Add a debugging exercise

Author: mstimberg

07-debugging-post-mortem.html

@@ -30,7 +30,7 @@
           <h2 class="subtitle">Post-mortem debugging</h2>
           <section class="objectives panel panel-warning">
 <div class="panel-heading">
-<h2><span class="glyphicon glyphicon-certificate"></span>Learning Objectives</h2>
+<h2 id="learning-objectives"><span class="glyphicon glyphicon-certificate"></span>Learning Objectives</h2>
 </div>
 <div class="panel-body">
 <ul>
@@ -61,7 +61,7 @@ <h2><span class="glyphicon glyphicon-certificate"></span>Learning Objectives</h2
     <span class="kw">return</span> -<span class="dv">1</span>
 
 <span class="kw">def</span> check_data(target):
-    test_data = [<span class="dv">3</span>, <span class="dv">2</span>, <span class="dv">8</span>, <span class="dv">9</span>, <span class="dv">3</span>, <span class="dv">4</span>, <span class="dv">7</span>, <span class="dv">5</span>]
+    test_data = [<span class="dv">3</span>, <span class="dv">2</span>, <span class="dv">8</span>, <span class="dv">9</span>, <span class="dv">3</span>, numpy.nan, <span class="dv">4</span>, <span class="dv">7</span>, <span class="dv">5</span>]
     <span class="co"># We look for a zero in the data</span>
     index = find_first(test_data, target)
     <span class="kw">if</span> index != -<span class="dv">1</span>:
@@ -71,7 +71,7 @@ <h2><span class="glyphicon glyphicon-certificate"></span>Learning Objectives</h2
 <p>The <code>find_first</code> function finds the first occurrence of an element in a list (or an array) of elements and returns its index. The <code>check_data</code> function uses this function to find a certain value in the data and prints the data until that point. It uses some fake “data” and takes an <code>target</code> argument to specify what value to look for.</p>
 <p>If we run the <code>check_data</code> function, all looks fine:</p>
 <pre class="sourceCode python"><code class="sourceCode python">check_data(<span class="dv">5</span>)</code></pre>
-<pre class="output"><code>Data until first occurrence of 5 : [3, 2, 8, 9, 3, 4, 7]</code></pre>
+<pre class="output"><code>Data until first occurrence of 5 : [3, 2, 8, 9, 3, nan, 4, 7]</code></pre>
 <p>Now let’s run the function again but use a different target value:</p>
 <pre class="sourceCode python"><code class="sourceCode python">check_data(<span class="dv">1</span>)</code></pre>
 <pre class="output"><code>---------------------------------------------------------------------------
@@ -80,7 +80,7 @@ <h2><span class="glyphicon glyphicon-certificate"></span>Learning Objectives</h2
 ----&gt; 1 check_data(1)
 
 &lt;ipython-input-56-610c1db03285&gt; in check_data(target)
-     16     test_data = [3, 2, 8, 9, 3, 4, 7, 5]
+     16     test_data = [3, 2, 8, 9, 3, numpy.nan, 4, 7, 5]
      17     # We look for a zero in the data
 ---&gt; 18     index = find_first(test_data, target)
      19     if index != -1:
@@ -118,13 +118,13 @@ <h2><span class="glyphicon glyphicon-certificate"></span>Learning Objectives</h2
 <p>To access the data, we can move “up” in the exception stack and investigate the variables there:</p>
 <pre class="output"><code>
ipdb&gt; 
up
 &gt; &lt;ipython-input-56-610c1db03285&gt;(18)check_data()
-     16     test_data = [3, 2, 8, 9, 3, 4, 7, 5]
+     16     test_data = [3, 2, 8, 9, 3, numpy.nan, 4, 7, 5]
      17     # We look for a zero in the data
 ---&gt; 18     index = find_first(test_data, target)
      19     if index != -1:
      20         print(&#39;Data until first occurrence of&#39;, target, &#39;:&#39;, test_data[:index])
 
