.. Copyright (C) 2001-2021 NLTK Project
.. For license information, see LICENSE.TXT

===============================
Unit test cases for ``toolbox``
===============================

    >>> from nltk import toolbox

--------------------------
``toolbox.StandardFormat``
--------------------------

    >>> f = toolbox.StandardFormat()

``toolbox.StandardFormat.open()``
---------------------------------
    >>> import os, tempfile
    >>> (fd, fname) = tempfile.mkstemp()
    >>> tf = os.fdopen(fd, "w")
    >>> _ = tf.write('\\lx a value\n\\lx another value\n')
    >>> tf.close()
    >>> f = toolbox.StandardFormat()
    >>> f.open(fname)
    >>> list(f.fields())
    [('lx', 'a value'), ('lx', 'another value')]
    >>> f.close()
    >>> os.unlink(fname)

``toolbox.StandardFormat.open_string()``
----------------------------------------
    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\n\\lx another value\n')
    >>> list(f.fields())
    [('lx', 'a value'), ('lx', 'another value')]
    >>> f.close()

``toolbox.StandardFormat.close()``
----------------------------------
    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\n\\lx another value\n')
    >>> list(f.fields())
    [('lx', 'a value'), ('lx', 'another value')]
    >>> f.close()

``toolbox.StandardFormat.line_num``
---------------------------------------

``StandardFormat.line_num`` contains the line number of the last line returned:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\n\\lx another value\n\\lx a third value\n')
    >>> line_nums = []
    >>> for l in f.raw_fields():
    ...     line_nums.append(f.line_num)
    >>> line_nums
    [1, 2, 3]

``StandardFormat.line_num`` contains the line number of the last line returned:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx two\nlines\n\\lx three\nlines\n\n\\lx two\nlines\n')
    >>> line_nums = []
    >>> for l in f.raw_fields():
    ...     line_nums.append(f.line_num)
    >>> line_nums
    [2, 5, 7]

``StandardFormat.line_num`` doesn't exist before opening or after closing
a file or string:

    >>> f = toolbox.StandardFormat()
    >>> f.line_num
    Traceback (most recent call last):
        ...
    AttributeError: 'StandardFormat' object has no attribute 'line_num'
    >>> f.open_string('\\lx two\nlines\n\\lx three\nlines\n\n\\lx two\nlines\n')
    >>> line_nums = []
    >>> for l in f.raw_fields():
    ...     line_nums.append(f.line_num)
    >>> line_nums
    [2, 5, 7]
    >>> f.close()
    >>> f.line_num
    Traceback (most recent call last):
        ...
    AttributeError: 'StandardFormat' object has no attribute 'line_num'

``toolbox.StandardFormat.raw_fields()``
---------------------------------------
``raw_fields()`` returns an iterator over tuples of two strings representing the
marker and its value. The marker is given without the backslash and the value
without its trailing newline:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\n\\lx another value\n')
    >>> list(f.raw_fields())
    [('lx', 'a value'), ('lx', 'another value')]

an empty file returns nothing:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('')
    >>> list(f.raw_fields())
    []

file with only a newline returns WHAT SHOULD IT RETURN???:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\n')
    >>> list(f.raw_fields())
    [(None, '')]

file with only one field should be parsed ok:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx one value\n')
    >>> list(f.raw_fields())
    [('lx', 'one value')]

file without a trailing newline should be parsed ok:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\n\\lx another value')
    >>> list(f.raw_fields())
    [('lx', 'a value'), ('lx', 'another value')]

trailing white space is preserved except for the final newline:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx trailing space \n\\lx trailing tab\t\n\\lx extra newline\n\n')
    >>> list(f.raw_fields())
    [('lx', 'trailing space '), ('lx', 'trailing tab\t'), ('lx', 'extra newline\n')]

line wrapping is preserved:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\nmore of the value\nand still more\n\\lc another val\n')
    >>> list(f.raw_fields())
    [('lx', 'a value\nmore of the value\nand still more'), ('lc', 'another val')]

file beginning with a multiline record should be parsed ok:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\nmore of the value\nand still more\n\\lc another val\n')
    >>> list(f.raw_fields())
    [('lx', 'a value\nmore of the value\nand still more'), ('lc', 'another val')]

file ending with a multiline record should be parsed ok:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lc a value\n\\lx another value\nmore of the value\nand still more\n')
    >>> list(f.raw_fields())
    [('lc', 'a value'), ('lx', 'another value\nmore of the value\nand still more')]

file beginning with a BOM should be parsed ok:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\xef\xbb\xbf\\lx a value\n\\lx another value\n')
    >>> list(f.raw_fields())
    [('lx', 'a value'), ('lx', 'another value')]

file beginning with two BOMs should ignore only the first one:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\xef\xbb\xbf\xef\xbb\xbf\\lx a value\n\\lx another value\n')
    >>> list(f.raw_fields())
    [(None, '\xef\xbb\xbf\\lx a value'), ('lx', 'another value')]

should not ignore a BOM not at the beginning of the file:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\n\xef\xbb\xbf\\lx another value\n')
    >>> list(f.raw_fields())
    [('lx', 'a value\n\xef\xbb\xbf\\lx another value')]

