<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> <title>Contributing — scikits.learn v0.6.0 documentation</title> <link rel="stylesheet" href="../_static/nature.css" type="text/css" /> <link rel="stylesheet" href="../_static/pygments.css" type="text/css" /> <script type="text/javascript"> var DOCUMENTATION_OPTIONS = { URL_ROOT: '../', VERSION: '0.6.0', COLLAPSE_INDEX: false, FILE_SUFFIX: '.html', HAS_SOURCE: true }; </script> <script type="text/javascript" src="../_static/jquery.js"></script> <script type="text/javascript" src="../_static/underscore.js"></script> <script type="text/javascript" src="../_static/doctools.js"></script> <link rel="shortcut icon" href="../_static/favicon.ico"/> <link rel="author" title="About these documents" href="../about.html" /> <link rel="top" title="scikits.learn v0.6.0 documentation" href="../index.html" /> <link rel="next" title="<no title>" href="../performance.html" /> <link rel="prev" title="SVM: Weighted samples" href="../auto_examples/svm/plot_weighted_samples.html" /> </head> <body> <div class="header-wrapper"> <div class="header"> <p class="logo"><a href="../index.html"> <img src="../_static/scikit-learn-logo-small.png" alt="Logo"/> </a> </p><div class="navbar"> <ul> <li><a href="../install.html">Download</a></li> <li><a href="../support.html">Support</a></li> <li><a href="../user_guide.html">User Guide</a></li> <li><a href="../auto_examples/index.html">Examples</a></li> <li><a href="#">Development</a></li> </ul> <div class="search_form"> <div id="cse" style="width: 100%;"></div> <script src="http://www.google.com/jsapi" type="text/javascript"></script> <script type="text/javascript"> google.load('search', '1', {language : 'en'}); google.setOnLoadCallback(function() { var customSearchControl = new google.search.CustomSearchControl('016639176250731907682:tjtqbvtvij0'); customSearchControl.setResultSetSize(google.search.Search.FILTERED_CSE_RESULTSET); var options = new google.search.DrawOptions(); options.setAutoComplete(true); customSearchControl.draw('cse', options); }, true); </script> </div> </div> <!-- end navbar --></div> </div> <div class="content-wrapper"> <!-- <div id="blue_tile"></div> --> <div class="sphinxsidebar"> <div class="rel"> <a href="../auto_examples/svm/plot_weighted_samples.html" title="SVM: Weighted samples" accesskey="P">previous</a> | <a href="../performance.html" title="<no title>" accesskey="N">next</a> | <a href="../genindex.html" title="General Index" accesskey="I">index</a> </div> <h3>Contents</h3> <ul> <li><a class="reference internal" href="#">Contributing</a><ul> <li><a class="reference internal" href="#submitting-a-bug-report">Submitting a bug report</a></li> <li><a class="reference internal" href="#retrieving-the-latest-code">Retrieving the latest code</a></li> <li><a class="reference internal" href="#contributing-code">Contributing code</a><ul> <li><a class="reference internal" href="#how-to-contribute">How to contribute</a></li> <li><a class="reference internal" href="#easyfix-issues">EasyFix Issues</a></li> <li><a class="reference internal" href="#roadmap">Roadmap</a></li> <li><a class="reference internal" href="#documentation">Documentation</a></li> <li><a class="reference internal" href="#developers-web-site">Developers web site</a></li> </ul> </li> <li><a class="reference internal" href="#coding-guidelines">Coding guidelines</a></li> <li><a class="reference internal" href="#apis-of-scikit-learn-objects">APIs of scikit learn objects</a><ul> <li><a class="reference internal" href="#different-objects">Different objects</a></li> <li><a class="reference internal" href="#estimators">Estimators</a><ul> <li><a class="reference internal" href="#instantiation">Instantiation</a></li> <li><a class="reference internal" href="#fitting">Fitting</a></li> <li><a class="reference internal" href="#python-tuples">Python tuples</a></li> <li><a class="reference internal" href="#optional-arguments">Optional Arguments</a></li> </ul> </li> <li><a class="reference internal" href="#unresolved-api-issues">Unresolved API issues</a></li> <li><a class="reference internal" href="#specific-models">Specific models</a></li> </ul> </li> </ul> </li> </ul> </div> <div class="content"> <div class="documentwrapper"> <div class="bodywrapper"> <div class="body"> <div class="section" id="contributing"> <h1>Contributing<a class="headerlink" href="#contributing" title="Permalink to this headline">¶</a></h1> <p>This project is a community effort, and everyone is welcomed to contribute.</p> <div class="section" id="submitting-a-bug-report"> <h2>Submitting a bug report<a class="headerlink" href="#submitting-a-bug-report" title="Permalink to this headline">¶</a></h2> <p>In case you experience difficulties using the package, do not hesitate to submit a ticket to the <a class="reference external" href="http://sourceforge.net/apps/trac/scikit-learn/report/1">Bug Tracker</a>.</p> <p>You are also welcomed to post there feature requests and patches.