the environment
variables ``BIGML_USERNAME`` and ``BIGML_API_KEY`` respectively.
Unix and MacOS
--------------
You can
add the following lines to your ``.bashrc`` or ``.bash_profile`` to set
those variables automatically when you log in:
.. code-block:: bash
export BIGML_USERNAME=myusername
export BIGML_API_KEY=ae579e7e53fb9abd646a6ff8aa99d4afe83ac291
refer to the next chapters to know how to do that in other operating systems.
With that environment set up, connecting to BigML is a breeze:
.. code-block:: python
from bigml.api import BigML
api = BigML()
Otherwise, you can initialize directly when instantiating the BigML
class as follows:
.. code-block:: python
api = BigML('myusername', 'ae579e7e53fb9abd646a6ff8aa99d4afe83ac291')
These credentials will allow you to manage any resource in your user
environment.
In BigML a user can also work for an ``organization``.
In this case, the organization administrator should previously assign
permissions for the user to access one or several particular projects
in the organization.
Once permissions are granted, the user can work with resources in a project
according to his permission level by creating a special constructor for
each project. The connection constructor in this case
should include the ``project ID``:
.. code-block:: python
api = BigML('myusername', 'ae579e7e53fb9abd646a6ff8aa99d4afe83ac291',
project='project/53739b98d994972da7001d4a')
If the project used in a connection object
does not belong to an existing organization but is one of the
projects under the user's account, all the resources
created or updated with that connection will also be assigned to the
specified project.
When the resource to be managed is a ``project`` itself, the connection
needs to include the corresponding``organization ID``:
.. code-block:: python
api = BigML('myusername', 'ae579e7e53fb9abd646a6ff8aa99d4afe83ac291',
organization='organization/53739b98d994972da7025d4a')
Authentication on Windows
-------------------------
The credentials should be permanently stored in your system using
.. code-block:: bash
setx BIGML_USERNAME myusername
setx BIGML_API_KEY ae579e7e53fb9abd646a6ff8aa99d4afe83ac291
Note that ``setx`` will not change the environment variables of your actual
console, so you will need to open a new one to start using them.
Authentication on Jupyter Notebook
----------------------------------
You can set the environment variables using the ``%env`` command in your
cells:
.. code-block:: bash
%env BIGML_USERNAME=myusername
%env BIGML_API_KEY=ae579e7e53fb9abd646a6ff8aa99d4afe83ac291
Alternative domains
-------------------
The main public domain for the API service is ``bigml.io``, but there are some
alternative domains, either for Virtual Private Cloud setups or
the australian subdomain (``au.bigml.io``). You can change the remote
server domain
to the VPC particular one by either setting the ``BIGML_DOMAIN`` environment
variable to your VPC subdomain:
.. code-block:: bash
export BIGML_DOMAIN=my_VPC.bigml.io
or setting it when instantiating your connection:
.. code-block:: python
api = BigML(domain="my_VPC.bigml.io")
The corresponding SSL REST calls will be directed to your private domain
henceforth.
You can also set up your connection to use a particular PredictServer
only for predictions. In order to do so, you'll need to specify a ``Domain``
object, where you can set up the general domain name as well as the
particular prediction domain name.
.. code-block:: python
from bigml.domain import Domain
from bigml.api import BigML
domain_info = Domain(prediction_domain="my_prediction_server.bigml.com",
prediction_protocol="http")
api = BigML(domain=domain_info)
Finally, you can combine all the options and change both the general domain
server, and the prediction domain server.
.. code-block:: python
from bigml.domain import Domain
from bigml.api import BigML
domain_info = Domain(domain="my_VPC.bigml.io",
prediction_domain="my_prediction_server.bigml.com",
prediction_protocol="https")
api = BigML(domain=domain_info)
Some arguments for the Domain constructor are more unsual, but they can also
be used to set your special service endpoints:
- protocol (string) Protocol for the service
(when different from HTTPS)
- verify (boolean) Sets on/off the SSL verification
- prediction_verify (boolean) Sets on/off the SSL verification
for the prediction server (when different from the general
SSL verification)
**Note** that the previously existing ``dev_mode`` flag:
.. code-block:: python
api = BigML(dev_mode=True)
that caused the connection to work with the Sandbox ``Development Environment``
has been **deprecated** because this environment does not longer exist.
The existing resources that were previously
created in this environment have been moved
to a special project in the now unique ``Production Environment``, so this
flag is no longer needed to work with them.
Quick Start
-----------
Imagine that you want to use `this csv
file <https://static.bigml.com/csv/iris.csv>`_ containing the `Iris
flower dataset <http://en.wikipedia.org/wiki/Iris_flower_data_set>`_ to
predict the species of a flower whose ``petal length`` is ``2.45`` and
whose ``petal width`` is ``1.75``. A preview of the dataset is shown
below. It has 4 numeric fields: ``sepal length``, ``sepal width``,
``petal length``, ``petal width`` and a categorical field: ``species``.
By default, BigML considers the last field in the dataset as the
objective field (i.e., the field that you want to generate predictions
for).
::
sepal length,sepal width,petal length,petal width,species
5.1,3.5,1.4,0.2,Iris-setosa
4.9,3.0,1.4,0.2,Iris-setosa
4.7,3.2,1.3,0.2,Iris-setosa
...
5.8,2.7,3.9,1.2,Iris-versicolor
6.0,2.7,5.1,1.6,Iris-versicolor
5.4,3.0,4.5,1.5,Iris-versicolor
...
6.8,3.0,5.5,2.1,Iris-virginica
5.7,2.5,5.0,2.0,Iris-virginica
5.8,2.8,5.1,2.4,Iris-virginica
You can easily generate a prediction following these steps:
.. code-block:: python
from bigml.api import BigML
api = BigML()
source = api.create_source('./data/iris.csv')
dataset = api.create_dataset(source)
model = api.create_model(dataset)
prediction = api.create_prediction(model, \
{"petal width": 1.75, "petal length": 2.45})
You can then print the prediction using the ``pprint`` method:
.. code-block:: python
>>> api.pprint(prediction)
species for {"petal width": 1.75, "petal length": 2.45} is Iris-setosa
Certainly, any of the resources created in BigML can be configured using
several arguments described in the `API documentation <https://bigml.com/api>`_.
Any of these configuration arguments can be added to the ``create`` method
as a dictionary in the last optional argument of the calls:
.. code-block:: python
from bigml.api import BigML
api = BigML()
source_args = {"name": "my source",
"source_parser": {"missing_tokens": ["NULL"]}}
source = api.create_source('./data/iris.csv', source_args)
dataset_args = {"name": "my dataset"}
dataset = api.create_dataset(source, dataset_args)
model_args = {"objective_field": "species"}
model = api.create_model(dataset, model_args)
prediction_args = {"name": "my prediction"}
prediction = api.create_prediction(model, \
{"petal width": 1.75, "petal length": 2.45},
prediction_args)
The ``iris`` dataset has a small number of instances, and usually will be
instantly created, so the ``api.create_`` calls will probably return the
finished resources outright. As BigML's API is asynchronous,
in general you will need to ensure
that objects are finished before using them by using ``api.ok``.
.. code-block:: python
from bigml.api import BigML
api = BigML()
source = api.create_source('./data/iris.csv')
api.ok(source)
dataset = api.create_dataset(source)
api.ok(dataset)
model = api.create_model(dataset)
api.ok(model)
prediction = api.create_prediction(model, \
{"petal width": 1.75, "petal length": 2.45})
Note that the prediction
call is not followed by the ``api.ok`` method. Predictions are so quick to be
generated that, unlike the
rest of resouces, will be generated synchronously as a finished object.
The example assumes that your objective field (the one you want to predict)
is the last field in the dataset. If that's not he case, you can explicitly
set the name of this field in the creation call using the ``objective_field``
argument:
.. code-block:: python
from bigml.api import BigML
api = BigML()
source = api.create_source('./data/iris.csv')
api.ok(source)
dataset = api.create_dataset(source)
api.ok(dataset)
model = api.create_model(dataset, {"objective_field": "species"})
api.ok(model)
prediction = api.create_prediction(model, \
{'sepal length': 5, 'sepal width': 2.5})
You can also generate an evaluation for the model by using:
.. code-block:: python
test_source = api.create_source('./data/test_iris.csv')
api.ok(test_source)
test_dataset = api.create_dataset(test_source)
api.ok(test_dataset)
evaluation = api.create_evaluation(model, test_dataset)
api.ok(evaluation)
If you set the ``storage`` argument in the ``api`` instantiation:
.. code-block:: python
api = BigML(storage='./storage')
all the generated, updated or retrieved resources will be automatically
saved to the chosen directory.
Alternatively, you can use the ``export`` method to explicitly
download the JSON information
that describes any of your resources in BigML to a particular file:
.. code-block:: python
api.export('model/5acea49a08b07e14b9001068',
filename="my_dir/my_model.json")
This example downloads the JSON for the model and stores it in
the ``my_dir/my_model.json`` file.
In the case of models that can be represented in a `PMML` syntax, the
export method can be used to produce the corresponding `PMML` file.
.. code-block:: python
api.export('model/5acea49a08b07e14b9001068',
filename="my_dir/my_model.pmml",
pmml=True)
You can also retrieve the last resource with some previously given tag:
.. code-block:: python
api.export_last("foo",
resource_type="ensemble",
filename="my_dir/my_ensemble.json")
which selects the last ensemble that has a ``foo`` tag. This mechanism can
be specially useful when retrieving retrained models that have been created
with a shared unique keyword as tag.
For a descriptive overview of the steps that you will usually need to
follow to model
your data and obtain predictions, please see the `basic Workflow sketch
<api_sketch.html>`_
document. You can also check other simple examples in the following documents:
- `model 101 <101_model.html>`_
- `logistic regression 101 <101_logistic_regression.html>`_
- `linear regression 101 <101_linear_regression.html>`_
- `ensemble 101 <101_ensemble.html>`_
- `cluster 101 <101_cluster>`_
- `anomaly detector 101 <101_anomaly.html>`_
- `association 101 <101_association.html>`_
- `topic model 101 <101_topic_model.html>`_
- `deepnet 101 <101_deepnet.html>`_
- `time series 101 <101_ts.html>`_
- `fusion 101 <101_fusion.html>`_
- `scripting 101 <101_scripting.html>`_
Additional Information
----------------------
We've just barely scratched the surface. For additional information, see
the `full documentation for the Python
bindings on Read the Docs <http://bigml.readthedocs.org>`_.
Alternatively, the same documentation can be built from a local checkout
of the source by installing `Sphinx <http://sphinx.pocoo.org>`_
(``$ pip install sphinx``) and then running
.. code-block:: bash
$ cd docs
$ make html
Then launch ``docs/_build/html/index.html`` in your browser.
How to Contribute
-----------------
Please follow the next steps:
1. Fork the project on github.com.
2. Create a new branch.
3. Commit changes to the new branch.
4. Send a `pull request <https://github.com/bigmlcom/python/pulls>`_.
For details on the underlying API, see the
`BigML API documentation <https://bigml.com/developers>`_.
.. :changelog:
History
-------
9.7.1 (2023-12-08)
------------------
- Fixing readthedocs configuration.
9.7.0 (2023-12-06)
------------------
- Changing query string separator in internall API calls.
9.6.2 (2023-08-02)
------------------
- Extending cloning to all available models and WhizzML scripts.
- Fixing shared resources cloning.
9.6.1 (2023-08-01)
------------------
- Adding shared resources cloning.
9.6.0 (2023-07-20)
------------------
- Adding ShapWrapper to enable local Shap values computation with the Shap
library.
- Adding Evaluation object.
- Improving Field class to allow field values encoding as numpy arrays.
9.5.0 (2023-06-16)
------------------
- Extending Local Fusions output to include confidence.
9.4.0 (2023-06-14)
------------------
- Extending LocalModel class to handle Time Series locally.
9.3.0 (2023-06-09)
------------------
- Adding a LocalModel class to handle any type of BigML model locally.
9.2.0 (2023-04-11)
------------------
- Extending all delete methods to allow additional query strings.
9.1.4 (2023-02-10)
------------------
- Providing local deepnet predictions default for Windows OS that cannot
handle images predictions.
9.1.3 (2022-12-22)
------------------
- Changing user's status endpoint retrieval to a lightweight version.
9.1.2 (2022-11-26)
------------------
- Removing all nose dependencies in tests.
9.1.1 (2022-11-18)
------------------
- Removing traces and refactoring Flatline interpreter invocation.
- Migrating tests to pytest.
9.1.0 (2022-11-09)
------------------
- Enabling local models to generate the transformations pipeline used in
training to be able to apply it locally to input data for predictions.
- Refactoring local models to bring the feature extraction capabilities to
the transformations pipeline.
- Rounding local numeric predictions and scores to 5 digits.
- Improving Pipelines and Image Processing documentation.
9.0.1 (2022-11-02)
------------------
- Fixing local batch_prediction method on DataFrames.
- Improving the training examples method in the Fields class to produce
examples of categorical field values following their original distribution.
9.0.0 (2022-10-12)
------------------
- Renaming Pipeline class to BMLPipeline and refactoring to allow scikit-learn
and Pandas pipelines to be also part of the Pipeline transformations.
- Adding several DataTransformers (BMLDataTransformer, SKDataTransformer and
DFDataTransformer) as wrappers for transformation generators.
8.2.3 (2022-10-07)
------------------
- Fixing dump methods for all local resources.
- Refactoring output options in batch_predict methods.
- Adding name and description to local resources.
8.2.2 (2022-09-29)
------------------
- Fixing locale check.
- Documenting the new ``.batch_predict`` method added to local models to
homogenize local batch predictions and accept Pandas' DataFrame as input.
8.2.1 (2022-09-23)
------------------
- Fixing update annotations function when creating annotated images sources.
8.2.0 (2022-09-22)
------------------
- Adding new option to api.ok to report progress via a function callback.
8.1.1 (2022-09-21)
------------------
- Fixing bug in api.ok: response to failing API calls.
8.1.0 (2022-09-16)
------------------
- Upgrading dependencies.
8.0.0 (2022-09-13)
------------------
- Adding new local Pipeline object to encapsulate and run prediction
workflows.
- Changing ``api.ok`` response to return as False when retrieving faulty
resources. Previously, only problems retrieving the API response cause
``api.ok`` returning False.
- Adding FaultyResource exception to inform about that type of failure when
using the ``raise_on_error`` flag in ``api.ok``.
- Fixing bug in local LogisticRegression when predicting with operating points.
7.7.0 (2022-08-30)
~~~~~~~~~~~~~~~~~~
- Adding local Dataset class to handle Flatline transformations.
7.6.0 (2022-08-18)
~~~~~~~~~~~~~~~~~~
- Adding feature expansion to image fields in local predictors.
7.5.1 (2022-08-08)
~~~~~~~~~~~~~~~~~~
- Adding utility to reproduce the backend images preprocess in local
deepnets.
7.5.0 (2022-07-09)
~~~~~~~~~~~~~~~~~~
- Adding the new capabilities the last API release: object detection.
7.4.2 (2022-05-26)
~~~~~~~~~~~~~~~~~~
- Improved webhooks hashing check.
- Adapting connection to allow API version setting, needed for compatibility
with the Predict Server syntax.
