.. image:: https://raw.githubusercontent.com/mars-project/mars/master/docs/source/images/mars-logo-title.png
|PyPI version| |Docs| |Build| |Coverage| |Quality| |License|
Mars is a tensor-based unified framework for large-scale data computation which scales numpy, pandas, scikit-learn and many other libraries.
Documentation, 中文文档
Mars is easy to install by
.. code-block:: bash
pip install pymarsInstallation for Developers
When you want to contribute code to Mars, you can follow the instructions below to install Mars
for development:
.. code-block:: bash
git clone https://github.com/mars-project/mars.git
cd mars
pip install -e ".[dev]"
More details about installing Mars can be found at
`installation <https://mars-project.readthedocs.io/en/latest/installation/index.html>`_ section in
Mars document.
Architecture Overview
---------------------
.. image:: https://raw.githubusercontent.com/mars-project/mars/master/docs/source/images/architecture.png
Getting Started
---------------
Starting a new runtime locally via:
.. code-block:: python
>>> import mars
>>> mars.new_session()
Or connecting to a Mars cluster which is already initialized.
.. code-block:: python
>>> import mars
>>> mars.new_session('http://<web_ip>:<ui_port>')
Mars Tensor
-----------
Mars tensor provides a familiar interface like Numpy.
+-----------------------------------------------+-----------------------------------------------+
| **Numpy** | **Mars tensor** |
+-----------------------------------------------+-----------------------------------------------+
|.. code-block:: python |.. code-block:: python |
| | |
| import numpy as np | import mars.tensor as mt |
| N = 200_000_000 | N = 200_000_000 |
| a = np.random.uniform(-1, 1, size=(N, 2)) | a = mt.random.uniform(-1, 1, size=(N, 2)) |
| print((np.linalg.norm(a, axis=1) < 1) | print(((mt.linalg.norm(a, axis=1) < 1) |
| .sum() * 4 / N) | .sum() * 4 / N).execute()) |
| | |
+-----------------------------------------------+-----------------------------------------------+
|.. code-block:: |.. code-block:: |
| | |
| 3.14174502 | 3.14161908 |
| CPU times: user 11.6 s, sys: 8.22 s, | CPU times: user 966 ms, sys: 544 ms, |
| total: 19.9 s | total: 1.51 s |
| Wall time: 22.5 s | Wall time: 3.77 s |
| | |
+-----------------------------------------------+-----------------------------------------------+
Mars can leverage multiple cores, even on a laptop, and could be even faster for a distributed setting.
Mars DataFrame
--------------
Mars DataFrame provides a familiar interface like pandas.
+-----------------------------------------+-----------------------------------------+
| **Pandas** | **Mars DataFrame** |
+-----------------------------------------+-----------------------------------------+
|.. code-block:: python |.. code-block:: python |
| | |
| import numpy as np | import mars.tensor as mt |
| import pandas as pd | import mars.dataframe as md |
| df = pd.DataFrame( | df = md.DataFrame( |
| np.random.rand(100000000, 4), | mt.random.rand(100000000, 4), |
| columns=list('abcd')) | columns=list('abcd')) |
| print(df.sum()) | print(df.sum().execute()) |
| | |
+-----------------------------------------+-----------------------------------------+
|.. code-block:: |.. code-block:: |
| | |
| CPU times: user 10.9 s, sys: 2.69 s, | CPU times: user 1.21 s, sys: 212 ms, |
| total: 13.6 s | total: 1.42 s |
| Wall time: 11 s | Wall time: 2.75 s |
+-----------------------------------------+-----------------------------------------+
Mars Learn
----------
Mars learn provides a familiar interface like scikit-learn.