</p> </div> <div class="section" id="retrieving-the-latest-code"> <span id="git-repo"></span><h2>Retrieving the latest code<a class="headerlink" href="#retrieving-the-latest-code" title="Permalink to this headline">¶</a></h2> <p>You can check the latest sources with the command:</p> <div class="highlight-python"><pre>git clone git://github.com/scikit-learn/scikit-learn.git</pre> </div> <p>or if you have write privileges:</p> <div class="highlight-python"><pre>git clone git@github.com:scikit-learn/scikit-learn.git</pre> </div> <p>You can also check out the sources online in the web page <a class="reference external" href="http://github.com/scikit-learn/scikit-learn">http://github.com/scikit-learn/scikit-learn</a></p> <p>If you run the development version, it is cumbersome to re-install the package each time you update the sources. It is thus preferred that you add the scikit-directory to your PYTHONPATH and build the extension in place:</p> <div class="highlight-python"><pre>python setup.py build_ext --inplace</pre> </div> </div> <div class="section" id="contributing-code"> <h2>Contributing code<a class="headerlink" href="#contributing-code" title="Permalink to this headline">¶</a></h2> <div class="section" id="how-to-contribute"> <h3>How to contribute<a class="headerlink" href="#how-to-contribute" title="Permalink to this headline">¶</a></h3> <p>The prefered way to contribute to <cite>scikit-learn</cite> is to fork the main repository on <a class="reference external" href="http://github.com/scikit-learn/scikit-learn/">github</a>:</p> <blockquote> <ol class="arabic"> <li><p class="first"><a class="reference external" href="https://github.com/signup/free">Create an account</a> on github if you don’t have one already.</p> </li> <li><p class="first">Fork the <a class="reference external" href="http://github.com/scikit-learn/scikit-learn">scikit-learn repo</a>: click on the ‘Fork’ button, at the top, center of the page. This creates a copy of the code on the github server where you can work.</p> </li> <li><p class="first">Clone this copy to your local disk (you need the <cite>git</cite> program to do this):</p> <div class="highlight-python"><pre>$ git clone git@github.com:YourLogin/scikit-learn.git</pre> </div> </li> <li><p class="first">Work on this copy, on your computer, using git to do the version control:</p> <div class="highlight-python"><pre>$ git add modified_files $ git commit $ git push origin master</pre> </div> <p>and so on.</p> </li> </ol> </blockquote> <p>When you are ready, and you have pushed your changes on your github repo, go the web page of the repo, and click on ‘Pull request’ to send us a pull request. Send us a mail with your pull request, and we can look at your changes, and integrate them.</p> <p><strong>Before asking for a pull or a review</strong>, be sure to read the <a class="reference internal" href="#coding-guidelines">coding-guidelines</a> (below).</p> <p>Also, make sure that your code is tested, and that all the tests for the scikit pass.</p> </div> <div class="section" id="easyfix-issues"> <h3>EasyFix Issues<a class="headerlink" href="#easyfix-issues" title="Permalink to this headline">¶</a></h3> <p>The best way to get your feet wet is to pick up an issue from the <a class="reference external" href="https://sourceforge.net/apps/trac/scikit-learn/report">issue tracker</a> that are labeled as EasyFix. This means that the knowledge needed to solve the issue is low, but still you are helping the project and letting more experienced developers concentrate on other issues.</p> </div> <div class="section" id="roadmap"> <h3>Roadmap<a class="headerlink" href="#roadmap" title="Permalink to this headline">¶</a></h3> <p><a class="reference external" href="http://sourceforge.net/apps/trac/scikit-learn/roadmap">Here</a> you will find a detailed roadmap, with a description on what’s planned to be implemented in the following releases.</p> </div> <div class="section" id="documentation"> <h3>Documentation<a class="headerlink" href="#documentation" title="Permalink to this headline">¶</a></h3> <p>We are glad to accept any sort of documentation: function docstrings, rst docs (like this one), tutorials, etc. Rst docs live in the source code repository, under directory doc/.</p> <p>You can edit them using any text editor and generate the html docs by typing from the doc/ directory <tt class="docutils literal"><span class="pre">make</span> <span class="pre">html</span></tt> (or <tt class="docutils literal"><span class="pre">make</span> <span class="pre">html-noplot</span></tt>, see README in that directory for more info). That should create a directory _build/html/ with html files that are viewable in a web browser.</p> </div> <div class="section" id="developers-web-site"> <h3>Developers web site<a class="headerlink" href="#developers-web-site" title="Permalink to this headline">¶</a></h3> <p>More information can be found at the <a class="reference external" href="https://github.com/scikit-learn/scikit-learn/wiki">developer’s wiki</a>.</p> </div> </div> <div class="section" id="coding-guidelines"> <span id="id1"></span><h2>Coding guidelines<a class="headerlink" href="#coding-guidelines" title="Permalink to this headline">¶</a></h2> <p>The following are some guidelines on how new code should be written. Of course, there are special cases and there will be exceptions to these rules. However, following these rules when submitting new code makes the review easier so new code can be integrated in less time.