7.4.1 (2022-05-16)
~~~~~~~~~~~~~~~~~~
- Adding support for webhooks hashing check.
7.4.0 (2022-05-13)
~~~~~~~~~~~~~~~~~~
- Adding extras to setup, to allow a basic installation and extra dependencies
handling for topics and images.
7.3.0 (2022-04-22)
~~~~~~~~~~~~~~~~~~
- Improving local ensembles to be generated from list of ensemble plus
models info.
- Allowing BigML connection class to retrieve models from storage without
any credentials.
7.2.2 (2022-04-20)
~~~~~~~~~~~~~~~~~~
- Enabling dump from SupervisedModel class.
7.2.1 (2022-04-19)
~~~~~~~~~~~~~~~~~~
- Fixing windows setup.
7.2.0 (2022-04-14)
~~~~~~~~~~~~~~~~~~
- Upgrading sensenet library and giving fallback for local deepnet predictions
on OS not supported by tensorflow.
- Adding support for pandas' DataFrame when creating sources.
7.1.2 (2022-03-17)
~~~~~~~~~~~~~~~~~~
- Fixing local deepnet predictions if no missing_count info is found.
7.1.1 (2022-03-11)
~~~~~~~~~~~~~~~~~~
- Improving local predictions for shared fusions and documenting.
7.1.0 (2022-03-10)
~~~~~~~~~~~~~~~~~~
- Adding the ability to predict locally from shared composed resources, like
Ensembles and Fusions.
7.0.1 (2022-03-04)
~~~~~~~~~~~~~~~~~~
- Fixing local Supervised Model to accept stored models as input.
- Improving tests to minimize resource creation.
- Updating dependencies versions.
7.0.0 (2021-08-03)
~~~~~~~~~~~~~~~~~~
- Adding composite sources to the API REST interface methods.
- Adding predictions for images.
6.3.2 (2021-12-02)
------------------
- Fixing local deepnet prediction method to fit new deepnet structure.
6.3.1 (2021-11-03)
------------------
- Improving local ensembles to handle faulty models with no root.
6.3.0 (2021-10-21)
------------------
- Adding activators to local deepnets prediction functions.
6.2.2 (2021-10-11)
------------------
- Fixing underlying dependency used in datetime fields.
6.2.1 (2021-08-10)
------------------
- Fixing predictions for local models with text fields.
6.2.0 (2021-07-22)
------------------
- Minimizing local model information when text or items fields are used.
6.1.2 (2021-06-16)
------------------
- Changing tests mechanism to github actions.
6.1.1 (2021-06-11)
~~~~~~~~~~~~~~~~~~
- Fixing api `get` method. The query string argument was lost.
- Improving api `get`, `create` and `update` methods to avoid unneeded
iterations.
- Improving docs.
6.1.0 (2021-03-06)
~~~~~~~~~~~~~~~~~~
- Fixing local anomalies when normalize_repeats has been used.
6.0.3 (2021-02-06)
~~~~~~~~~~~~~~~~~~
- Improving local models with default numeric values to automatically fill in
missings in input data.
- Raising exceptions if trying to update a failed resource.
6.0.2 (2020-12-21)
~~~~~~~~~~~~~~~~~~
- Fixing bug in local anomaly filter.
6.0.1 (2020-11-11)
~~~~~~~~~~~~~~~~~~
- Changing local anomaly filter.
- Fixing update and download methods to wait for the resource to be finished.
6.0.0 (2020-09-02)
~~~~~~~~~~~~~~~~~~
- New compact format for local models.
- Adding cache managing capacities for all local predictors.
5.2.0 (2020-08-19)
~~~~~~~~~~~~~~~~~~
- Adding syntactic sugar methods to the BigML class to help managing resources.
5.1.1 (2020-08-11)
~~~~~~~~~~~~~~~~~~
- Fixing module directory inclusion and improving docs on local anomalies.
5.1.0 (2020-08-07)
~~~~~~~~~~~~~~~~~~
- Refactoring local anomaly to reduce memory requirements.
5.0.1 (2020-08-05)
~~~~~~~~~~~~~~~~~~
- Fixing bug in get_tasks_status to get information about transient net
errors.
5.0.0 (2020-07-31)
~~~~~~~~~~~~~~~~~~
- Deprecating support for Python 2.7.X versions. Only Python 3 supported
from this version on.
4.32.3 (2020-07-15)
~~~~~~~~~~~~~~~~~~~
- Extending the Fields class to check the attributes that can be updated in
a source or dataset fields structure to avoid failing fields updates.
4.32.2 (2020-06-15)
~~~~~~~~~~~~~~~~~~~
- Fixing local anomaly scores for new anomaly detectors with feedback and
setting the maximum input data precision to five digits.
4.32.1 (2020-06-10)
~~~~~~~~~~~~~~~~~~~
- Fixing local anomaly scores prediction for corner cases of samples with
one row.
4.32.0 (2020-05-19)
~~~~~~~~~~~~~~~~~~~
- Allowing scripts to be created from gists using the create_script method.
- Improving training examples generation in Fields class.
4.31.2 (2020-05-14)
~~~~~~~~~~~~~~~~~~~
- Fixing problems creating ephemeral prediction resources.
4.31.1 (2020-05-06)
~~~~~~~~~~~~~~~~~~~
- Improving the api.ok method to add an estimated wait time.
- Improving docs and adding TOC for new structure.
4.31.0 (2020-04-22)
~~~~~~~~~~~~~~~~~~~
- Adding REST methods to manage external data connections.
4.30.2 (2020-04-20)
~~~~~~~~~~~~~~~~~~~
- Fixing local anomaly scores for datasets with significant amounts of missings.
- Fixing input data modification for local predictions when fields are not
used in the models.
4.30.1 (2020-04-16)
~~~~~~~~~~~~~~~~~~~
- Fixing tasks status info for organizations.
4.30.0 (2020-04-10)
~~~~~~~~~~~~~~~~~~~
- Allowing the BigML class to retrieve any resource from local storage and
extract its fields.
4.29.2 (2020-03-20)
~~~~~~~~~~~~~~~~~~~
- Improving exception handling when retrieving resources.
4.29.1 (2020-03-03)
~~~~~~~~~~~~~~~~~~~
- Fixing bug when disabling SSL verification in predictions only.
4.29.0 (2020-02-29)
~~~~~~~~~~~~~~~~~~~
- Improving api.ok method to allow retries to avoid transient HTTP failures.
- Deprecating the `retries` argument in api.ok.
- Fixing local predictions confidence for weighted models.
4.28.1 (2020-02-04)
~~~~~~~~~~~~~~~~~~~
- Changing api.ok method to avoid raising exceptions when retrieving a
faulty resource.
- Adding call stack info to local Execution class.
- Fixing docs builder.
4.28.0 (2020-01-23)
~~~~~~~~~~~~~~~~~~~
- Adding Execution local utility to extract the outputs and results from an
execution.
4.27.3 (2020-01-15)
~~~~~~~~~~~~~~~~~~~
- Fixing local Fusion class to allow using linear regressions.
4.27.2 (2020-01-03)
~~~~~~~~~~~~~~~~~~~
- Fixing warning message and template files in generated code for hadoop
actionable models.
- Fixing local ensembles that asked for credentials before needing them.
4.27.1 (2019-12-19)
~~~~~~~~~~~~~~~~~~~
- Avoiding asking for credential in classes that predict locally when the
complete information is provided so no connection is needed.
4.27.0 (2019-12-03)
~~~~~~~~~~~~~~~~~~~
- Extending the custom formats for datetimes allowed as input for local
predictions.
- Fixing datetimes allowed as input for local predictions. They can be
provided by name or ID.
4.26.0 (2019-11-27)
~~~~~~~~~~~~~~~~~~~
- Extending the ability to use an alternative url to all predictions,
centroids, anomaly scores, etc. Also to their batch versions.
4.25.3 (2019-11-26)
~~~~~~~~~~~~~~~~~~~
- Changing bigml-chronos dependency version according to its new internal
structure. The previous version caused problems when used in some external
projects.
4.25.2 (2019-11-06)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local Cluster object when using text or item fields.
4.25.1 (2019-08-28)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local Fusion object when retrienving from storage.
4.25.0 (2019-08-18)
~~~~~~~~~~~~~~~~~~~
- Adding the ability to parse datetime fields locally for local predictions
(uses bigml-chronos as a dependency).
4.24.3 (2019-08-08)
~~~~~~~~~~~~~~~~~~~
- Fixing local LinearRegression to work even if `numpy` and `scipy` are
not installed.
4.24.2 (2019-07-30)
~~~~~~~~~~~~~~~~~~~
- Fixing local EnsemblePredictor code to avoid crash when using deep trees.
4.24.1 (2019-07-05)
~~~~~~~~~~~~~~~~~~~
- Adding missing tokens handling to local models.
4.24.0 (2019-06-28)
~~~~~~~~~~~~~~~~~~~
- Refactoring for multipackage compatibility.
- Deprecating ``ensemble_id`` attribute in local ensembles.
- Extending the BigML class to export model's alternative output formats.
4.23.1 (2019-06-06)
~~~~~~~~~~~~~~~~~~~
- Fixing local predictions for models with unpreferred and datetime fields.
4.23.0 (2019-05-24)
~~~~~~~~~~~~~~~~~~~
- Adding access to tasks information in the API connection object.
4.22.1 (2019-05-23)
~~~~~~~~~~~~~~~~~~~
- Improving the local Ensemble and Fusion classes to use the component
models when a local JSON file is used as argument.
4.22.0 (2019-05-11)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local linear regressions for non-invertible confidence bounds
matrices.
- Adding the option of cloning model resources from shared clonable ones.
- Fixing Fields object for timeseries.
4.21.2 (2019-04-09)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local fusion regression predictions.
4.21.1 (2019-04-06)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local linear regression predictions.
4.21.0 (2019-03-22)
~~~~~~~~~~~~~~~~~~~
- Adding REST and local methods for linear regression.
4.20.2 (2019-02-02)
~~~~~~~~~~~~~~~~~~~
- Adding new format for the list of datasets to create a multidataset from.
4.20.1 (2019-02-01)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local ensemble when used with externally defined connection,
as found by @KamalGalrani.
4.20.0 (2018-12-01)
~~~~~~~~~~~~~~~~~~~
- Adding PCA REST call methods.
- Adding local PCAs and Projections.
4.19.10 (2018-12-01)
~~~~~~~~~~~~~~~~~~~~
- Fixing local Deepnet predictions for regressions without numpy.
4.19.9 (2018-10-24)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in create datasets for a list of one dataset only.
4.19.8 (2018-09-18)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in create evaluation for timeseries.
4.19.7 (2018-09-13)
~~~~~~~~~~~~~~~~~~~
- Fixing bug when exporting fusions with weights.
- Local fusions now caching all models in the constructor.
4.19.6 (2018-09-12)
~~~~~~~~~~~~~~~~~~~
- Fixing bug when exporting fusions.
4.19.5 (2018-08-23)
~~~~~~~~~~~~~~~~~~~
- Changing source upload `async` parameter to ensure Python 3.7 compatibility.
4.19.4 (2018-07-18)
~~~~~~~~~~~~~~~~~~~
- Fixing local logistic regression predictions with weight field missing in
input data.
4.19.3 (2018-06-26)
~~~~~~~~~~~~~~~~~~~
- Modifying local fusion object to adapt to logistic regressions with
no missing numerics allowed.
4.19.2 (2018-06-25)
~~~~~~~~~~~~~~~~~~~
- Removing left over comment.
4.19.1 (2018-06-23)
~~~~~~~~~~~~~~~~~~~
- Refactoring the local classes that manage models information to create
predictions. Now all of them allow a path, an ID or a dictionary to be
the first argument in the constructor.
4.19.0 (2018-06-20)
~~~~~~~~~~~~~~~~~~~
- Adding local fusion object and predict methods.
- Fixing error handling in local objects.
- Fixing bug in local logistic regressions when using a local stored file.
4.18.3 (2018-06-03)
~~~~~~~~~~~~~~~~~~~
- Adding batch predictions for fusion resources.
4.18.2 (2018-05-28)
~~~~~~~~~~~~~~~~~~~
- Adding predictions and evaluations for fusion resources.
4.18.1 (2018-05-19)
~~~~~~~~~~~~~~~~~~~
- Fixing bug when unused field IDs are used in local prediction inputs.
4.18.0 (2018-05-19)
~~~~~~~~~~~~~~~~~~~
- Adding methods for the REST calls to OptiMLs and Fusions.
4.17.1 (2018-05-15)
~~~~~~~~~~~~~~~~~~~
- Adding the option to export PMML models when available.
- Fixing bug in local deepnets for regressions.
- Adapting local Cluster and Anomaly detector to not include summary fields
information.
4.17.0 (2018-05-02)
~~~~~~~~~~~~~~~~~~~
- Adding the local Supervised Model class to allow local predictions with
any supervised model resource.
4.16.2 (2018-04-31)
~~~~~~~~~~~~~~~~~~~
- Adding the `export` and `export_last` methods to download and save the
remote resources in the local file system.
4.16.1 (2018-04-24)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local deepnet predictions.
4.16.0 (2018-04-03)
~~~~~~~~~~~~~~~~~~~
- Deprecating local predictions formatting arguments. Formatting is available
through the `cast_prediction` function.
4.15.2 (2018-02-24)
~~~~~~~~~~~~~~~~~~~
- Local predictions for regression ensembles corrected for strange models
whose nodes lack the confidence attribute.
4.15.1 (2018-02-07)
~~~~~~~~~~~~~~~~~~~
- Removing logs left in local ensemble object.
4.15.0 (2018-02-07)
~~~~~~~~~~~~~~~~~~~
- Adding organizations support for all the API calls.
4.14.0 (2018-01-22)
~~~~~~~~~~~~~~~~~~~
- Deprecating `dev_mode` flag from BigML's API connection. The development
environment has been deprecated.
- Fixing bug in local cluster output to CSV.
- Improving docs with local batch predictions examples.
- Adding operating kind support for local predictions in models and ensembles.
- Fixing bug in ensembles local predictions with probability.
- Fixing bug in logistic regression local predictions with operating points.
4.13.7 (2018-01-02)
~~~~~~~~~~~~~~~~~~~
- Changing local predictions with threshold to meet changes in backend.
- Adding support for configurations REST API calls.
4.13.6 (2017-12-05)
~~~~~~~~~~~~~~~~~~~
- Fixing predict confidence method in local ensembles.
4.13.5 (2017-11-23)
~~~~~~~~~~~~~~~~~~~
- Adding operating point local predictions to deepnets.
4.13.4 (2017-11-21)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local ensemble predictions with operating points.
- Fixing bug for local EnsemblePredictor class.
4.13.3 (2017-11-14)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in local ensemble predictions for inputs that don't match the
expected field types.
4.13.2 (2017-11-14)
~~~~~~~~~~~~~~~~~~~
- Adding left out static files for local ensemble predictor functions.
4.13.1 (2017-11-10)
~~~~~~~~~~~~~~~~~~~
- Refactoring local BoostedTrees and adding the EnsemblePredictor to
use the local predict functions of each model to generate the ensemble
prediction.
4.13.0 (2017-11-07)
~~~~~~~~~~~~~~~~~~~
- Adding operating point thresholds to local model, ensemble and logistic
regression predictions.
4.12.1 (2017-10-12)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in the local Deepnet predictions when numpy is not installed.
