Python doctest：多行结果

Question

我无法让 doctest 处理包含多行且开头可能包含空行的结果。这可能是由缩进和解析问题引起的。我已经想出了一些解决方案：

• 将所需的结果写入文件，并doctest结果与文件内容之间的比较。
• 将结果的哈希值与已知的哈希值进行比较。这种方法的主要缺点是，doctest 的读者对所需结果知之甚少。
• 找到一种方法使 doctest 能够处理多行结果。
• 使用 unittest 而不是 doctest。

有什么想法吗？

代码：

    >>> data_lists=[ {"Average execution" : [1, 2, 3, 2, 3]},
    ...                              {"Top execution"     : [3, 4, 5, 7, 8, 11, 6]},
    ...                              {"Current execution" : [1, 2, 1, 2, 1, 5]}       ]
    >>> c=Chart(data_lists,
    ...         ("Large<br>rooster", "Rsync rooster", "Pool<br>Chicken", "Disease"),
    ...         width=700, height=300)
    >>> print c.html.strip()
    <div id="placeholder3" style="width:700px;height:300px"></div>

    <script id="source" language="javascript" type="text/javascript">
    $(function () {

    var d0 = [[0, 1], [4, 2], [8, 3], [12, 2], [16, 3]];
    var d1 = [[1, 3], [5, 4], [9, 5], [13, 7], [17, 8], [21, 11], [25, 6]];
    var d2 = [[2, 1], [6, 2], [10, 1], [14, 2], [18, 1], [22, 5]];

        $.plot($("#placeholder3"), [

        {   label: "Average execution",  data: d0,   bars: { show: true }  },
        {   label: "Top execution",  data: d1,   bars: { show: true }  },
        {   label: "Current execution",  data: d2,   bars: { show: true }  }

        ],
        {
            xaxis: { ticks: [[1.5, "Large<br>rooster"], [5.5, "Rsync<br>rooster"], [9.5, "Pool<br>Chicken"], [13.5, "Disease"]] }
        }
        );
    });
    </script>

错误：

**********************************************************************
File "HTML.py", line 28, in __main__.Chart.__init__
Failed example:
    print c.html.strip()
Expected:
    <div id="placeholder3" style="width:700px;height:300px"></div>
Got:
    <div id="placeholder3" style="width:700px;height:300px"></div>
    <BLANKLINE>
        <script id="source" language="javascript" type="text/javascript">
        $(function () {
    <BLANKLINE>
        var d0 = [[0, 1], [4, 2], [8, 3], [12, 2], [16, 3]];
        var d1 = [[1, 3], [5, 4], [9, 5], [13, 7], [17, 8], [21, 11], [25, 6]];
        var d2 = [[2, 1], [6, 2], [10, 1], [14, 2], [18, 1], [22, 5]];
    <BLANKLINE>
            $.plot($("#placeholder3"), [
    <BLANKLINE>
            {   label: "Average execution",  data: d0,   bars: { show: true }  },
            {   label: "Top execution",  data: d1,   bars: { show: true }  },
            {   label: "Current execution",  data: d2,   bars: { show: true }  }
    <BLANKLINE>
            ],
            {
                xaxis: { ticks: [[1.5, "Large<br>rooster"], [5.5, "Rsync rooster"], [9.5, "Pool<br>Chicken"], [13.5, "Disease"]] }
            }
            );
        });
        </script>
**********************************************************************
1 items had failures:
   1 of   3 in __main__.Chart.

__init__
***Test Failed*** 1 failures.

Answer 1

将 放入预期输出中，就像错误消息中显示的那样。那么测试应该可以正常工作。预期的输入在第一个仅空白的行处终止，这就是为什么您必须专门标记它：

>>> data_lists=[ {"Average execution" : [1, 2, 3, 2, 3]},
...                              {"Top execution"     : [3, 4, 5, 7, 8, 11, 6]},
...                              {"Current execution" : [1, 2, 1, 2, 1, 5]}       ]
>>> c=Chart(data_lists,
...         ("Large<br>rooster", "Rsync rooster", "Pool<br>Chicken", "Disease"),
...         width=700, height=300)
>>> print c.html.strip()
<div id="placeholder3" style="width:700px;height:300px"></div>
<BLANKLINE>
<script id="source" language="javascript" type="text/javascript">
$(function () {
<BLANKLINE>
var d0 = [[0, 1], [4, 2], [8, 3], [12, 2], [16, 3]];
var d1 = [[1, 3], [5, 4], [9, 5], [13, 7], [17, 8], [21, 11], [25, 6]];
var d2 = [[2, 1], [6, 2], [10, 1], [14, 2], [18, 1], [22, 5]];
<BLANKLINE>
    $.plot($("#placeholder3"), [
<BLANKLINE>
    {   label: "Average execution",  data: d0,   bars: { show: true }  },
    {   label: "Top execution",  data: d1,   bars: { show: true }  },
    {   label: "Current execution",  data: d2,   bars: { show: true }  }
<BLANKLINE>
    ],
    {
        xaxis: { ticks: [[1.5, "Large<br>rooster"], [5.5, "Rsync<br>rooster"], [9.5, "Pool<br>Chicken"], [13.5, "Disease"]] }
    }
    );
});
</script>

请参阅 doctest 文档，其中解释了这一点： http://docs.python.org/library/doctest.html#how-are-docstring-examples-recognized

Answer 2

使用 指示输出中的空行，就像失败输出中的 ;-)

Answer 3

除了按照其他答案中的建议使用 之外，此选项也很有用：

>>> print c.html.strip()  # doctest: +NORMALIZE_WHITESPACE

特别是如果预期输出包含 \r （回车符）字符，在这种情况下如果没有设置 NORMALIZE_WHITESPACE， 将不会匹配。