ipdb&gt; 
test_data
-     [3, 2, 8, 9, 3, 4, 7, 5]</code></pre>
+     [3, 2, 8, 9, 3, nan, 4, 7, 5]</code></pre>
 <p>To get back to where we were before, we use <code>down</code> or alternatively <code>d</code> (instead of <code>up</code> we can also use <code>u</code>):</p>
 <pre class="output"><code>ipdb&gt; d
 &gt; &lt;ipython-input-56-610c1db03285&gt;(10)find_first()
@@ -149,7 +149,7 @@ <h2><span class="glyphicon glyphicon-certificate"></span>Learning Objectives</h2
 <p>Using a counter and indexing into a list at every iteration of a loop is a often considered to be a so-called “<a href="https://en.wikipedia.org/wiki/Anti-pattern">anti-pattern</a>” in Python.</p>
 <aside class="callout panel panel-info">
 <div class="panel-heading">
-<h2><span class="glyphicon glyphicon-pushpin"></span>Post-mortem debugging with pytest</h2>
+<h2 id="post-mortem-debugging-with-pytest"><span class="glyphicon glyphicon-pushpin"></span>Post-mortem debugging with pytest</h2>
 </div>
 <div class="panel-body">
 <p>You can run your test suite with pytest and make pytest open a debugger session for you as soon as an error occurs:</p>
@@ -181,7 +181,33 @@ <h2><span class="glyphicon glyphicon-pushpin"></span>Post-mortem debugging with
 (Pdb)</code></pre>
 </div>
 </aside>
-<p>TODO: Exercise</p>
+<section class="challenge panel panel-success">
+<div class="panel-heading">
+<h2 id="error-debugging"><span class="glyphicon glyphicon-pencil"></span>Error debugging</h2>
+</div>
+<div class="panel-body">
+<p>Copy &amp; paste the following code and run <code>test_code_array</code>. It will fail with an error, use the debugger to find out why.</p>
+<pre class="sourceCode python"><code class="sourceCode python"><span class="ch">import</span> numpy
+<span class="ch">from</span> numpy.testing.utils <span class="ch">import</span> assert_equal
+
+<span class="kw">def</span> code_array(codes):
+    <span class="co">&#39;&#39;&#39;Store airport codes in an array in numerical form&#39;&#39;&#39;</span>
+    to_numerical_code = numpy.vectorize(<span class="kw">lambda</span> code: numpy.array([<span class="dt">ord</span>(c) <span class="kw">for</span> c in code]))
+    <span class="kw">return</span> numpy.vstack([to_numerical_code(c) <span class="kw">for</span> c in codes])
+
+<span class="kw">def</span> test_code_array():
+    airport_codes = [<span class="st">&#39;TXL&#39;</span>, <span class="st">&#39;LAX&#39;</span>, <span class="st">&#39;PHX&#39;</span>, <span class="st">&#39;CDG&#39;</span> <span class="st">&#39;ORY&#39;</span>, <span class="st">&#39;JFK&#39;</span>, <span class="st">&#39;JNB&#39;</span>, <span class="st">&#39;WAW&#39;</span>]
+    expected = numpy.array([[<span class="dv">84</span>, <span class="dv">88</span>, <span class="dv">76</span>],
+                            [<span class="dv">76</span>, <span class="dv">65</span>, <span class="dv">88</span>],
+                            [<span class="dv">80</span>, <span class="dv">72</span>, <span class="dv">88</span>],
+                            [<span class="dv">67</span>, <span class="dv">68</span>, <span class="dv">71</span>],
+                            [<span class="dv">79</span>, <span class="dv">82</span>, <span class="dv">89</span>],
+                            [<span class="dv">74</span>, <span class="dv">70</span>, <span class="dv">75</span>],
+                            [<span class="dv">74</span>, <span class="dv">78</span>, <span class="dv">66</span>],
+                            [<span class="dv">87</span>, <span class="dv">65</span>, <span class="dv">87</span>]])
+    assert_equal(code_array(airport_codes), expected)</code></pre>
+</div>
+</section>
         </div>
       </div>
       </article>

07-debugging-post-mortem.md

@@ -50,7 +50,7 @@ def find_first(data, element):
     return -1
 
 def check_data(target):
-    test_data = [3, 2, 8, 9, 3, 4, 7, 5]
+    test_data = [3, 2, 8, 9, 3, numpy.nan, 4, 7, 5]
     # We look for a zero in the data
     index = find_first(test_data, target)
     if index != -1:
@@ -71,7 +71,7 @@ If we run the `check_data` function, all looks fine:
 check_data(5)
 ~~~
 ~~~ {.output}
-Data until first occurrence of 5 : [3, 2, 8, 9, 3, 4, 7]
+Data until first occurrence of 5 : [3, 2, 8, 9, 3, nan, 4, 7]
 ~~~
 
 Now let's run the function again but use a different target value:
@@ -86,7 +86,7 @@ IndexError                                Traceback (most recent call last)
 ----> 1 check_data(1)
 
 <ipython-input-56-610c1db03285> in check_data(target)
-     16     test_data = [3, 2, 8, 9, 3, 4, 7, 5]
+     16     test_data = [3, 2, 8, 9, 3, numpy.nan, 4, 7, 5]
      17     # We look for a zero in the data
 ---> 18     index = find_first(test_data, target)
      19     if index != -1:
@@ -159,13 +159,13 @@ variables there:
 ~~~ {.output}
 
ipdb> 
up
 > <ipython-input-56-610c1db03285>(18)check_data()
-     16     test_data = [3, 2, 8, 9, 3, 4, 7, 5]
+     16     test_data = [3, 2, 8, 9, 3, numpy.nan, 4, 7, 5]
      17     # We look for a zero in the data
 ---> 18     index = find_first(test_data, target)
      19     if index != -1:
      20         print('Data until first occurrence of', target, ':', test_data[:index])
 
ipdb> 
test_data
-     [3, 2, 8, 9, 3, 4, 7, 5]
+     [3, 2, 8, 9, 3, nan, 4, 7, 5]
 ~~~
 
 To get back to where we were before, we use `down` or alternatively `d`
@@ -250,4 +250,28 @@ often considered to be a so-called
 > (Pdb)
 > ~~~
 
-TODO: Exercise
+> ## Error debugging {.challenge}
+> Copy & paste the following code and run `test_code_array`. It will fail with
+> an error, use the debugger to find out why.
+>
+> ~~~ {.python}
+> import numpy
+> from numpy.testing.utils import assert_equal
+>
+> def code_array(codes):
+>     '''Store airport codes in an array in numerical form'''
+>     to_numerical_code = numpy.vectorize(lambda code: numpy.array([ord(c) for c in code]))
+>     return numpy.vstack([to_numerical_code(c) for c in codes])
+>
+> def test_code_array():
+>     airport_codes = ['TXL', 'LAX', 'PHX', 'CDG' 'ORY', 'JFK', 'JNB', 'WAW']
+>     expected = numpy.array([[84, 88, 76],
+>                             [76, 65, 88],
+>                             [80, 72, 88],
+>                             [67, 68, 71],
+>                             [79, 82, 89],
+>                             [74, 70, 75],
+>                             [74, 78, 66],
+>                             [87, 65, 87]])
+>     assert_equal(code_array(airport_codes), expected)
+> ~~~