</p> <p>Uniformly formated code makes it easier to share code ownership. The scikit learn tries to follow closely the officiel Python guidelines detailed in <a class="reference external" href="http://www.python.org/dev/peps/pep-0008/">PEP8</a> that details how code should be formatted, and indented. Please read it and follow it.</p> <p>In addition, we add the following guidelines:</p> <blockquote> <ul class="simple"> <li>Use underscores to separate words in non class names: <cite>n_samples</cite> rather than <cite>nsamples</cite>.</li> <li>Avoid multiple statements on one line. Prefer a line return after a control flow statement (<cite>if</cite>/<cite>for</cite>).</li> <li>Use relative imports for references inside scikits.learn.</li> <li><strong>Please don’t use `import *` in any case</strong>. It is considered harmful by the <a class="reference external" href="http://docs.python.org/howto/doanddont.html#from-module-import">official Python recommandations</a>. It makes the code harder to read as the origin of symbols is no longer explicitely referenced, but most important, it prevents using a static analysis tool like <a class="reference external" href="http://www.divmod.org/trac/wiki/DivmodPyflakes">pyflakes</a> to automatically find bugs in the scikit.</li> </ul> </blockquote> <p>A good example of code that we like can be found <a class="reference external" href="https://svn.enthought.com/enthought/browser/sandbox/docs/coding_standard.py">here</a>.</p> </div> <div class="section" id="apis-of-scikit-learn-objects"> <h2>APIs of scikit learn objects<a class="headerlink" href="#apis-of-scikit-learn-objects" title="Permalink to this headline">¶</a></h2> <p>To have a uniform API, we try to have a common basic API for all the objects. In addition, to avoid the proliferation of framework code, we try to adopt simple conventions and limit to a minimum the number of methods an object has to implement.</p> <div class="section" id="different-objects"> <h3>Different objects<a class="headerlink" href="#different-objects" title="Permalink to this headline">¶</a></h3> <p>The main objects of the scikit learn are (one class can implement multiple interfaces):</p> <table class="docutils field-list" frame="void" rules="none"> <col class="field-name" /> <col class="field-body" /> <tbody valign="top"> <tr class="field"><th class="field-name">Estimator:</th><td class="field-body"><p class="first">The base object, implements:</p> <div class="highlight-python"><div class="highlight"><pre><span class="n">estimator</span> <span class="o">=</span> <span class="n">obj</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">data</span><span class="p">)</span> </pre></div> </div> </td> </tr> <tr class="field"><th class="field-name">Predictor:</th><td class="field-body"><p class="first">For suppervised learning, or some unsupervised problems, implements:</p> <div class="highlight-python"><div class="highlight"><pre><span class="n">prediction</span> <span class="o">=</span> <span class="n">obj</span><span class="o">.</span><span class="n">predict</span><span class="p">(</span><span class="n">data</span><span class="p">)</span> </pre></div> </div> </td> </tr> <tr class="field"><th class="field-name">Transformer:</th><td class="field-body"><p class="first">For filtering or modifying the data, in a supervised or unsupervised way, implements:</p> <div class="highlight-python"><div class="highlight"><pre><span class="n">new_data</span> <span class="o">=</span> <span class="n">obj</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">data</span><span class="p">)</span> </pre></div> </div> <p>When fitting and transforming can be performed much more efficiently together than separately, implements:</p> <p>new_data = obj.fit_transform(data)</p> </td> </tr> <tr class="field"><th class="field-name">Model:</th><td class="field-body"><p class="first">A model that can give a goodness of fit or a likelihood of unseen data, implements (higher is better):</p> <div class="last highlight-python"><div class="highlight"><pre><span class="n">score</span> <span class="o">=</span> <span class="n">obj</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">data</span><span class="p">)</span> </pre></div> </div> </td> </tr> </tbody> </table> </div> <div class="section" id="estimators"> <h3>Estimators<a class="headerlink" href="#estimators" title="Permalink to this headline">¶</a></h3> <p>The API has one predominant object: the estimator. A estimator is an object that fits a model based on some training data and is capable of inferring some properties on new data. It can be for instance a classifier or a regressor. All estimators implement the fit method:</p> <div class="highlight-python"><div class="highlight"><pre><span class="n">estimator</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span> </pre></div> </div> <div class="section" id="instantiation"> <h4>Instantiation<a class="headerlink" href="#instantiation" title="Permalink to this headline">¶</a></h4> <p>This concerns the object creation. The object’s __init__ method might accept as arguments constants that determine the estimator behavior (like the C constant in SVMs).