4.12.0 (2017-10-04)
~~~~~~~~~~~~~~~~~~~
- Adding support for Deepnets REST API calls and local predictions using
the local Deepnet object.
4.11.3 (2017-09-29)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in the local Ensemble object. Failed to use the
stored ensemble object.
4.11.2 (2017-07-29)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in source uploads using Python3 when reading data from stdin.
4.11.1 (2017-06-23)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in source uploads using Python3 when a category is set.
4.11.0 (2017-06-23)
~~~~~~~~~~~~~~~~~~~
- Adding REST methods for managing time-series and local time-series object
to create forecasts.
4.10.5 (2017-07-13)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in the sources upload using Python3. Server changes need the
content-type of the file to be sent.
4.10.4 (2017-06-21)
~~~~~~~~~~~~~~~~~~~
- Fixing bug in the local model predicted distributions for weighted models.
- Fixing bug in predicted probability for local model predictions
using weighted models.
4.10.3 (2017-06-07)
~~~~~~~~~~~~~~~~~~~
- Changing boosted local ensembles predictions to match the improvements in
API.
- Fixing bug in association rules export to CSV and lisp for rules with numeric
attributes.
4.10.2 (2017-05-23)
~~~~~~~~~~~~~~~~~~~
- Fixing bug: local Model object failed when retrieving old JSON models from
local storage.
4.10.1 (2017-05-15)
~~~~~~~~~~~~~~~~~~~
- Internal refactoring preparing for extensions in BigMLer.
4.10.0 (2017-05-05)
~~~~~~~~~~~~~~~~~~~
- Adding predic_probability and predict_confidence methods to local model and
ensemble.
- Internal refactoring of local model classes preparing for extensions
in BigMLer.
4.9.2 (2017-03-26)
~~~~~~~~~~~~~~~~~~
- Fixing bug: local model slugifying fails when fields have empty names.
4.9.1 (2017-03-23)
~~~~~~~~~~~~~~~~~~
- Adding methods to local cluster: closest data points from a
reference point and centroids ordered from a reference point.
- Modifying internal codes in MultiVote class.
4.9.0 (2017-03-21)
~~~~~~~~~~~~~~~~~~
- Adding boosted ensembles to the local Ensemble object.
4.8.3 (2017-03-01)
~~~~~~~~~~~~~~~~~~
- Fixing bug in local logistic regression predictions when a constant field is
forced as input field.
4.8.2 (2017-02-09)
~~~~~~~~~~~~~~~~~~
- Fixing bug: Adapting to changes in Python 3.6 which cause the connection to
the API using SSL to fail.
4.8.1 (2017-01-11)
~~~~~~~~~~~~~~~~~~
- Changing local association parameters to adapt to API docs specifications.
4.8.0 (2017-01-08)
~~~~~~~~~~~~~~~~~~
- Adapting to final format of local association sets and adding tests.
4.7.3 (2016-12-03)
~~~~~~~~~~~~~~~~~~
- Bug fixing: query string is allowed also for project get calls.
4.7.2 (2016-12-02)
~~~~~~~~~~~~~~~~~~
- Allowing a query string to be added to get calls for all the resource types.
4.7.1 (2016-12-01)
~~~~~~~~~~~~~~~~~~
- Improving the Fields object: extracting fields structure from topic models.
- Bug fixing: Local Topic Distributions failed when tokenizing inputs with
sequences of separators.
4.7.0 (2016-11-30)
~~~~~~~~~~~~~~~~~~
- Adding REST methods for the new resource types: Topic Model,
Topic Distribution, Batch Topic Distribution.
- Adding local Topic Model object.
4.6.10 (2016-10-26)
~~~~~~~~~~~~~~~~~~~
- Improving local cluster object to fill in missing numerics for clusters
with default numeric values.
4.6.9 (2016-09-27)
~~~~~~~~~~~~~~~~~~
- Fixing bug in tests for anomaly detector and ill-formatted comments.
- Adapting tests to new logistic regression default value for balance_fields.
4.6.8 (2016-09-22)
~~~~~~~~~~~~~~~~~~
- Adding optional information to local predictions.
- Improving casting for booleans in local predictions.
- Improving the retrieval of stored or remote resources in local
predictor objects.
4.6.7 (2016-09-15)
~~~~~~~~~~~~~~~~~~
- Changing the type for the bias attribute to create logistic regressions to
boolean.
4.6.6 (2016-08-02)
~~~~~~~~~~~~~~~~~~
- Improving message for unauthorized API calls adding information about the
current domain.
4.6.5 (2016-07-16)
~~~~~~~~~~~~~~~~~~
- Fixing bug in local model. Fixing predictions for weighted models.
4.6.4 (2016-07-06)
~~~~~~~~~~~~~~~~~~
- Fixing bug in delete_execution method. The delete call now has a
query_string.
4.6.3 (2016-06-25)
~~~~~~~~~~~~~~~~~~
- Fixing bug in local logistic regression predictions' format.
4.6.2 (2016-06-20)
~~~~~~~~~~~~~~~~~~
- Adding local logistic regression as argument for evaluations.
4.6.1 (2016-06-12)
~~~~~~~~~~~~~~~~~~
- Adapting local logistic regression object to new coefficients format and
adding field_codings attribute.
4.6.0 (2016-05-19)
~~~~~~~~~~~~~~~~~~
- Adding REST methods to manage new types of whizzml resources: scripts,
executions and libraries.
- Fixing bug in logistic regression predictions for datases with text fields.
When input data has only one term and `all` token mode is used, local and
remote predictions didn't match.
4.5.3 (2016-05-04)
~~~~~~~~~~~~~~~~~~
- Improving the cluster report information.
- Fixing bug in logistic regression predictions. Results differred from
the backend predictions when date-time fields were present.
4.5.2 (2016-03-24)
~~~~~~~~~~~~~~~~~~
- Fixing bug in model's local predictions. When the model uses text fields and
the field contents are missing in the input data, the prediction does
not return the last prediction and stop. It now follows the
"does not contain" branch.
4.5.1 (2016-03-12)
~~~~~~~~~~~~~~~~~~
- Adding method to Fields object to produce CSV summary files.
- Adding method to Fields object to import changes in updatable attributes
from CSV files or strings.
4.5.0 (2016-02-08)
~~~~~~~~~~~~~~~~~~
- Adapting association object to the new syntax of missing values.
- Improving docs and comments for the proportional strategy in predictions.
- Fixing bug: centroid input data datetime fields are optional.
4.4.2 (2016-01-06)
~~~~~~~~~~~~~~~~~~
- Adapting logistic regression local object to the new missing_numeric
parameter.
4.4.1 (2015-12-18)
~~~~~~~~~~~~~~~~~~
- Fixing bug: summarized path output failed when adding missing operators.
4.4.0 (2015-12-15)
~~~~~~~~~~~~~~~~~~
- Adding REST API calls for association rules and local Association object.
- Adapting local model, cluster, anomaly and logistic regression objects
to new field type: items.
- Fixing bug: wrong value of giny impurity
- Fixing bug: local model summary failed occasionally when missings were used
in a numeric predicate.
- Fixing bug: wrong syntax in flatline filter method of the tree object.
4.3.4 (2015-12-10)
~~~~~~~~~~~~~~~~~~
- Fixing bug: Logistic regression object failed to build when using input
fields or non-preferred fields in dataset.
4.3.3 (2015-11-30)
~~~~~~~~~~~~~~~~~~
- Fixing bug: Anomaly object failed to generate the filter for new datasets
when text empty values were found.
4.3.2 (2015-11-24)
~~~~~~~~~~~~~~~~~~
- Adding verify and protocol options to the existing Domain class constructor
to handle special installs.
4.3.1 (2015-11-07)
~~~~~~~~~~~~~~~~~~
- Fixing bug: Local logistic regression predictions differ when input data
has contents in a text field but the terms involved do not appear in the
bag of words.
4.3.0 (2015-10-16)
~~~~~~~~~~~~~~~~~~
- Adding logistic regression as a new prediction model.
4.2.2 (2015-10-14)
~~~~~~~~~~~~~~~~~~
- Fixing bug: Fields object failed to store the correct objective id when the
objective was in the first column.
4.2.1 (2015-10-14)
~~~~~~~~~~~~~~~~~~
- Fixing bug: Improving error handling in download_dataset method.
4.2.0 (2015-07-27)
~~~~~~~~~~~~~~~~~~
- Adding REST methods to manage new type of resource: correlations.
- Adding REST methods to manage new type of resource: tests.
- Adding min and max values predictions for regression models and ensembles.
- Fixing bug: Fields object was not retrieving objective id from the
resource info.
4.1.7 (2015-08-15)
~~~~~~~~~~~~~~~~~~
- Fixing bug: console messages failed when used with Python3 on Windows.
4.1.6 (2015-06-25)
~~~~~~~~~~~~~~~~~~
- Fixing bug: Removing id fields from the filter to select the anomalies listed
in the Anomaly object from the origin dataset.
4.1.5 (2015-06-06)
~~~~~~~~~~~~~~~~~~
- Fixing bug: create_source method failed when unicode literals were used in
args.
4.1.4 (2015-05-27)
~~~~~~~~~~~~~~~~~~
- Ensuring unique ordering in MultiVote categorical combinations (only
needed in Python 3).
4.1.3 (2015-05-19)
~~~~~~~~~~~~~~~~~~
- Adapting code to handle uploading from String objects.
- Adding models creation new origin resources: clusters and centroids.
4.1.2 (2015-04-28)
~~~~~~~~~~~~~~~~~~
- Fixing bug in summarize method for local models. Ensuring unicode use and
adding tests for generated outputs.
4.1.1 (2015-04-26)
~~~~~~~~~~~~~~~~~~
- Fixing bug in method to print the fields in the anomaly trees.
- Fixing bug in the create_source method for Python3. Creation failed when
the `tags` argument was used.
4.1.0 (2015-04-14)
~~~~~~~~~~~~~~~~~~
- Adding median based predictions to ensembles.
4.0.2 (2015-04-12)
~~~~~~~~~~~~~~~~~~
- Fixing bug: multimodels median predictions failed.
4.0.1 (2015-04-10)
~~~~~~~~~~~~~~~~~~
- Adding support for median-based predictions in MultiModels.
4.0.0 (2015-04-10)
~~~~~~~~~~~~~~~~~~
- Python 3 added to supported Python versions.
- Test suite migrated to nose.
3.0.3 (2015-04-08)
~~~~~~~~~~~~~~~~~~
- Changing setup to ensure compatible Python and requests versions.
- Hiding warnings when SSL verification is disabled.
3.0.2 (2015-03-26)
~~~~~~~~~~~~~~~~~~
- Adding samples as Fields generator resources
3.0.1 (2015-03-17)
~~~~~~~~~~~~~~~~~~
- Changing the Ensemble object init method to use the max_models argument
also when loading the ensemble fields to trigger garbage collection.
3.0.0 (2015-03-04)
~~~~~~~~~~~~~~~~~~
- Adding Google App Engine support for remote REST calls.
- Adding cache_get argument to Ensemble constructor to allow getting
local model objects from cache.
2.2.0 (2015-02-26)
~~~~~~~~~~~~~~~~~~
- Adding lists of local models as argument for the local ensemble
constructor.
2.1.0 (2015-02-22)
~~~~~~~~~~~~~~~~~~
- Adding distribution and median to ensembles' predictions output.
2.0.0 (2015-02-12)
~~~~~~~~~~~~~~~~~~
- Adding REST API calls for samples.
1.10.8 (2015-02-10)
~~~~~~~~~~~~~~~~~~~
- Adding distribution units to the predict method output of the local model.
1.10.7 (2015-02-07)
~~~~~~~~~~~~~~~~~~~
- Extending the predict method in local models to get multiple predictions.
- Changing the local model object to add the units used in the distribution
and the add_median argument in the predict method.
1.10.6 (2015-02-06)
~~~~~~~~~~~~~~~~~~~
- Adding the median as prediction for the local model object.
1.10.5 (2014-01-29)
~~~~~~~~~~~~~~~~~~~
- Fixing bug: centroids failed when predicted from local clusters with
summary fields.
1.10.4 (2014-01-17)
~~~~~~~~~~~~~~~~~~~
- Improvements in docs presentation and content.
- Adding tree_CSV method to local model to output the nodes information
in CSV format.
1.10.3 (2014-01-16)
~~~~~~~~~~~~~~~~~~~
- Fixing bug: local ensembles were not retrieved from the stored JSON file.
- Adding the ability to construct local ensembles from any existing JSON file
describing an ensemble structure.
1.10.2 (2014-01-15)
~~~~~~~~~~~~~~~~~~~
- Source creation from inline data.
1.10.1 (2014-12-29)
~~~~~~~~~~~~~~~~~~~
- Fixing bug: source upload failed in old Python versions.
1.10.0 (2014-12-29)
~~~~~~~~~~~~~~~~~~~
- Refactoring the BigML class before adding the new project resource.
- Changing the ok and check_resource methods to download lighter resources.
- Fixing bug: cluster summarize for 1-centroid clusters.
- Fixing bug: adapting to new SSL verification in Python 2.7.9.
1.9.8 (2014-12-01)
~~~~~~~~~~~~~~~~~~
- Adding impurity to Model leaves, and a new method to select impure leaves.
- Fixing bug: the Model, Cluster and Anomaly objects had no resource_id
attribute when built from a local resource JSON structure.
1.9.7 (2014-11-24)
~~~~~~~~~~~~~~~~~~
- Adding method in Anomaly object to build the filter to exclude anomalies
from the original dataset.
- Basic code refactorization for initial resources structure.
1.9.6 (2014-11-09)
~~~~~~~~~~~~~~~~~~
- Adding BIGML_PROTOCOL, BIGML_SSL_VERIFY and BIGML_PREDICTION_SSL_VERIFY
environment variables to change the default corresponding values in
customized private environments.
1.9.5 (2014-11-03)
~~~~~~~~~~~~~~~~~~
- Fixing bug: summarize method breaks for clusters with text fields.
1.9.4 (2014-10-27)
~~~~~~~~~~~~~~~~~~
- Changing MultiModel class to return in-memory list of predictions.
1.9.3 (2014-10-23)
~~~~~~~~~~~~~~~~~~
- Improving Fields and including the new Cluster and
Anomalies fields structures as fields resources.
- Improving ModelFields to filter missing values from input data.
- Forcing garbage collection in local ensemble to lower memory usage.
1.9.2 (2014-10-13)
~~~~~~~~~~~~~~~~~~
- Changing some Fields exceptions handling.
- Refactoring api code to handle create, update and delete methods dynamically.
- Adding connection info string for printing.
- Improving tests information.
1.9.1 (2014-10-10)
~~~~~~~~~~~~~~~~~~
- Adding the summarize and statistics_CSV methods to local cluster object.
1.9.0 (2014-10-02)
~~~~~~~~~~~~~~~~~~
- Adding the batch anomaly score REST API calls.
1.8.0 (2014-09-09)
~~~~~~~~~~~~~~~~~~
- Adding the anomaly detector and anomaly score REST API calls.
- Adding the local anomaly detector.
1.7.0 (2014-08-29)
~~~~~~~~~~~~~~~~~~
- Adding to local model predictions the ability to use the new
missing-combined operators.