``toolbox.StandardFormat.fields()``
-----------------------------------
trailing white space is not preserved:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx trailing space \n\\lx trailing tab\t\n\\lx extra newline\n\n')
    >>> list(f.fields())
    [('lx', 'trailing space'), ('lx', 'trailing tab'), ('lx', 'extra newline')]

multiline fields are unwrapped:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\lx a value\nmore of the value\nand still more\n\\lc another val\n')
    >>> list(f.fields())
    [('lx', 'a value more of the value and still more'), ('lc', 'another val')]

markers
-------
A backslash in the first position on a new line indicates the start of a
marker. The backslash is not part of the marker:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\mk a value\n')
    >>> list(f.fields())
    [('mk', 'a value')]

If the backslash occurs later in the line it does not indicate the start
of a marker:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\mk a value\n \\mk another one\n')
    >>> list(f.raw_fields())
    [('mk', 'a value\n \\mk another one')]

There is no specific limit to the length of a marker:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\this_is_an_extremely_long_marker value\n')
    >>> list(f.fields())
    [('this_is_an_extremely_long_marker', 'value')]

A marker can contain any non white space character:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\`~!@#$%^&*()_-=+[{]}\\|,<.>/?;:"0123456789 value\n')
    >>> list(f.fields())
    [('`~!@#$%^&*()_-=+[{]}\\|,<.>/?;:"0123456789', 'value')]

A marker is terminated by any white space character:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\mk a value\n\\mk\tanother one\n\\mk\rthird one\n\\mk\ffourth one')
    >>> list(f.fields())
    [('mk', 'a value'), ('mk', 'another one'), ('mk', 'third one'), ('mk', 'fourth one')]

Consecutive whitespace characters (except newline) are treated the same as one:

    >>> f = toolbox.StandardFormat()
    >>> f.open_string('\\mk \t\r\fa value\n')
    >>> list(f.fields())
    [('mk', 'a value')]

-----------------------
``toolbox.ToolboxData``
-----------------------

    >>> db = toolbox.ToolboxData()

``toolbox.ToolboxData.parse()``
-------------------------------
check that normal parsing works:

    >>> from xml.etree import ElementTree
    >>> td = toolbox.ToolboxData()
    >>> s = """\\_sh v3.0  400  Rotokas Dictionary
    ... \\_DateStampHasFourDigitYear
    ...
    ... \\lx kaa
    ... \\ps V.A
    ... \\ge gag
    ... \\gp nek i pas
    ...
    ... \\lx kaa
    ... \\ps V.B
    ... \\ge strangle
    ... \\gp pasim nek
    ... """
    >>> td.open_string(s)
    >>> tree = td.parse(key='lx')
    >>> tree.tag
    'toolbox_data'
    >>> ElementTree.tostring(list(tree)[0]).decode('utf8')
    '<header><_sh>v3.0  400  Rotokas Dictionary</_sh><_DateStampHasFourDigitYear /></header>'
    >>> ElementTree.tostring(list(tree)[1]).decode('utf8')
    '<record><lx>kaa</lx><ps>V.A</ps><ge>gag</ge><gp>nek i pas</gp></record>'
    >>> ElementTree.tostring(list(tree)[2]).decode('utf8')
    '<record><lx>kaa</lx><ps>V.B</ps><ge>strangle</ge><gp>pasim nek</gp></record>'

check that guessing the key marker works:

    >>> from xml.etree import ElementTree
    >>> td = toolbox.ToolboxData()
    >>> s = """\\_sh v3.0  400  Rotokas Dictionary
    ... \\_DateStampHasFourDigitYear
    ...
    ... \\lx kaa
    ... \\ps V.A
    ... \\ge gag
    ... \\gp nek i pas
    ...
    ... \\lx kaa
    ... \\ps V.B
    ... \\ge strangle
    ... \\gp pasim nek
    ... """
    >>> td.open_string(s)
    >>> tree = td.parse()
    >>> ElementTree.tostring(list(tree)[0]).decode('utf8')
    '<header><_sh>v3.0  400  Rotokas Dictionary</_sh><_DateStampHasFourDigitYear /></header>'
    >>> ElementTree.tostring(list(tree)[1]).decode('utf8')
    '<record><lx>kaa</lx><ps>V.A</ps><ge>gag</ge><gp>nek i pas</gp></record>'
    >>> ElementTree.tostring(list(tree)[2]).decode('utf8')
    '<record><lx>kaa</lx><ps>V.B</ps><ge>strangle</ge><gp>pasim nek</gp></record>'

-----------------------
``toolbox`` functions
-----------------------

``toolbox.to_sfm_string()``
-------------------------------