+---------------------------------------------+----------------------------------------------------+
| **Scikit-learn** | **Mars learn** |
+---------------------------------------------+----------------------------------------------------+
|.. code-block:: python |.. code-block:: python |
| | |
| from sklearn.datasets import make_blobs | from mars.learn.datasets import make_blobs |
| from sklearn.decomposition import PCA | from mars.learn.decomposition import PCA |
| X, y = make_blobs( | X, y = make_blobs( |
| n_samples=100000000, n_features=3, | n_samples=100000000, n_features=3, |
| centers=[[3, 3, 3], [0, 0, 0], | centers=[[3, 3, 3], [0, 0, 0], |
| [1, 1, 1], [2, 2, 2]], | [1, 1, 1], [2, 2, 2]], |
| cluster_std=[0.2, 0.1, 0.2, 0.2], | cluster_std=[0.2, 0.1, 0.2, 0.2], |
| random_state=9) | random_state=9) |
| pca = PCA(n_components=3) | pca = PCA(n_components=3) |
| pca.fit(X) | pca.fit(X) |
| print(pca.explained_variance_ratio_) | print(pca.explained_variance_ratio_) |
| print(pca.explained_variance_) | print(pca.explained_variance_) |
| | |
+---------------------------------------------+----------------------------------------------------+
Mars learn also integrates with many libraries:
- `TensorFlow <https://mars-project.readthedocs.io//en/latest/user_guide/learn/tensorflow.html>`_
- `PyTorch <https://mars-project.readthedocs.io/en/latest/user_guide/learn/pytorch.html>`_
- `XGBoost <https://mars-project.readthedocs.io/en/latest/user_guide/learn/xgboost.html>`_
- `LightGBM <https://mars-project.readthedocs.io/en/latest/user_guide/learn/lightgbm.html>`_
- `Joblib <https://mars-project.readthedocs.io/en/latest/user_guide/learn/joblib.html>`_
- `Statsmodels <https://mars-project.readthedocs.io/en/latest/user_guide/learn/statsmodels.html>`_
Mars remote
-----------
Mars remote allows users to execute functions in parallel.
+-------------------------------------------+--------------------------------------------+
| **Vanilla function calls** | **Mars remote** |
+-------------------------------------------+--------------------------------------------+
|.. code-block:: python |.. code-block:: python |
| | |
| import numpy as np | import numpy as np |
| | import mars.remote as mr |
| | |
| def calc_chunk(n, i): | def calc_chunk(n, i): |
| rs = np.random.RandomState(i) | rs = np.random.RandomState(i) |
| a = rs.uniform(-1, 1, size=(n, 2)) | a = rs.uniform(-1, 1, size=(n, 2)) |
| d = np.linalg.norm(a, axis=1) | d = np.linalg.norm(a, axis=1) |
| return (d < 1).sum() | return (d < 1).sum() |
| | |
| def calc_pi(fs, N): | def calc_pi(fs, N): |
| return sum(fs) * 4 / N | return sum(fs) * 4 / N |
| | |
| N = 200_000_000 | N = 200_000_000 |
| n = 10_000_000 | n = 10_000_000 |
| | |
| fs = [calc_chunk(n, i) | fs = [mr.spawn(calc_chunk, args=(n, i)) |
| for i in range(N // n)] | for i in range(N // n)] |
| pi = calc_pi(fs, N) | pi = mr.spawn(calc_pi, args=(fs, N)) |
| print(pi) | print(pi.execute().fetch()) |
| | |
+-------------------------------------------+--------------------------------------------+
|.. code-block:: |.. code-block:: |
| | |
| 3.1416312 | 3.1416312 |
| CPU times: user 32.2 s, sys: 4.86 s, | CPU times: user 616 ms, sys: 307 ms, |
| total: 37.1 s | total: 923 ms |
| Wall time: 12.4 s | Wall time: 3.99 s |
| | |
+-------------------------------------------+--------------------------------------------+
DASK on Mars
------------
Refer to `DASK on Mars`_ for more information.
Eager ModeMars supports eager mode which makes it friendly for developing and easy to debug.
Users can enable the eager mode by options, set options at the beginning of the program or console session.
.. code-block:: python
>>> from mars.config import options
>>> options.eager_mode = TrueOr use a context.
.. code-block:: python
>>> from mars.config import option_context
>>> with option_context() as options:
>>> options.eager_mode = True
>>> # the eager mode is on only for the with statement
>>> ...If eager mode is on, tensor, DataFrame etc will be executed immediately by default session once it is created.
.. code-block:: python
>>> import mars.tensor as mt
>>> import mars.dataframe as md
>>> from mars.config import options
>>> options.eager_mode = True
>>> t = mt.arange(6).reshape((2, 3))
>>> t
array([[0, 1, 2],
[3, 4, 5]])
>>> df = md.DataFrame(t)
>>> df.sum()
0 3
1 5
2 7
dtype: int64Mars also has deep integration with Ray and can run on Ray <https://docs.ray.io/en/latest/>_ efficiently and
interact with the large ecosystem of machine learning and distributed systems built on top of the core Ray.