1.6.7 (2014-08-05)
~~~~~~~~~~~~~~~~~~
- Fixing bug in corner case of model predictions using proportional missing
strategy.
- Adding the unique path to the first missing split to the predictions using
proportional missing strategy.
1.6.6 (2014-07-31)
~~~~~~~~~~~~~~~~~~
- Improving the locale handling to avoid problems when logging to console under
Windows.
1.6.5 (2014-07-26)
~~~~~~~~~~~~~~~~~~
- Adding stats method to Fields to show fields statistics.
- Adding api method to create a source from a batch prediction.
1.6.4 (2014-07-25)
~~~~~~~~~~~~~~~~~~
- Changing the create methods to check if origin resources are finished
by downloading no fields information.
1.6.3 (2014-07-24)
~~~~~~~~~~~~~~~~~~
- Changing some variable names in the predict method (add_count, add_path) and
the prediction structure to follow other bindigns naming.
1.6.2 (2014-07-19)
~~~~~~~~~~~~~~~~~~
- Building local model from a JSON model file.
- Predictions output can contain confidence, distribution, instances and/or
rules.
1.6.1 (2014-07-09)
~~~~~~~~~~~~~~~~~~
- Fixing bug: download_dataset method did not return content when no filename
was provided.
1.6.0 (2014-07-03)
~~~~~~~~~~~~~~~~~~
- Fixing bug: check valid parameter in distribution merge function.
- Adding downlod_dataset method to api to export datasets to CSV.
1.5.1 (2014-06-13)
~~~~~~~~~~~~~~~~~~
- Fixing bug: local clusters' centroid method crashes when text or categorical
fields are not present in input data.
1.5.0 (2014-06-05)
~~~~~~~~~~~~~~~~~~
- Adding local cluster to produce centroid predictions locally.
1.4.4 (2014-05-23)
~~~~~~~~~~~~~~~~~~
- Adding shared urls to datasets.
- Fixing bug: error renaming variables.
1.4.3 (2014-05-22)
~~~~~~~~~~~~~~~~~~
- Adding the ability to change the remote server domain in the API
connection constructor (for VPCs).
- Adding the ability to generate datasets from clusters.
1.4.2 (2014-05-20)
~~~~~~~~~~~~~~~~~~
- Fixing bug when using api.ok method for centroids and batch centroids.
1.4.1 (2014-05-19)
~~~~~~~~~~~~~~~~~~
- Docs and test updates.
1.4.0 (2014-05-14)
~~~~~~~~~~~~~~~~~~
- Adding REST methods to manage clusters, centroids and batch centroids.
1.3.1 (2014-05-06)
~~~~~~~~~~~~~~~~~~
- Adding the average_confidence method to local models.
- Fixing bug in pprint for predictions with input data keyed by field names.
1.3.0 (2014-04-07)
~~~~~~~~~~~~~~~~~~
- Changing Fields object constructor to accept also source, dataset or model
resources.
1.2.2 (2014-04-01)
~~~~~~~~~~~~~~~~~~
- Changing error message when create_source calls result in http errors
to standarize them.
- Simplifying create_prediction calls because now API accepts field names
as input_data keys.
- Adding missing_counts and error_counts to report the missing values and
error counts per field in the dataset.
1.2.1 (2014-03-19)
~~~~~~~~~~~~~~~~~~
- Adding error to regression local predictions using proportional missing
strategy.
1.2.0 (2014-03-07)
~~~~~~~~~~~~~~~~~~
- Adding proportional missing strategy to MultiModel and solving tie breaks
in remote predictions.
- Adding new output options to model's python, rules and tableau outputs:
ability to extract the branch of the model leading to a certain node with
or without the hanging subtree.
- Adding HTTP_TOO_MANY_REQUESTS error handling in REST API calls.
1.1.0 (2014-02-10)
~~~~~~~~~~~~~~~~~~
- Adding Tableau-ready ouput to local model code generators.
1.0.6 (2014-02-03)
~~~~~~~~~~~~~~~~~~
- Fixing getters: getter for batch predictions was missing.
1.0.5 (2014-01-22)
~~~~~~~~~~~~~~~~~~
- Improving BaseModel and Model. If they receive a partial model
structure with a correct model id, the needed model resource is downloaded
and stored (if storage is enabled in the given api connection).
- Improving local ensemble. Adding a new `fields` attribute that
contains all the fields used in its models.
1.0.4 (2014-01-21)
~~~~~~~~~~~~~~~~~~
- Adding a summarize method to local ensembles with data distribution
and field importance information.
1.0.3 (2014-01-21)
~~~~~~~~~~~~~~~~~~
- Fixes bug in regressions predictions with ensembles and plurality without
confidence information. Predictions values were not normalized.
- Updating copyright information.
1.0.2 (2014-01-20)
~~~~~~~~~~~~~~~~~~
- Fixes bug in create calls: the user provided args dictionaries were
updated inside the calls.
1.0.1 (2014-01-05)
~~~~~~~~~~~~~~~~~~
- Changing the source for ensemble field importance computations.
- Fixes bug in http_ok adding the valid state for updates.
1.0.0 (2013-12-09)
~~~~~~~~~~~~~~~~~~
- Adding more info to error messages in REST methods.
- Adding new missing fields strategy in predict method.
- Fixes bug in shared models: credentials where not properly set.
- Adding batch predictions REST methods.
0.10.3 (2013-12-19)
~~~~~~~~~~~~~~~~~~~
- Fixes bug in local ensembles with more than 200 fields.
0.10.2 (2013-12-02)
~~~~~~~~~~~~~~~~~~~
- Fixes bug in summarize method of local models: field importance report
crashed.
- Fixes bug in status method of the BigML connection object: status for
async uploads of source files crashed while uploading.
0.10.1 (2013-11-25)
~~~~~~~~~~~~~~~~~~~
- Adding threshold combiner to MultiModel objects.
0.10.0 (2013-11-21)
~~~~~~~~~~~~~~~~~~~
- Adding a function printing field importance to ensembles.
- Changing Model to add a lightweight BaseModel class with no Tree
information.
- Adding function to get resource type from resource id or structure.
- Adding resource type checks to REST functions.
- Adding threshold as new combination method for local ensembles.
0.9.1 (2013-10-17)
~~~~~~~~~~~~~~~~~~
- Fixes duplication changing field names in local model if they are not unique.
0.9.0 (2013-10-08)
~~~~~~~~~~~~~~~~~~
- Adds the environment variables and adapts the create_prediction method
to create predictions using a different prediction server.
- Support for shared models.
0.8.0 (2013-08-10)
~~~~~~~~~~~~~~~~~~
- Adds text analysis local predict function
- Modifies outputs for text analysis: rules, summary, python, hadoop
0.7.5 (2013-08-22)
~~~~~~~~~~~~~~~~~~
- Fixes temporarily problems in predictions for regression models and
ensembles
- Adds en-gb to the list of available locales, avoiding spurious warnings
0.7.4 (2013-08-17)
~~~~~~~~~~~~~~~~~~
- Changes warning logger level to info
0.7.3 (2013-08-09)
~~~~~~~~~~~~~~~~~~
- Adds fields method to retrieve only preferred fields
- Fixes error message when no valid resource id is provided in check_resource
0.7.2 (2013-07-04)
~~~~~~~~~~~~~~~~~~
- Fixes check_resource method that was not using query-string data
- Add list of models as argument in Ensemble constructor
- MultiModel has BigML connection as a new optional argument
0.7.1 (2013-06-19)
~~~~~~~~~~~~~~~~~~
- Fixes Multimodel list_models method
- Fixes check_resource method for predictions
- Adds local configuration environment variable BIGML_DOMAIN replacing
BIGML_URL and BIGML_DEV_URL
- Refactors Ensemble and Model's predict method
0.7.0 (2013-05-01)
~~~~~~~~~~~~~~~~~~
- Adds splits in datasets to generate new datasets
- Adds evaluations for ensembles
0.6.0 (2013-04-27)
~~~~~~~~~~~~~~~~~~
- REST API methods for model ensembles
- New method returning the leaves of tree models
- Improved error handling in GET methods
0.5.2 (2013-03-03)
~~~~~~~~~~~~~~~~~~
- Adds combined confidence to combined predictions
- Fixes get_status for resources that have no status info
- Fixes bug: public datasets, that should be downloadable, weren't
0.5.1 (2013-02-12)
~~~~~~~~~~~~~~~~~~
- Fixes bug: no status info in public models, now shows FINISHED status code
- Adds more file-like objects (e.g. stdin) support in create_source input
- Refactoring Fields pair method and Model predict method to increase
- Adds some more locale aliases
0.5.0 (2013-01-16)
~~~~~~~~~~~~~~~~~~
- Adds evaluation api functions
- New prediction combination method: probability weighted
- Refactors MultiModels lists of predictions into MultiVote
- Multimodels partial predictions: new format
0.4.8 (2012-12-21)
~~~~~~~~~~~~~~~~~~
- Improved locale management
- Adds new features to MultiModel to allow local batch predictions
- Improved combined predictions
- Adds local predictions options: plurality, confidence weighted
0.4.7 (2012-12-06)
~~~~~~~~~~~~~~~~~~
- Warning message to inform of locale default if verbose mode
0.4.6 (2012-12-06)
~~~~~~~~~~~~~~~~~~
- Fix locale code for windows
0.4.5 (2012-12-05)
~~~~~~~~~~~~~~~~~~
- Fix remote predictions for input data containing fields not included in rules
0.4.4 (2012-12-02)
~~~~~~~~~~~~~~~~~~
- Tiny fixes
- Fix local predictions for input data containing fields not included in rules
- Overall clean up
0.4.3 (2012-11-07)
~~~~~~~~~~~~~~~~~~
- A few tiny fixes
- Multi models to generate predictions from multiple local models
- Adds hadoop-python code generation to create local predictions
0.4.2 (2012-09-19)
~~~~~~~~~~~~~~~~~~
- Fix Python generation
- Add a debug flag to log https requests and responses
- Type conversion in fields pairing
0.4.1 (2012-09-17)
~~~~~~~~~~~~~~~~~~
- Fix missing distribution field in new models
- Add new Field class to deal with BigML auto-generated ids
- Add by_name flag to predict methods to avoid reverse name lookups
- Add summarize method in models to generate class grouped printed output
0.4.0 (2012-08-20)
~~~~~~~~~~~~~~~~~~
- Development Mode
- Remote Sources
- Bigger files streamed with Poster
- Asynchronous Uploading
- Local Models
- Local Predictions
- Rule Generation
- Python Generation
- Overall clean up
0.3.1 (2012-07-05)
~~~~~~~~~~~~~~~~~~
- Initial release for the "andromeda" version of BigML.io.
Raw data
{
"_id": null,
"home_page": "https://bigml.com/developers",
"name": "bigml",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "",
"author": "The BigML Team",
"author_email": "bigml@bigml.com",
"download_url": "https://files.pythonhosted.org/packages/7a/b1/de6fee8bdc9f595aa947f0dd80b8aa40d8bb95a72ab8659bc1a3479167fe/bigml-9.7.1.tar.gz",
"platform": null,
"description": "BigML Python Bindings\n=====================\n\n`BigML <https://bigml.com>`_ makes machine learning easy by taking care\nof the details required to add data-driven decisions and predictive\npower to your company. Unlike other machine learning services, BigML\ncreates\n`beautiful predictive models <https://bigml.com/gallery/models>`_ that\ncan be easily understood and interacted with.\n\nThese BigML Python bindings allow you to interact with\n`BigML.io <https://bigml.io/>`_, the API\nfor BigML. You can use it to easily create, retrieve, list, update, and\ndelete BigML resources (i.e., sources, datasets, models and,\npredictions). For additional information, see\nthe `full documentation for the Python\nbindings on Read the Docs <http://bigml.readthedocs.org>`_.\n\nThis module is licensed under the `Apache License, Version\n2.0 <http://www.apache.org/licenses/LICENSE-2.0.html>`_.\n\nSupport\n-------\n\nPlease report problems and bugs to our `BigML.io issue\ntracker <https://github.com/bigmlcom/io/issues>`_.\n\nDiscussions about the different bindings take place in the general\n`BigML mailing list <http://groups.google.com/group/bigml>`_. Or join us\nin our `Campfire chatroom <https://bigmlinc.campfirenow.com/f20a0>`_.\n\nRequirements\n------------\n\nOnly ``Python 3`` versions are currently supported by these bindings.\nSupport for Python 2.7.X ended in version ``4.32.3``.\n\nThe basic third-party dependencies are the\n`requests <https://github.com/kennethreitz/requests>`_,\n`unidecode <http://pypi.python.org/pypi/Unidecode/#downloads>`_,\n`requests-toolbelt <https://pypi.python.org/pypi/requests-toolbelt>`_,\n`bigml-chronos <https://pypi.org/project/bigml-chronos>`_,\n`msgpack <https://pypi.org/project/msgpack>`_,\n`numpy <http://www.numpy.org/>`_ and\n`scipy <http://www.scipy.org/>`_ libraries. These\nlibraries are automatically installed during the basic setup.\nSupport for Google App Engine has been added as of version 3.0.0,\nusing the `urlfetch` package instead of `requests`.\n\nThe bindings will also use ``simplejson`` if you happen to have it\ninstalled, but that is optional: we fall back to Python's built-in JSON\nlibraries is ``simplejson`` is not found.\n\nThe bindings provide support to use the ``BigML`` platform to create, update,\nget and delete resources, but also to produce local predictions using the\nmodels created in ``BigML``. Most of them will be actionable with the basic\ninstallation, but some additional dependencies are needed to use local\n``Topic Models`` and Image Processing models. Please, refer to the\n`Installation <#installation>`_ section for details.\n\nOS Requirements\n~~~~~~~~~~~~~~~\n\nThe basic installation of the bindings is compatible and can be used\non Linux and Windows based Operating Systems.\nHowever, the extra options that allow working with\nimage processing models (``[images]`` and ``[full]``) are only supported\nand tested on Linux-based Operating Systems.\nFor image models, Windows OS is not recommended and cannot be supported out of\nthe box, because the specific compiler versions or dlls required are\nunavailable in general.\n\nInstallation\n------------\n\nTo install the basic latest stable release with\n`pip <http://www.pip-installer.org/>`_, please use:\n\n.. code-block:: bash\n\n $ pip install bigml\n\nSupport for local Topic Distributions (Topic Models' predictions)\nand local predictions for datasets that include Images will only be\navailable as extras, because the libraries used for that are not\nusually available in all Operative Systems. If you need to support those,\nplease check the `Installation Extras <#installation-extras>`_ section.\n\nInstallation Extras\n-------------------\n\nLocal Topic Distributions support can be installed using:\n\n.. code-block:: bash\n\n pip install bigml[topics]\n\nImages local predictions support can be installed using:\n\n.. code-block:: bash\n\n pip install bigml[images]\n\nThe full set of features can be installed using:\n\n.. code-block:: bash\n\n pip install bigml[full]\n\n\nWARNING: Mind that installing these extras can require some extra work, as\nexplained in the `Requirements <#requirements>`_ section.\n\nYou can also install the development version of the bindings directly\nfrom the Git repository\n\n.. code-block:: bash\n\n $ pip install -e git://github.com/bigmlcom/python.git#egg=bigml_python\n\n\nRunning the Tests\n-----------------\n\nThe tests will be run using `pytest <https://docs.pytest.org/en/7.2.x/>`_.\nYou'll need to set up your authentication\nvia environment variables, as explained\nin the authentication section. Also some of the tests need other environment\nvariables like ``BIGML_ORGANIZATION`` to test calls when used by Organization\nmembers and ``BIGML_EXTERNAL_CONN_HOST``, ``BIGML_EXTERNAL_CONN_PORT``,\n``BIGML_EXTERNAL_CONN_DB``, ``BIGML_EXTERNAL_CONN_USER``,\n``BIGML_EXTERNAL_CONN_PWD`` and ``BIGML_EXTERNAL_CONN_SOURCE``\nin order to test external data connectors.