</p> <p>It should not, however, take the actual training data as argument, as this is left to the <tt class="docutils literal"><span class="pre">fit()</span></tt> method:</p> <div class="highlight-python"><div class="highlight"><pre><span class="n">clf2</span> <span class="o">=</span> <span class="n">SVC</span><span class="p">(</span><span class="n">C</span><span class="o">=</span><span class="mf">2.3</span><span class="p">)</span> <span class="n">clf3</span> <span class="o">=</span> <span class="n">SVC</span><span class="p">([[</span><span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">],</span> <span class="p">[</span><span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">]],</span> <span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">])</span> <span class="c"># WRONG!</span> </pre></div> </div> <p>The arguments that go in the <cite>__init__</cite> should all be keyword arguments with a default value. In other words, a user should be able to instanciate an estimator without passing to it any arguments.</p> <p>The arguments in given at instanciation of an estimator should all correspond to hyper parameters describing the model or the optimisation problem that estimator tries to solve. They should however not be parameters of the estimation routine: these are passed directly to the <cite>fit</cite> method.</p> <p>In addition, <strong>every keyword argument given to the `__init__` should correspond to an attribute on the instance</strong>. The scikit relies on this to find what are the relevent attributes to set on an estimator when doing model selection.</p> <p>All estimators should inherit from <cite>scikit.learn.base.BaseEstimator</cite></p> </div> <div class="section" id="fitting"> <h4>Fitting<a class="headerlink" href="#fitting" title="Permalink to this headline">¶</a></h4> <p>The next thing you’ll probably want to do is to estimate some parameters in the model. This is implemented in the .fit() method.</p> <p>The fit method takes as argument the training data, which can be one array in the case of unsupervised learning, or two arrays in the case of supervised learning.</p> <p>Note that the model is fitted using X and y but the object holds no reference to X, y. There are however some exceptions to this, as in the case of precomputed kernels where you need to store access these data in the predict method.</p> <blockquote> <p>Parameters</p> <blockquote> <ul class="simple"> <li>X : array-like, with shape = [N, D], where N is the number of samples and D is the number of features.</li> <li>Y : array, with shape = [N], where N is the number of samples.</li> <li>args, kwargs. Parameters can also be set in the fit method.</li> </ul> </blockquote> </blockquote> <p>X.shape[0] should be the same as Y.shape[0]. If this requisite is not met, an exception should be raised.</p> <p>Y might be dropped in the case of unsupervised learning.</p> <p>The method should return the object (self).</p> </div> <div class="section" id="python-tuples"> <h4>Python tuples<a class="headerlink" href="#python-tuples" title="Permalink to this headline">¶</a></h4> <p>In addition to numpy arrays, all methods should be able to accept python tuples as arguments. In practice, this means you should call numpy.asanyarray at the beginning at each public method that accepts arrays.</p> </div> <div class="section" id="optional-arguments"> <h4>Optional Arguments<a class="headerlink" href="#optional-arguments" title="Permalink to this headline">¶</a></h4> <p>In iterative algorithms, number of iterations should be specified by an int called <tt class="docutils literal"><span class="pre">n_iter</span></tt>.</p> </div> </div> <div class="section" id="unresolved-api-issues"> <h3>Unresolved API issues<a class="headerlink" href="#unresolved-api-issues" title="Permalink to this headline">¶</a></h3> <p>Some things are must still be decided:</p> <blockquote> <ul class="simple"> <li>what should happen when predict is called before than fit() ?</li> <li>which exception should be raised when arrays’ shape do not match in fit() ?</li> </ul> </blockquote> </div> <div class="section" id="specific-models"> <h3>Specific models<a class="headerlink" href="#specific-models" title="Permalink to this headline">¶</a></h3> <p>In linear models, coefficients are stored in an array called <tt class="docutils literal"><span class="pre">coef_</span></tt>, and independent term is stored in <tt class="docutils literal"><span class="pre">intercept_</span></tt>.</p> </div> </div> </div> </div> </div> </div> <div class="clearer"></div> </div> </div> <div class="footer"> <p style="text-align: center">This documentation is relative to scikits.learn version 0.6.0<p> © 2010, scikits.learn developers (BSD Lincense). Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 1.0.5. Design by <a href="http://webylimonada.com">Web y Limonada</a>. </div> </body> </html>