Starting a new Mars on Ray runtime locally via:
.. code-block:: python
import mars
mars.new_session(backend='ray')
# Perform computationInteract with Ray Dataset:
.. code-block:: python
import mars.tensor as mt
import mars.dataframe as md
df = md.DataFrame(
mt.random.rand(1000_0000, 4),
columns=list('abcd'))
# Convert mars dataframe to ray dataset
ds = md.to_ray_dataset(df)
print(ds.schema(), ds.count())
ds.filter(lambda row: row["a"] > 0.5).show(5)
# Convert ray dataset to mars dataframe
df2 = md.read_ray_dataset(ds)
print(df2.head(5).execute())Refer to Mars on Ray_ for more information.
Mars can scale in to a single machine, and scale out to a cluster with thousands of machines. It's fairly simple to migrate from a single machine to a cluster to process more data or gain a better performance.
Bare Metal Deployment
Mars is easy to scale out to a cluster by starting different components of
mars distributed runtime on different machines in the cluster.
A node can be selected as supervisor which integrated a web service,
leaving other nodes as workers. The supervisor can be started with the following command:
.. code-block:: bash
mars-supervisor -h <host_name> -p <supervisor_port> -w <web_port>
Workers can be started with the following command:
.. code-block:: bash
mars-worker -h <host_name> -p <worker_port> -s <supervisor_endpoint>
After all mars processes are started, users can run
.. code-block:: python
>>> sess = new_session('http://<web_ip>:<ui_port>')
>>> # perform computation
Kubernetes DeploymentRefer to Run on Kubernetes_ for more information.
Yarn Deployment
Refer to `Run on Yarn`_ for more information.
Getting involved
----------------
- Read `development guide <https://mars-project.readthedocs.io/en/latest/development/index.html>`_.
- Join our Slack workgroup: `Slack <https://join.slack.com/t/mars-computing/shared_invite/zt-17pw2cfua-NRb2H4vrg77pr9T4g3nQOQ>`_.
- Join the mailing list: send an email to `mars-dev@googlegroups.com`_.
- Please report bugs by submitting a `GitHub issue`_.
- Submit contributions using `pull requests`_.
Thank you in advance for your contributions!
.. |Build| image:: https://github.com/mars-project/mars/workflows/Mars%20CI%20Core/badge.svg
:target: https://github.com/mars-project/mars/actions
.. |Coverage| image:: https://codecov.io/gh/mars-project/mars/branch/master/graph/badge.svg
:target: https://codecov.io/gh/mars-project/mars
.. |Quality| image:: https://img.shields.io/codacy/grade/6a80bb4659ed410eb33795f580c8615e.svg
:target: https://app.codacy.com/project/mars-project/mars/dashboard
.. |PyPI version| image:: https://img.shields.io/pypi/v/pymars.svg
:target: https://pypi.python.org/pypi/pymars
.. |Docs| image:: https://img.shields.io/badge/docs-latest-brightgreen.svg
:target: `Documentation`_
.. |License| image:: https://img.shields.io/pypi/l/pymars.svg
:target: https://github.com/mars-project/mars/blob/master/LICENSE
.. _`mars-dev@googlegroups.com`: https://groups.google.com/forum/#!forum/mars-dev
.. _`GitHub issue`: https://github.com/mars-project/mars/issues
.. _`pull requests`: https://github.com/mars-project/mars/pulls
.. _`Documentation`: https://mars-project.readthedocs.io
.. _`中文文档`: https://mars-project.readthedocs.io/zh_CN/latest/
.. _`Mars on Ray`: https://mars-project.readthedocs.io/en/latest/installation/ray.html
.. _`Run on Kubernetes`: https://mars-project.readthedocs.io/en/latest/installation/kubernetes.html
.. _`Run on Yarn`: https://mars-project.readthedocs.io/en/latest/installation/yarn.html
.. _`DASK on Mars`: https://mars-project.readthedocs.io/en/latest/user_guide/contrib/dask.html