\n\nWith that in place, you can run the test suite simply by issuing\n\n.. code-block:: bash\n\n $ pytest\n\nAdditionally, `Tox <http://tox.testrun.org/>`_ can be used to\nautomatically run the test suite in virtual environments for all\nsupported Python versions. To install Tox:\n\n.. code-block:: bash\n\n $ pip install tox\n\nThen run the tests from the top-level project directory:\n\n.. code-block:: bash\n\n $ tox\n\nImporting the module\n--------------------\n\nTo import the module:\n\n.. code-block:: python\n\n import bigml.api\n\nAlternatively you can just import the BigML class:\n\n.. code-block:: python\n\n from bigml.api import BigML\n\nAuthentication\n--------------\n\nAll the requests to BigML.io must be authenticated using your username\nand `API key <https://bigml.com/account/apikey>`_ and are always\ntransmitted over HTTPS.\n\nThis module will look for your username and API key in the environment\nvariables ``BIGML_USERNAME`` and ``BIGML_API_KEY`` respectively.\n\nUnix and MacOS\n--------------\n\nYou can\nadd the following lines to your ``.bashrc`` or ``.bash_profile`` to set\nthose variables automatically when you log in:\n\n.. code-block:: bash\n\n export BIGML_USERNAME=myusername\n export BIGML_API_KEY=ae579e7e53fb9abd646a6ff8aa99d4afe83ac291\n\nrefer to the next chapters to know how to do that in other operating systems.\n\nWith that environment set up, connecting to BigML is a breeze:\n\n.. code-block:: python\n\n from bigml.api import BigML\n api = BigML()\n\nOtherwise, you can initialize directly when instantiating the BigML\nclass as follows:\n\n.. code-block:: python\n\n api = BigML('myusername', 'ae579e7e53fb9abd646a6ff8aa99d4afe83ac291')\n\nThese credentials will allow you to manage any resource in your user\nenvironment.\n\nIn BigML a user can also work for an ``organization``.\nIn this case, the organization administrator should previously assign\npermissions for the user to access one or several particular projects\nin the organization.\nOnce permissions are granted, the user can work with resources in a project\naccording to his permission level by creating a special constructor for\neach project. The connection constructor in this case\nshould include the ``project ID``:\n\n.. code-block:: python\n\n api = BigML('myusername', 'ae579e7e53fb9abd646a6ff8aa99d4afe83ac291',\n project='project/53739b98d994972da7001d4a')\n\nIf the project used in a connection object\ndoes not belong to an existing organization but is one of the\nprojects under the user's account, all the resources\ncreated or updated with that connection will also be assigned to the\nspecified project.\n\nWhen the resource to be managed is a ``project`` itself, the connection\nneeds to include the corresponding``organization ID``:\n\n.. code-block:: python\n\n api = BigML('myusername', 'ae579e7e53fb9abd646a6ff8aa99d4afe83ac291',\n organization='organization/53739b98d994972da7025d4a')\n\n\nAuthentication on Windows\n-------------------------\n\nThe credentials should be permanently stored in your system using\n\n.. code-block:: bash\n\n setx BIGML_USERNAME myusername\n setx BIGML_API_KEY ae579e7e53fb9abd646a6ff8aa99d4afe83ac291\n\nNote that ``setx`` will not change the environment variables of your actual\nconsole, so you will need to open a new one to start using them.\n\n\nAuthentication on Jupyter Notebook\n----------------------------------\n\nYou can set the environment variables using the ``%env`` command in your\ncells:\n\n.. code-block:: bash\n\n %env BIGML_USERNAME=myusername\n %env BIGML_API_KEY=ae579e7e53fb9abd646a6ff8aa99d4afe83ac291\n\n\nAlternative domains\n-------------------\n\n\nThe main public domain for the API service is ``bigml.io``, but there are some\nalternative domains, either for Virtual Private Cloud setups or\nthe australian subdomain (``au.bigml.io``). You can change the remote\nserver domain\nto the VPC particular one by either setting the ``BIGML_DOMAIN`` environment\nvariable to your VPC subdomain:\n\n.. code-block:: bash\n\n export BIGML_DOMAIN=my_VPC.bigml.io\n\nor setting it when instantiating your connection:\n\n.. code-block:: python\n\n api = BigML(domain=\"my_VPC.bigml.io\")\n\nThe corresponding SSL REST calls will be directed to your private domain\nhenceforth.\n\nYou can also set up your connection to use a particular PredictServer\nonly for predictions. In order to do so, you'll need to specify a ``Domain``\nobject, where you can set up the general domain name as well as the\nparticular prediction domain name.\n\n.. code-block:: python\n\n from bigml.domain import Domain\n from bigml.api import BigML\n\n domain_info = Domain(prediction_domain=\"my_prediction_server.bigml.com\",\n prediction_protocol=\"http\")\n\n api = BigML(domain=domain_info)\n\nFinally, you can combine all the options and change both the general domain\nserver, and the prediction domain server.\n\n.. code-block:: python\n\n from bigml.domain import Domain\n from bigml.api import BigML\n domain_info = Domain(domain=\"my_VPC.bigml.io\",\n prediction_domain=\"my_prediction_server.bigml.com\",\n prediction_protocol=\"https\")\n\n api = BigML(domain=domain_info)\n\nSome arguments for the Domain constructor are more unsual, but they can also\nbe used to set your special service endpoints:\n\n- protocol (string) Protocol for the service\n (when different from HTTPS)\n- verify (boolean) Sets on/off the SSL verification\n- prediction_verify (boolean) Sets on/off the SSL verification\n for the prediction server (when different from the general\n SSL verification)\n\n**Note** that the previously existing ``dev_mode`` flag:\n\n.. code-block:: python\n\n api = BigML(dev_mode=True)\n\nthat caused the connection to work with the Sandbox ``Development Environment``\nhas been **deprecated** because this environment does not longer exist.\nThe existing resources that were previously\ncreated in this environment have been moved\nto a special project in the now unique ``Production Environment``, so this\nflag is no longer needed to work with them.\n\n\nQuick Start\n-----------\n\nImagine that you want to use `this csv\nfile <https://static.bigml.com/csv/iris.csv>`_ containing the `Iris\nflower dataset <http://en.wikipedia.org/wiki/Iris_flower_data_set>`_ to\npredict the species of a flower whose ``petal length`` is ``2.45`` and\nwhose ``petal width`` is ``1.75``. A preview of the dataset is shown\nbelow. It has 4 numeric fields: ``sepal length``, ``sepal width``,\n``petal length``, ``petal width`` and a categorical field: ``species``.\nBy default, BigML considers the last field in the dataset as the\nobjective field (i.e., the field that you want to generate predictions\nfor).\n\n::\n\n sepal length,sepal width,petal length,petal width,species\n 5.1,3.5,1.4,0.2,Iris-setosa\n 4.9,3.0,1.4,0.2,Iris-setosa\n 4.7,3.2,1.3,0.2,Iris-setosa\n ...\n 5.8,2.7,3.9,1.2,Iris-versicolor\n 6.0,2.7,5.1,1.6,Iris-versicolor\n 5.4,3.0,4.5,1.5,Iris-versicolor\n ...\n 6.8,3.0,5.5,2.1,Iris-virginica\n 5.7,2.5,5.0,2.0,Iris-virginica\n 5.8,2.8,5.1,2.4,Iris-virginica\n\nYou can easily generate a prediction following these steps:\n\n.. code-block:: python\n\n from bigml.api import BigML\n\n api = BigML()\n\n source = api.create_source('./data/iris.csv')\n dataset = api.create_dataset(source)\n model = api.create_model(dataset)\n prediction = api.create_prediction(model, \\\n {\"petal width\": 1.75, \"petal length\": 2.45})\n\nYou can then print the prediction using the ``pprint`` method:\n\n.. code-block:: python\n\n >>> api.pprint(prediction)\n species for {\"petal width\": 1.75, \"petal length\": 2.45} is Iris-setosa\n\nCertainly, any of the resources created in BigML can be configured using\nseveral arguments described in the `API documentation <https://bigml.com/api>`_.\nAny of these configuration arguments can be added to the ``create`` method\nas a dictionary in the last optional argument of the calls:\n\n.. code-block:: python\n\n from bigml.api import BigML\n\n api = BigML()\n\n source_args = {\"name\": \"my source\",\n \"source_parser\": {\"missing_tokens\": [\"NULL\"]}}\n source = api.create_source('./data/iris.csv', source_args)\n dataset_args = {\"name\": \"my dataset\"}\n dataset = api.create_dataset(source, dataset_args)\n model_args = {\"objective_field\": \"species\"}\n model = api.create_model(dataset, model_args)\n prediction_args = {\"name\": \"my prediction\"}\n prediction = api.create_prediction(model, \\\n {\"petal width\": 1.75, \"petal length\": 2.45},\n prediction_args)\n\nThe ``iris`` dataset has a small number of instances, and usually will be\ninstantly created, so the ``api.create_`` calls will probably return the\nfinished resources outright. As BigML's API is asynchronous,\nin general you will need to ensure\nthat objects are finished before using them by using ``api.ok``.\n\n.. code-block:: python\n\n from bigml.api import BigML\n\n api = BigML()\n\n source = api.create_source('./data/iris.csv')\n api.ok(source)\n dataset = api.create_dataset(source)\n api.ok(dataset)\n model = api.create_model(dataset)\n api.ok(model)\n prediction = api.create_prediction(model, \\\n {\"petal width\": 1.75, \"petal length\": 2.45})\n\nNote that the prediction\ncall is not followed by the ``api.ok`` method. Predictions are so quick to be\ngenerated that, unlike the\nrest of resouces, will be generated synchronously as a finished object.\n\nThe example assumes that your objective field (the one you want to predict)\nis the last field in the dataset. If that's not he case, you can explicitly\nset the name of this field in the creation call using the ``objective_field``\nargument:\n\n\n.. code-block:: python\n\n from bigml.api import BigML\n\n api = BigML()\n\n source = api.create_source('./data/iris.csv')\n api.ok(source)\n dataset = api.create_dataset(source)\n api.ok(dataset)\n model = api.create_model(dataset, {\"objective_field\": \"species\"})\n api.ok(model)\n prediction = api.create_prediction(model, \\\n {'sepal length': 5, 'sepal width': 2.5})\n\n\nYou can also generate an evaluation for the model by using:\n\n.. code-block:: python\n\n test_source = api.create_source('./data/test_iris.csv')\n api.ok(test_source)\n test_dataset = api.create_dataset(test_source)\n api.ok(test_dataset)\n evaluation = api.create_evaluation(model, test_dataset)\n api.ok(evaluation)\n\nIf you set the ``storage`` argument in the ``api`` instantiation:\n\n.. code-block:: python\n\n api = BigML(storage='./storage')\n\nall the generated, updated or retrieved resources will be automatically\nsaved to the chosen directory.\n\nAlternatively, you can use the ``export`` method to explicitly\ndownload the JSON information\nthat describes any of your resources in BigML to a particular file:\n\n.. code-block:: python\n\n api.export('model/5acea49a08b07e14b9001068',\n filename=\"my_dir/my_model.json\")\n\nThis example downloads the JSON for the model and stores it in\nthe ``my_dir/my_model.json`` file.\n\nIn the case of models that can be represented in a `PMML` syntax, the\nexport method can be used to produce the corresponding `PMML` file.\n\n.. code-block:: python\n\n api.export('model/5acea49a08b07e14b9001068',\n filename=\"my_dir/my_model.pmml\",\n pmml=True)\n\nYou can also retrieve the last resource with some previously given tag:\n\n.. code-block:: python\n\n api.export_last(\"foo\",\n resource_type=\"ensemble\",\n filename=\"my_dir/my_ensemble.json\")\n\nwhich selects the last ensemble that has a ``foo`` tag. This mechanism can\nbe specially useful when retrieving retrained models that have been created\nwith a shared unique keyword as tag.\n\nFor a descriptive overview of the steps that you will usually need to\nfollow to model\nyour data and obtain predictions, please see the `basic Workflow sketch\n<api_sketch.html>`_\ndocument. You can also check other simple examples in the following documents:\n\n- `model 101 <101_model.html>`_\n- `logistic regression 101 <101_logistic_regression.html>`_\n- `linear regression 101 <101_linear_regression.html>`_\n- `ensemble 101 <101_ensemble.html>`_\n- `cluster 101 <101_cluster>`_\n- `anomaly detector 101 <101_anomaly.html>`_\n- `association 101 <101_association.html>`_\n- `topic model 101 <101_topic_model.html>`_\n- `deepnet 101 <101_deepnet.html>`_\n- `time series 101 <101_ts.html>`_\n- `fusion 101 <101_fusion.html>`_\n- `scripting 101 <101_scripting.html>`_\n\nAdditional Information\n----------------------\n\nWe've just barely scratched the surface. For additional information, see\nthe `full documentation for the Python\nbindings on Read the Docs <http://bigml.readthedocs.org>`_.\nAlternatively, the same documentation can be built from a local checkout\nof the source by installing `Sphinx <http://sphinx.pocoo.org>`_\n(``$ pip install sphinx``) and then running\n\n.. code-block:: bash\n\n $ cd docs\n $ make html\n\nThen launch ``docs/_build/html/index.html`` in your browser.\n\nHow to Contribute\n-----------------\n\nPlease follow the next steps:\n\n 1. Fork the project on github.com.\n 2. Create a new branch.\n 3. Commit changes to the new branch.\n 4. Send a `pull request <https://github.com/bigmlcom/python/pulls>`_.\n\n\nFor details on the underlying API, see the\n`BigML API documentation <https://bigml.com/developers>`_.\n\n\n.. :changelog:\n\nHistory\n-------\n\n9.7.1 (2023-12-08)\n------------------\n\n- Fixing readthedocs configuration.\n\n9.7.0 (2023-12-06)\n------------------\n\n- Changing query string separator in internall API calls.\n\n9.6.2 (2023-08-02)\n------------------\n\n- Extending cloning to all available models and WhizzML scripts.\n- Fixing shared resources cloning.\n\n9.6.1 (2023-08-01)\n------------------\n\n- Adding shared resources cloning.\n\n9.6.0 (2023-07-20)\n------------------\n\n- Adding ShapWrapper to enable local Shap values computation with the Shap\n library.\n- Adding Evaluation object.\n- Improving Field class to allow field values encoding as numpy arrays.\n\n9.5.0 (2023-06-16)\n------------------\n\n- Extending Local Fusions output to include confidence.\n\n9.4.0 (2023-06-14)\n------------------\n\n- Extending LocalModel class to handle Time Series locally.\n\n9.3.0 (2023-06-09)\n------------------\n\n- Adding a LocalModel class to handle any type of BigML model locally.\n\n9.2.0 (2023-04-11)\n------------------\n\n- Extending all delete methods to allow additional query strings.\n\n9.1.4 (2023-02-10)\n------------------\n\n- Providing local deepnet predictions default for Windows OS that cannot\n handle images predictions.\n\n9.1.3 (2022-12-22)\n------------------\n\n- Changing user's status endpoint retrieval to a lightweight version.\n\n9.1.2 (2022-11-26)\n------------------\n\n- Removing all nose dependencies in tests.\n\n9.1.1 (2022-11-18)\n------------------\n\n- Removing traces and refactoring Flatline interpreter invocation.\n- Migrating tests to pytest.\n\n9.1.0 (2022-11-09)\n------------------\n\n- Enabling local models to generate the transformations pipeline used in\n training to be able to apply it locally to input data for predictions.\n- Refactoring local models to bring the feature extraction capabilities to\n the transformations pipeline.\n- Rounding local numeric predictions and scores to 5 digits.\n- Improving Pipelines and Image Processing documentation.\n\n9.0.1 (2022-11-02)\n------------------\n\n- Fixing local batch_prediction method on DataFrames.\n- Improving the training examples method in the Fields class to produce\n examples of categorical field values following their original distribution.\n\n9.0.0 (2022-10-12)\n------------------\n\n- Renaming Pipeline class to BMLPipeline and refactoring to allow scikit-learn\n and Pandas pipelines to be also part of the Pipeline transformations.\n- Adding several DataTransformers (BMLDataTransformer, SKDataTransformer and\n DFDataTransformer) as wrappers for transformation generators.\n\n8.2.3 (2022-10-07)\n------------------\n\n- Fixing dump methods for all local resources.\n- Refactoring output options in batch_predict methods.\n- Adding name and description to local resources.\n\n8.2.2 (2022-09-29)\n------------------\n\n- Fixing locale check.\n- Documenting the new ``.batch_predict`` method added to local models to\n homogenize local batch predictions and accept Pandas' DataFrame as input.\n\n8.2.1 (2022-09-23)\n------------------\n\n- Fixing update annotations function when creating annotated images sources.\n\n8.2.0 (2022-09-22)\n------------------\n\n- Adding new option to api.ok to report progress via a function callback.\n\n8.1.1 (2022-09-21)\n------------------\n\n- Fixing bug in api.ok: response to failing API calls.\n\n8.1.0 (2022-09-16)\n------------------\n\n- Upgrading dependencies.\n\n8.0.0 (2022-09-13)\n------------------\n\n- Adding new local Pipeline object to encapsulate and run prediction\n workflows.\n- Changing ``api.ok`` response to return as False when retrieving faulty\n resources. Previously, only problems retrieving the API response cause\n ``api.ok`` returning False.\n- Adding FaultyResource exception to inform about that type of failure when\n using the ``raise_on_error`` flag in ``api.ok``.\n- Fixing bug in local LogisticRegression when predicting with operating points.\n\n7.7.0 (2022-08-30)\n~~~~~~~~~~~~~~~~~~\n\n- Adding local Dataset class to handle Flatline transformations.\n\n7.6.0 (2022-08-18)\n~~~~~~~~~~~~~~~~~~\n\n- Adding feature expansion to image fields in local predictors.\n\n7.5.1 (2022-08-08)\n~~~~~~~~~~~~~~~~~~\n\n- Adding utility to reproduce the backend images preprocess in local\n deepnets.\n\n7.5.0 (2022-07-09)\n~~~~~~~~~~~~~~~~~~\n\n- Adding the new capabilities the last API release: object detection.\n\n7.4.2 (2022-05-26)\n~~~~~~~~~~~~~~~~~~\n\n- Improved webhooks hashing check.\n- Adapting connection to allow API version setting, needed for compatibility\n with the Predict Server syntax.\n\n7.4.1 (2022-05-16)\n~~~~~~~~~~~~~~~~~~\n\n- Adding support for webhooks hashing check.\n\n7.4.0 (2022-05-13)\n~~~~~~~~~~~~~~~~~~\n\n- Adding extras to setup, to allow a basic installation and extra dependencies\n handling for topics and images.\n\n7.3.0 (2022-04-22)\n~~~~~~~~~~~~~~~~~~\n\n- Improving local ensembles to be generated from list of ensemble plus\n models info.\n- Allowing BigML connection class to retrieve models from storage without\n any credentials.\n\n7.2.2 (2022-04-20)\n~~~~~~~~~~~~~~~~~~\n\n- Enabling dump from SupervisedModel class.\n\n\n7.2.1 (2022-04-19)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing windows setup.\n\n7.2.0 (2022-04-14)\n~~~~~~~~~~~~~~~~~~\n\n- Upgrading sensenet library and giving fallback for local deepnet predictions\n on OS not supported by tensorflow.\n- Adding support for pandas' DataFrame when creating sources.\n\n7.1.2 (2022-03-17)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing local deepnet predictions if no missing_count info is found.\n\n7.1.1 (2022-03-11)\n~~~~~~~~~~~~~~~~~~\n\n- Improving local predictions for shared fusions and documenting.\n\n7.1.0 (2022-03-10)\n~~~~~~~~~~~~~~~~~~\n\n- Adding the ability to predict locally from shared composed resources, like\n Ensembles and Fusions.\n\n7.0.1 (2022-03-04)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing local Supervised Model to accept stored models as input.\n- Improving tests to minimize resource creation.\n- Updating dependencies versions.\n\n7.0.0 (2021-08-03)\n~~~~~~~~~~~~~~~~~~\n\n- Adding composite sources to the API REST interface methods.\n- Adding predictions for images.\n\n6.3.2 (2021-12-02)\n------------------\n\n- Fixing local deepnet prediction method to fit new deepnet structure.\n\n6.3.1 (2021-11-03)\n------------------\n\n- Improving local ensembles to handle faulty models with no root.\n\n6.3.0 (2021-10-21)\n------------------\n\n- Adding activators to local deepnets prediction functions.\n\n6.2.2 (2021-10-11)\n------------------\n\n- Fixing underlying dependency used in datetime fields.\n\n6.2.1 (2021-08-10)\n------------------\n\n- Fixing predictions for local models with text fields.\n\n6.2.0 (2021-07-22)\n------------------\n\n- Minimizing local model information when text or items fields are used.\n\n6.1.2 (2021-06-16)\n------------------\n\n- Changing tests mechanism to github actions.\n\n6.1.1 (2021-06-11)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing api `get` method. The query string argument was lost.\n- Improving api `get`, `create` and `update` methods to avoid unneeded\n iterations.\n- Improving docs.\n\n6.1.0 (2021-03-06)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing local anomalies when normalize_repeats has been used.\n\n6.0.3 (2021-02-06)\n~~~~~~~~~~~~~~~~~~\n\n- Improving local models with default numeric values to automatically fill in\n missings in input data.\n- Raising exceptions if trying to update a failed resource.\n\n6.0.2 (2020-12-21)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local anomaly filter.\n\n6.0.1 (2020-11-11)\n~~~~~~~~~~~~~~~~~~\n\n- Changing local anomaly filter.\n- Fixing update and download methods to wait for the resource to be finished.\n\n6.0.0 (2020-09-02)\n~~~~~~~~~~~~~~~~~~\n\n- New compact format for local models.\n- Adding cache managing capacities for all local predictors.\n\n5.2.0 (2020-08-19)\n~~~~~~~~~~~~~~~~~~\n\n- Adding syntactic sugar methods to the BigML class to help managing resources.\n\n5.1.1 (2020-08-11)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing module directory inclusion and improving docs on local anomalies.\n\n5.1.0 (2020-08-07)\n~~~~~~~~~~~~~~~~~~\n\n- Refactoring local anomaly to reduce memory requirements.\n\n\n5.0.1 (2020-08-05)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in get_tasks_status to get information about transient net\n errors.\n\n5.0.0 (2020-07-31)\n~~~~~~~~~~~~~~~~~~\n\n- Deprecating support for Python 2.7.X versions. Only Python 3 supported\n from this version on.\n\n4.32.3 (2020-07-15)\n~~~~~~~~~~~~~~~~~~~\n\n- Extending the Fields class to check the attributes that can be updated in\n a source or dataset fields structure to avoid failing fields updates.\n\n4.32.2 (2020-06-15)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local anomaly scores for new anomaly detectors with feedback and\n setting the maximum input data precision to five digits.\n\n4.32.1 (2020-06-10)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local anomaly scores prediction for corner cases of samples with\n one row.\n\n4.32.0 (2020-05-19)\n~~~~~~~~~~~~~~~~~~~\n\n- Allowing scripts to be created from gists using the create_script method.\n- Improving training examples generation in Fields class.\n\n4.31.2 (2020-05-14)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing problems creating ephemeral prediction resources.\n\n4.31.1 (2020-05-06)\n~~~~~~~~~~~~~~~~~~~\n\n- Improving the api.ok method to add an estimated wait time.\n- Improving docs and adding TOC for new structure.\n\n4.31.0 (2020-04-22)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding REST methods to manage external data connections.\n\n4.30.2 (2020-04-20)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local anomaly scores for datasets with significant amounts of missings.\n- Fixing input data modification for local predictions when fields are not\n used in the models.\n\n4.30.1 (2020-04-16)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing tasks status info for organizations.\n\n4.30.0 (2020-04-10)\n~~~~~~~~~~~~~~~~~~~\n\n- Allowing the BigML class to retrieve any resource from local storage and\n extract its fields.\n\n4.29.2 (2020-03-20)\n~~~~~~~~~~~~~~~~~~~\n\n- Improving exception handling when retrieving resources.\n\n4.29.1 (2020-03-03)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug when disabling SSL verification in predictions only.\n\n4.29.0 (2020-02-29)\n~~~~~~~~~~~~~~~~~~~\n\n- Improving api.ok method to allow retries to avoid transient HTTP failures.\n- Deprecating the `retries` argument in api.ok.\n- Fixing local predictions confidence for weighted models.\n\n4.28.1 (2020-02-04)\n~~~~~~~~~~~~~~~~~~~\n\n- Changing api.ok method to avoid raising exceptions when retrieving a\n faulty resource.\n- Adding call stack info to local Execution class.\n- Fixing docs builder.\n\n4.28.0 (2020-01-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding Execution local utility to extract the outputs and results from an\n execution.\n\n4.27.3 (2020-01-15)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local Fusion class to allow using linear regressions.\n\n4.27.2 (2020-01-03)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing warning message and template files in generated code for hadoop\n actionable models.\n- Fixing local ensembles that asked for credentials before needing them.\n\n4.27.1 (2019-12-19)\n~~~~~~~~~~~~~~~~~~~\n\n- Avoiding asking for credential in classes that predict locally when the\n complete information is provided so no connection is needed.\n\n4.27.0 (2019-12-03)\n~~~~~~~~~~~~~~~~~~~\n\n- Extending the custom formats for datetimes allowed as input for local\n predictions.\n- Fixing datetimes allowed as input for local predictions. They can be\n provided by name or ID.\n\n4.26.0 (2019-11-27)\n~~~~~~~~~~~~~~~~~~~\n\n- Extending the ability to use an alternative url to all predictions,\n centroids, anomaly scores, etc. Also to their batch versions.\n\n4.25.3 (2019-11-26)\n~~~~~~~~~~~~~~~~~~~\n\n- Changing bigml-chronos dependency version according to its new internal\n structure. The previous version caused problems when used in some external\n projects.\n\n4.25.2 (2019-11-06)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local Cluster object when using text or item fields.\n\n4.25.1 (2019-08-28)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local Fusion object when retrienving from storage.\n\n4.25.0 (2019-08-18)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding the ability to parse datetime fields locally for local predictions\n (uses bigml-chronos as a dependency).\n\n4.24.3 (2019-08-08)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local LinearRegression to work even if `numpy` and `scipy` are\n not installed.\n\n4.24.2 (2019-07-30)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local EnsemblePredictor code to avoid crash when using deep trees.\n\n4.24.1 (2019-07-05)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding missing tokens handling to local models.\n\n4.24.0 (2019-06-28)\n~~~~~~~~~~~~~~~~~~~\n\n- Refactoring for multipackage compatibility.\n- Deprecating ``ensemble_id`` attribute in local ensembles.\n- Extending the BigML class to export model's alternative output formats.\n\n4.23.1 (2019-06-06)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local predictions for models with unpreferred and datetime fields.\n\n4.23.0 (2019-05-24)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding access to tasks information in the API connection object.\n\n4.22.1 (2019-05-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Improving the local Ensemble and Fusion classes to use the component\n models when a local JSON file is used as argument.\n\n4.22.0 (2019-05-11)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local linear regressions for non-invertible confidence bounds\n matrices.\n- Adding the option of cloning model resources from shared clonable ones.\n- Fixing Fields object for timeseries.\n\n4.21.2 (2019-04-09)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local fusion regression predictions.\n\n4.21.1 (2019-04-06)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local linear regression predictions.\n\n4.21.0 (2019-03-22)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding REST and local methods for linear regression.\n\n\n4.20.2 (2019-02-02)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding new format for the list of datasets to create a multidataset from.\n\n4.20.1 (2019-02-01)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local ensemble when used with externally defined connection,\n as found by @KamalGalrani.\n\n4.20.0 (2018-12-01)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding PCA REST call methods.\n- Adding local PCAs and Projections.\n\n4.19.10 (2018-12-01)\n~~~~~~~~~~~~~~~~~~~~\n\n- Fixing local Deepnet predictions for regressions without numpy.\n\n4.19.9 (2018-10-24)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in create datasets for a list of one dataset only.\n\n4.19.8 (2018-09-18)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in create evaluation for timeseries.\n\n4.19.7 (2018-09-13)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug when exporting fusions with weights.\n- Local fusions now caching all models in the constructor.\n\n4.19.6 (2018-09-12)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug when exporting fusions.\n\n4.19.5 (2018-08-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Changing source upload `async` parameter to ensure Python 3.7 compatibility.\n\n4.19.4 (2018-07-18)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing local logistic regression predictions with weight field missing in\n input data.\n\n4.19.3 (2018-06-26)\n~~~~~~~~~~~~~~~~~~~\n\n- Modifying local fusion object to adapt to logistic regressions with\n no missing numerics allowed.\n\n4.19.2 (2018-06-25)\n~~~~~~~~~~~~~~~~~~~\n\n- Removing left over comment.\n\n4.19.1 (2018-06-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Refactoring the local classes that manage models information to create\n predictions. Now all of them allow a path, an ID or a dictionary to be\n the first argument in the constructor.\n\n4.19.0 (2018-06-20)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding local fusion object and predict methods.\n- Fixing error handling in local objects.\n- Fixing bug in local logistic regressions when using a local stored file.\n\n4.18.3 (2018-06-03)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding batch predictions for fusion resources.\n\n4.18.2 (2018-05-28)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding predictions and evaluations for fusion resources.\n\n4.18.1 (2018-05-19)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug when unused field IDs are used in local prediction inputs.\n\n4.18.0 (2018-05-19)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding methods for the REST calls to OptiMLs and Fusions.\n\n4.17.1 (2018-05-15)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding the option to export PMML models when available.\n- Fixing bug in local deepnets for regressions.\n- Adapting local Cluster and Anomaly detector to not include summary fields\n information.\n\n4.17.0 (2018-05-02)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding the local Supervised Model class to allow local predictions with\n any supervised model resource.\n\n4.16.2 (2018-04-31)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding the `export` and `export_last` methods to download and save the\n remote resources in the local file system.\n\n4.16.1 (2018-04-24)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local deepnet predictions.\n\n4.16.0 (2018-04-03)\n~~~~~~~~~~~~~~~~~~~\n\n- Deprecating local predictions formatting arguments. Formatting is available\n through the `cast_prediction` function.\n\n4.15.2 (2018-02-24)\n~~~~~~~~~~~~~~~~~~~\n\n- Local predictions for regression ensembles corrected for strange models\n whose nodes lack the confidence attribute.\n\n4.15.1 (2018-02-07)\n~~~~~~~~~~~~~~~~~~~\n\n- Removing logs left in local ensemble object.\n\n4.15.0 (2018-02-07)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding organizations support for all the API calls.\n\n4.14.0 (2018-01-22)\n~~~~~~~~~~~~~~~~~~~\n\n- Deprecating `dev_mode` flag from BigML's API connection. The development\n environment has been deprecated.\n- Fixing bug in local cluster output to CSV.\n- Improving docs with local batch predictions examples.\n- Adding operating kind support for local predictions in models and ensembles.\n- Fixing bug in ensembles local predictions with probability.\n- Fixing bug in logistic regression local predictions with operating points.\n\n4.13.7 (2018-01-02)\n~~~~~~~~~~~~~~~~~~~\n\n- Changing local predictions with threshold to meet changes in backend.\n- Adding support for configurations REST API calls.\n\n4.13.6 (2017-12-05)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing predict confidence method in local ensembles.\n\n4.13.5 (2017-11-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding operating point local predictions to deepnets.\n\n4.13.4 (2017-11-21)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local ensemble predictions with operating points.\n- Fixing bug for local EnsemblePredictor class.\n\n4.13.3 (2017-11-14)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local ensemble predictions for inputs that don't match the\n expected field types.\n\n4.13.2 (2017-11-14)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding left out static files for local ensemble predictor functions.\n\n4.13.1 (2017-11-10)\n~~~~~~~~~~~~~~~~~~~\n\n- Refactoring local BoostedTrees and adding the EnsemblePredictor to\n use the local predict functions of each model to generate the ensemble\n prediction.\n\n4.13.0 (2017-11-07)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding operating point thresholds to local model, ensemble and logistic\n regression predictions.\n\n4.12.1 (2017-10-12)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in the local Deepnet predictions when numpy is not installed.\n\n4.12.0 (2017-10-04)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding support for Deepnets REST API calls and local predictions using\n the local Deepnet object.\n\n4.11.3 (2017-09-29)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in the local Ensemble object. Failed to use the\n stored ensemble object.\n\n4.11.2 (2017-07-29)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in source uploads using Python3 when reading data from stdin.\n\n4.11.1 (2017-06-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in source uploads using Python3 when a category is set.\n\n4.11.0 (2017-06-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding REST methods for managing time-series and local time-series object\n to create forecasts.\n\n4.10.5 (2017-07-13)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in the sources upload using Python3. Server changes need the\n content-type of the file to be sent.\n\n4.10.4 (2017-06-21)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in the local model predicted distributions for weighted models.\n- Fixing bug in predicted probability for local model predictions\n using weighted models.\n\n4.10.3 (2017-06-07)\n~~~~~~~~~~~~~~~~~~~\n\n- Changing boosted local ensembles predictions to match the improvements in\n API.\n- Fixing bug in association rules export to CSV and lisp for rules with numeric\n attributes.\n\n4.10.2 (2017-05-23)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: local Model object failed when retrieving old JSON models from\n local storage.\n\n4.10.1 (2017-05-15)\n~~~~~~~~~~~~~~~~~~~\n\n- Internal refactoring preparing for extensions in BigMLer.\n\n4.10.0 (2017-05-05)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding predic_probability and predict_confidence methods to local model and\n ensemble.\n- Internal refactoring of local model classes preparing for extensions\n in BigMLer.\n\n4.9.2 (2017-03-26)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: local model slugifying fails when fields have empty names.\n\n4.9.1 (2017-03-23)\n~~~~~~~~~~~~~~~~~~\n\n- Adding methods to local cluster: closest data points from a\n reference point and centroids ordered from a reference point.\n- Modifying internal codes in MultiVote class.\n\n4.9.0 (2017-03-21)\n~~~~~~~~~~~~~~~~~~\n\n- Adding boosted ensembles to the local Ensemble object.\n\n4.8.3 (2017-03-01)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local logistic regression predictions when a constant field is\n forced as input field.\n\n4.8.2 (2017-02-09)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: Adapting to changes in Python 3.6 which cause the connection to\n the API using SSL to fail.\n\n4.8.1 (2017-01-11)\n~~~~~~~~~~~~~~~~~~\n\n- Changing local association parameters to adapt to API docs specifications.\n\n4.8.0 (2017-01-08)\n~~~~~~~~~~~~~~~~~~\n\n- Adapting to final format of local association sets and adding tests.\n\n4.7.3 (2016-12-03)\n~~~~~~~~~~~~~~~~~~\n\n- Bug fixing: query string is allowed also for project get calls.\n\n4.7.2 (2016-12-02)\n~~~~~~~~~~~~~~~~~~\n\n- Allowing a query string to be added to get calls for all the resource types.\n\n4.7.1 (2016-12-01)\n~~~~~~~~~~~~~~~~~~\n\n- Improving the Fields object: extracting fields structure from topic models.\n- Bug fixing: Local Topic Distributions failed when tokenizing inputs with\n sequences of separators.\n\n4.7.0 (2016-11-30)\n~~~~~~~~~~~~~~~~~~\n\n- Adding REST methods for the new resource types: Topic Model,\n Topic Distribution, Batch Topic Distribution.\n- Adding local Topic Model object.\n\n4.6.10 (2016-10-26)\n~~~~~~~~~~~~~~~~~~~\n\n- Improving local cluster object to fill in missing numerics for clusters\n with default numeric values.\n\n4.6.9 (2016-09-27)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in tests for anomaly detector and ill-formatted comments.\n- Adapting tests to new logistic regression default value for balance_fields.\n\n4.6.8 (2016-09-22)\n~~~~~~~~~~~~~~~~~~\n\n- Adding optional information to local predictions.\n- Improving casting for booleans in local predictions.\n- Improving the retrieval of stored or remote resources in local\n predictor objects.\n\n4.6.7 (2016-09-15)\n~~~~~~~~~~~~~~~~~~\n\n- Changing the type for the bias attribute to create logistic regressions to\n boolean.\n\n4.6.6 (2016-08-02)\n~~~~~~~~~~~~~~~~~~\n\n- Improving message for unauthorized API calls adding information about the\n current domain.\n\n4.6.5 (2016-07-16)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local model. Fixing predictions for weighted models.\n\n4.6.4 (2016-07-06)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in delete_execution method. The delete call now has a\n query_string.\n\n4.6.3 (2016-06-25)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in local logistic regression predictions' format.\n\n4.6.2 (2016-06-20)\n~~~~~~~~~~~~~~~~~~\n\n- Adding local logistic regression as argument for evaluations.\n\n4.6.1 (2016-06-12)\n~~~~~~~~~~~~~~~~~~\n\n- Adapting local logistic regression object to new coefficients format and\n adding field_codings attribute.\n\n4.6.0 (2016-05-19)\n~~~~~~~~~~~~~~~~~~\n\n- Adding REST methods to manage new types of whizzml resources: scripts,\n executions and libraries.\n- Fixing bug in logistic regression predictions for datases with text fields.\n When input data has only one term and `all` token mode is used, local and\n remote predictions didn't match.\n\n4.5.3 (2016-05-04)\n~~~~~~~~~~~~~~~~~~\n\n- Improving the cluster report information.\n- Fixing bug in logistic regression predictions. Results differred from\n the backend predictions when date-time fields were present.\n\n4.5.2 (2016-03-24)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in model's local predictions. When the model uses text fields and\n the field contents are missing in the input data, the prediction does\n not return the last prediction and stop. It now follows the\n \"does not contain\" branch.\n\n4.5.1 (2016-03-12)\n~~~~~~~~~~~~~~~~~~\n\n- Adding method to Fields object to produce CSV summary files.\n- Adding method to Fields object to import changes in updatable attributes\n from CSV files or strings.\n\n4.5.0 (2016-02-08)\n~~~~~~~~~~~~~~~~~~\n\n- Adapting association object to the new syntax of missing values.\n- Improving docs and comments for the proportional strategy in predictions.\n- Fixing bug: centroid input data datetime fields are optional.\n\n4.4.2 (2016-01-06)\n~~~~~~~~~~~~~~~~~~\n\n- Adapting logistic regression local object to the new missing_numeric\n parameter.\n\n4.4.1 (2015-12-18)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: summarized path output failed when adding missing operators.\n\n4.4.0 (2015-12-15)\n~~~~~~~~~~~~~~~~~~\n\n- Adding REST API calls for association rules and local Association object.\n- Adapting local model, cluster, anomaly and logistic regression objects\n to new field type: items.\n- Fixing bug: wrong value of giny impurity\n- Fixing bug: local model summary failed occasionally when missings were used\n in a numeric predicate.\n- Fixing bug: wrong syntax in flatline filter method of the tree object.\n\n4.3.4 (2015-12-10)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: Logistic regression object failed to build when using input\n fields or non-preferred fields in dataset.\n\n4.3.3 (2015-11-30)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: Anomaly object failed to generate the filter for new datasets\n when text empty values were found.\n\n4.3.2 (2015-11-24)\n~~~~~~~~~~~~~~~~~~\n\n- Adding verify and protocol options to the existing Domain class constructor\n to handle special installs.\n\n4.3.1 (2015-11-07)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: Local logistic regression predictions differ when input data\n has contents in a text field but the terms involved do not appear in the\n bag of words.\n\n4.3.0 (2015-10-16)\n~~~~~~~~~~~~~~~~~~\n\n- Adding logistic regression as a new prediction model.\n\n4.2.2 (2015-10-14)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: Fields object failed to store the correct objective id when the\n objective was in the first column.\n\n4.2.1 (2015-10-14)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: Improving error handling in download_dataset method.\n\n4.2.0 (2015-07-27)\n~~~~~~~~~~~~~~~~~~\n\n- Adding REST methods to manage new type of resource: correlations.\n- Adding REST methods to manage new type of resource: tests.\n- Adding min and max values predictions for regression models and ensembles.\n- Fixing bug: Fields object was not retrieving objective id from the\n resource info.\n\n4.1.7 (2015-08-15)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: console messages failed when used with Python3 on Windows.\n\n4.1.6 (2015-06-25)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: Removing id fields from the filter to select the anomalies listed\n in the Anomaly object from the origin dataset.\n\n4.1.5 (2015-06-06)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: create_source method failed when unicode literals were used in\n args.\n\n4.1.4 (2015-05-27)\n~~~~~~~~~~~~~~~~~~\n\n- Ensuring unique ordering in MultiVote categorical combinations (only\n needed in Python 3).\n\n4.1.3 (2015-05-19)\n~~~~~~~~~~~~~~~~~~\n\n- Adapting code to handle uploading from String objects.\n- Adding models creation new origin resources: clusters and centroids.\n\n4.1.2 (2015-04-28)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in summarize method for local models. Ensuring unicode use and\n adding tests for generated outputs.\n\n4.1.1 (2015-04-26)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in method to print the fields in the anomaly trees.\n- Fixing bug in the create_source method for Python3. Creation failed when\n the `tags` argument was used.\n\n4.1.0 (2015-04-14)\n~~~~~~~~~~~~~~~~~~\n\n- Adding median based predictions to ensembles.\n\n4.0.2 (2015-04-12)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: multimodels median predictions failed.\n\n4.0.1 (2015-04-10)\n~~~~~~~~~~~~~~~~~~\n\n- Adding support for median-based predictions in MultiModels.\n\n4.0.0 (2015-04-10)\n~~~~~~~~~~~~~~~~~~\n\n- Python 3 added to supported Python versions.\n- Test suite migrated to nose.\n\n\n3.0.3 (2015-04-08)\n~~~~~~~~~~~~~~~~~~\n\n- Changing setup to ensure compatible Python and requests versions.\n- Hiding warnings when SSL verification is disabled.\n\n3.0.2 (2015-03-26)\n~~~~~~~~~~~~~~~~~~\n\n- Adding samples as Fields generator resources\n\n3.0.1 (2015-03-17)\n~~~~~~~~~~~~~~~~~~\n\n- Changing the Ensemble object init method to use the max_models argument\n also when loading the ensemble fields to trigger garbage collection.\n\n3.0.0 (2015-03-04)\n~~~~~~~~~~~~~~~~~~\n\n- Adding Google App Engine support for remote REST calls.\n- Adding cache_get argument to Ensemble constructor to allow getting\n local model objects from cache.\n\n2.2.0 (2015-02-26)\n~~~~~~~~~~~~~~~~~~\n\n- Adding lists of local models as argument for the local ensemble\n constructor.\n\n2.1.0 (2015-02-22)\n~~~~~~~~~~~~~~~~~~\n\n- Adding distribution and median to ensembles' predictions output.\n\n2.0.0 (2015-02-12)\n~~~~~~~~~~~~~~~~~~\n\n- Adding REST API calls for samples.\n\n1.10.8 (2015-02-10)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding distribution units to the predict method output of the local model.\n\n1.10.7 (2015-02-07)\n~~~~~~~~~~~~~~~~~~~\n\n- Extending the predict method in local models to get multiple predictions.\n- Changing the local model object to add the units used in the distribution\n and the add_median argument in the predict method.\n\n1.10.6 (2015-02-06)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding the median as prediction for the local model object.\n\n1.10.5 (2014-01-29)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: centroids failed when predicted from local clusters with\n summary fields.\n\n1.10.4 (2014-01-17)\n~~~~~~~~~~~~~~~~~~~\n\n- Improvements in docs presentation and content.\n- Adding tree_CSV method to local model to output the nodes information\n in CSV format.\n\n1.10.3 (2014-01-16)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: local ensembles were not retrieved from the stored JSON file.\n- Adding the ability to construct local ensembles from any existing JSON file\n describing an ensemble structure.\n\n1.10.2 (2014-01-15)\n~~~~~~~~~~~~~~~~~~~\n\n- Source creation from inline data.\n\n1.10.1 (2014-12-29)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: source upload failed in old Python versions.\n\n1.10.0 (2014-12-29)\n~~~~~~~~~~~~~~~~~~~\n\n- Refactoring the BigML class before adding the new project resource.\n- Changing the ok and check_resource methods to download lighter resources.\n- Fixing bug: cluster summarize for 1-centroid clusters.\n- Fixing bug: adapting to new SSL verification in Python 2.7.9.\n\n1.9.8 (2014-12-01)\n~~~~~~~~~~~~~~~~~~\n\n- Adding impurity to Model leaves, and a new method to select impure leaves.\n- Fixing bug: the Model, Cluster and Anomaly objects had no resource_id\n attribute when built from a local resource JSON structure.\n\n1.9.7 (2014-11-24)\n~~~~~~~~~~~~~~~~~~\n\n- Adding method in Anomaly object to build the filter to exclude anomalies\n from the original dataset.\n- Basic code refactorization for initial resources structure.\n\n1.9.6 (2014-11-09)\n~~~~~~~~~~~~~~~~~~\n\n- Adding BIGML_PROTOCOL, BIGML_SSL_VERIFY and BIGML_PREDICTION_SSL_VERIFY\n environment variables to change the default corresponding values in\n customized private environments.\n\n1.9.5 (2014-11-03)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: summarize method breaks for clusters with text fields.\n\n1.9.4 (2014-10-27)\n~~~~~~~~~~~~~~~~~~\n\n- Changing MultiModel class to return in-memory list of predictions.\n\n1.9.3 (2014-10-23)\n~~~~~~~~~~~~~~~~~~\n\n- Improving Fields and including the new Cluster and\n Anomalies fields structures as fields resources.\n- Improving ModelFields to filter missing values from input data.\n- Forcing garbage collection in local ensemble to lower memory usage.\n\n1.9.2 (2014-10-13)\n~~~~~~~~~~~~~~~~~~\n\n- Changing some Fields exceptions handling.\n- Refactoring api code to handle create, update and delete methods dynamically.\n- Adding connection info string for printing.\n- Improving tests information.\n\n1.9.1 (2014-10-10)\n~~~~~~~~~~~~~~~~~~\n\n- Adding the summarize and statistics_CSV methods to local cluster object.\n\n1.9.0 (2014-10-02)\n~~~~~~~~~~~~~~~~~~\n\n- Adding the batch anomaly score REST API calls.\n\n1.8.0 (2014-09-09)\n~~~~~~~~~~~~~~~~~~\n\n- Adding the anomaly detector and anomaly score REST API calls.\n- Adding the local anomaly detector.\n\n1.7.0 (2014-08-29)\n~~~~~~~~~~~~~~~~~~\n\n- Adding to local model predictions the ability to use the new\n missing-combined operators.\n\n1.6.7 (2014-08-05)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug in corner case of model predictions using proportional missing\n strategy.\n- Adding the unique path to the first missing split to the predictions using\n proportional missing strategy.\n\n1.6.6 (2014-07-31)\n~~~~~~~~~~~~~~~~~~\n\n- Improving the locale handling to avoid problems when logging to console under\n Windows.\n\n1.6.5 (2014-07-26)\n~~~~~~~~~~~~~~~~~~\n\n- Adding stats method to Fields to show fields statistics.\n- Adding api method to create a source from a batch prediction.\n\n1.6.4 (2014-07-25)\n~~~~~~~~~~~~~~~~~~\n\n- Changing the create methods to check if origin resources are finished\n by downloading no fields information.\n\n1.6.3 (2014-07-24)\n~~~~~~~~~~~~~~~~~~\n\n- Changing some variable names in the predict method (add_count, add_path) and\n the prediction structure to follow other bindigns naming.\n\n1.6.2 (2014-07-19)\n~~~~~~~~~~~~~~~~~~\n\n- Building local model from a JSON model file.\n- Predictions output can contain confidence, distribution, instances and/or\n rules.\n\n1.6.1 (2014-07-09)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: download_dataset method did not return content when no filename\n was provided.\n\n1.6.0 (2014-07-03)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: check valid parameter in distribution merge function.\n- Adding downlod_dataset method to api to export datasets to CSV.\n\n1.5.1 (2014-06-13)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug: local clusters' centroid method crashes when text or categorical\n fields are not present in input data.\n\n1.5.0 (2014-06-05)\n~~~~~~~~~~~~~~~~~~\n\n- Adding local cluster to produce centroid predictions locally.\n\n1.4.4 (2014-05-23)\n~~~~~~~~~~~~~~~~~~\n\n- Adding shared urls to datasets.\n- Fixing bug: error renaming variables.\n\n1.4.3 (2014-05-22)\n~~~~~~~~~~~~~~~~~~\n\n- Adding the ability to change the remote server domain in the API\n connection constructor (for VPCs).\n- Adding the ability to generate datasets from clusters.\n\n1.4.2 (2014-05-20)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing bug when using api.ok method for centroids and batch centroids.\n\n1.4.1 (2014-05-19)\n~~~~~~~~~~~~~~~~~~\n\n- Docs and test updates.\n\n1.4.0 (2014-05-14)\n~~~~~~~~~~~~~~~~~~\n\n- Adding REST methods to manage clusters, centroids and batch centroids.\n\n1.3.1 (2014-05-06)\n~~~~~~~~~~~~~~~~~~\n\n- Adding the average_confidence method to local models.\n- Fixing bug in pprint for predictions with input data keyed by field names.\n\n1.3.0 (2014-04-07)\n~~~~~~~~~~~~~~~~~~\n\n- Changing Fields object constructor to accept also source, dataset or model\n resources.\n\n1.2.2 (2014-04-01)\n~~~~~~~~~~~~~~~~~~\n\n- Changing error message when create_source calls result in http errors\n to standarize them.\n- Simplifying create_prediction calls because now API accepts field names\n as input_data keys.\n- Adding missing_counts and error_counts to report the missing values and\n error counts per field in the dataset.\n\n1.2.1 (2014-03-19)\n~~~~~~~~~~~~~~~~~~\n\n- Adding error to regression local predictions using proportional missing\n strategy.\n\n1.2.0 (2014-03-07)\n~~~~~~~~~~~~~~~~~~\n\n- Adding proportional missing strategy to MultiModel and solving tie breaks\n in remote predictions.\n- Adding new output options to model's python, rules and tableau outputs:\n ability to extract the branch of the model leading to a certain node with\n or without the hanging subtree.\n- Adding HTTP_TOO_MANY_REQUESTS error handling in REST API calls.\n\n1.1.0 (2014-02-10)\n~~~~~~~~~~~~~~~~~~\n\n- Adding Tableau-ready ouput to local model code generators.\n\n1.0.6 (2014-02-03)\n~~~~~~~~~~~~~~~~~~\n\n- Fixing getters: getter for batch predictions was missing.\n\n1.0.5 (2014-01-22)\n~~~~~~~~~~~~~~~~~~\n\n- Improving BaseModel and Model. If they receive a partial model\n structure with a correct model id, the needed model resource is downloaded\n and stored (if storage is enabled in the given api connection).\n- Improving local ensemble. Adding a new `fields` attribute that\n contains all the fields used in its models.\n\n1.0.4 (2014-01-21)\n~~~~~~~~~~~~~~~~~~\n\n- Adding a summarize method to local ensembles with data distribution\n and field importance information.\n\n1.0.3 (2014-01-21)\n~~~~~~~~~~~~~~~~~~\n\n- Fixes bug in regressions predictions with ensembles and plurality without\n confidence information. Predictions values were not normalized.\n- Updating copyright information.\n\n1.0.2 (2014-01-20)\n~~~~~~~~~~~~~~~~~~\n\n- Fixes bug in create calls: the user provided args dictionaries were\n updated inside the calls.\n\n1.0.1 (2014-01-05)\n~~~~~~~~~~~~~~~~~~\n\n- Changing the source for ensemble field importance computations.\n- Fixes bug in http_ok adding the valid state for updates.\n\n1.0.0 (2013-12-09)\n~~~~~~~~~~~~~~~~~~\n\n- Adding more info to error messages in REST methods.\n- Adding new missing fields strategy in predict method.\n- Fixes bug in shared models: credentials where not properly set.\n- Adding batch predictions REST methods.\n\n0.10.3 (2013-12-19)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixes bug in local ensembles with more than 200 fields.\n\n0.10.2 (2013-12-02)\n~~~~~~~~~~~~~~~~~~~\n\n- Fixes bug in summarize method of local models: field importance report\n crashed.\n- Fixes bug in status method of the BigML connection object: status for\n async uploads of source files crashed while uploading.\n\n0.10.1 (2013-11-25)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding threshold combiner to MultiModel objects.\n\n0.10.0 (2013-11-21)\n~~~~~~~~~~~~~~~~~~~\n\n- Adding a function printing field importance to ensembles.\n- Changing Model to add a lightweight BaseModel class with no Tree\n information.\n- Adding function to get resource type from resource id or structure.\n- Adding resource type checks to REST functions.\n- Adding threshold as new combination method for local ensembles.\n\n0.9.1 (2013-10-17)\n~~~~~~~~~~~~~~~~~~\n\n- Fixes duplication changing field names in local model if they are not unique.\n\n0.9.0 (2013-10-08)\n~~~~~~~~~~~~~~~~~~\n\n- Adds the environment variables and adapts the create_prediction method\n to create predictions using a different prediction server.\n- Support for shared models.\n\n0.8.0 (2013-08-10)\n~~~~~~~~~~~~~~~~~~\n\n- Adds text analysis local predict function\n- Modifies outputs for text analysis: rules, summary, python, hadoop\n\n0.7.5 (2013-08-22)\n~~~~~~~~~~~~~~~~~~\n\n- Fixes temporarily problems in predictions for regression models and\n ensembles\n- Adds en-gb to the list of available locales, avoiding spurious warnings\n\n0.7.4 (2013-08-17)\n~~~~~~~~~~~~~~~~~~\n\n- Changes warning logger level to info\n\n0.7.3 (2013-08-09)\n~~~~~~~~~~~~~~~~~~\n\n- Adds fields method to retrieve only preferred fields\n- Fixes error message when no valid resource id is provided in check_resource\n\n0.7.2 (2013-07-04)\n~~~~~~~~~~~~~~~~~~\n\n- Fixes check_resource method that was not using query-string data\n- Add list of models as argument in Ensemble constructor\n- MultiModel has BigML connection as a new optional argument\n\n0.7.1 (2013-06-19)\n~~~~~~~~~~~~~~~~~~\n\n- Fixes Multimodel list_models method\n- Fixes check_resource method for predictions\n- Adds local configuration environment variable BIGML_DOMAIN replacing\n BIGML_URL and BIGML_DEV_URL\n- Refactors Ensemble and Model's predict method\n\n0.7.0 (2013-05-01)\n~~~~~~~~~~~~~~~~~~\n\n- Adds splits in datasets to generate new datasets\n- Adds evaluations for ensembles\n\n0.6.0 (2013-04-27)\n~~~~~~~~~~~~~~~~~~\n\n- REST API methods for model ensembles\n- New method returning the leaves of tree models\n- Improved error handling in GET methods\n\n0.5.2 (2013-03-03)\n~~~~~~~~~~~~~~~~~~\n\n- Adds combined confidence to combined predictions\n- Fixes get_status for resources that have no status info\n- Fixes bug: public datasets, that should be downloadable, weren't\n\n0.5.1 (2013-02-12)\n~~~~~~~~~~~~~~~~~~\n\n- Fixes bug: no status info in public models, now shows FINISHED status code\n- Adds more file-like objects (e.g. stdin) support in create_source input\n- Refactoring Fields pair method and Model predict method to increase\n- Adds some more locale aliases\n\n0.5.0 (2013-01-16)\n~~~~~~~~~~~~~~~~~~\n\n- Adds evaluation api functions\n- New prediction combination method: probability weighted\n- Refactors MultiModels lists of predictions into MultiVote\n- Multimodels partial predictions: new format\n\n0.4.8 (2012-12-21)\n~~~~~~~~~~~~~~~~~~\n\n- Improved locale management\n- Adds new features to MultiModel to allow local batch predictions\n- Improved combined predictions\n- Adds local predictions options: plurality, confidence weighted\n\n0.4.7 (2012-12-06)\n~~~~~~~~~~~~~~~~~~\n\n- Warning message to inform of locale default if verbose mode\n\n0.4.6 (2012-12-06)\n~~~~~~~~~~~~~~~~~~\n\n- Fix locale code for windows\n\n0.4.5 (2012-12-05)\n~~~~~~~~~~~~~~~~~~\n\n- Fix remote predictions for input data containing fields not included in rules\n\n0.4.4 (2012-12-02)\n~~~~~~~~~~~~~~~~~~\n\n- Tiny fixes\n- Fix local predictions for input data containing fields not included in rules\n- Overall clean up\n\n0.4.3 (2012-11-07)\n~~~~~~~~~~~~~~~~~~\n\n- A few tiny fixes\n- Multi models to generate predictions from multiple local models\n- Adds hadoop-python code generation to create local predictions\n\n0.4.2 (2012-09-19)\n~~~~~~~~~~~~~~~~~~\n\n- Fix Python generation\n- Add a debug flag to log https requests and responses\n- Type conversion in fields pairing\n\n0.4.1 (2012-09-17)\n~~~~~~~~~~~~~~~~~~\n\n- Fix missing distribution field in new models\n- Add new Field class to deal with BigML auto-generated ids\n- Add by_name flag to predict methods to avoid reverse name lookups\n- Add summarize method in models to generate class grouped printed output\n\n0.4.0 (2012-08-20)\n~~~~~~~~~~~~~~~~~~\n\n- Development Mode\n- Remote Sources\n- Bigger files streamed with Poster\n- Asynchronous Uploading\n- Local Models\n- Local Predictions\n- Rule Generation\n- Python Generation\n- Overall clean up\n\n\n0.3.1 (2012-07-05)\n~~~~~~~~~~~~~~~~~~\n\n- Initial release for the \"andromeda\" version of BigML.io.\n",
"bugtrack_url": null,
"license": "http://www.apache.org/licenses/LICENSE-2.0",
"summary": "An open source binding to BigML.io, the public BigML API",
"version": "9.7.1",
"project_urls": {
"Download": "https://github.com/bigmlcom/python",
"Homepage": "https://bigml.com/developers"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "adca8bc283072cf7aed18aca4b330ae1ea217a20e1d6ff1117888eef4baac07f",
"md5": "e11d260a3a6af64171d28a2e23d3de96",
"sha256": "7506d24b17bb19e475880cd28b85941f5393d0d2d3c8039365358821edeed8e1"
},
"downloads": -1,
"filename": "bigml-9.7.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "e11d260a3a6af64171d28a2e23d3de96",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 791274,
"upload_time": "2023-12-08T16:56:49",
"upload_time_iso_8601": "2023-12-08T16:56:49.637136Z",
"url": "https://files.pythonhosted.org/packages/ad/ca/8bc283072cf7aed18aca4b330ae1ea217a20e1d6ff1117888eef4baac07f/bigml-9.7.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "7ab1de6fee8bdc9f595aa947f0dd80b8aa40d8bb95a72ab8659bc1a3479167fe",
"md5": "303a19032aec75a514e1d7f510cd0da7",
"sha256": "e0f08ec8f15347c25b371798549017af7c012b6a0eebb5cf9a519dd20486ae82"
},
"downloads": -1,
"filename": "bigml-9.7.1.tar.gz",
"has_sig": false,
"md5_digest": "303a19032aec75a514e1d7f510cd0da7",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 25140159,
"upload_time": "2023-12-08T16:57:05",
"upload_time_iso_8601": "2023-12-08T16:57:05.120265Z",
"url": "https://files.pythonhosted.org/packages/7a/b1/de6fee8bdc9f595aa947f0dd80b8aa40d8bb95a72ab8659bc1a3479167fe/bigml-9.7.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-12-08 16:57:05",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "bigmlcom",
"github_project": "python",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "bigml"
}
JSON
