Skip to content

docconverter

Convert doc files to the Giella xml format.

DocError

Bases: Exception

Use this when errors occur in this module.

Source code in /home/anders/projects/CorpusTools/corpustools/docconverter.py 
28
29
class DocError(Exception):
    """Use this when errors occur in this module."""

doc_to_unicodehtml(filename)

Convert a doc file to xhtml.

Parameters:

Name Type Description Default
filename str

path to the file

 required

Returns:

Type Description
str

A string containing the xhtml version of the doc file.
Source code in /home/anders/projects/CorpusTools/corpustools/docconverter.py 
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
def doc_to_unicodehtml(filename):
    """Convert a doc file to xhtml.

    Args:
        filename (str): path to the file

    Returns:
        (str): A string containing the xhtml version of the doc file.
    """
    text = extract_text(filename)
    try:
        return text.decode("utf8")
    except UnicodeDecodeError:
        # remove control characters
        remove_re = re.compile("[\x00-\x08\x0B-\x0C\x0E-\x1F\x7F{}]")

        return remove_re.subn("", text.decode("windows-1252"))[0]

extract_text(filename)

Extract the text from a document.

Parameters:

Name Type Description Default
filename str

path to the document

 required

Returns:

Type Description
bytes

the output of the program
Source code in /home/anders/projects/CorpusTools/corpustools/docconverter.py 
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
def extract_text(filename):
    """Extract the text from a document.

    Args:
        filename (str): path to the document

    Returns:
        (bytes): the output of the program
    """
    command = ["wvHtml", filename, "-"]
    runner = util.ExternalCommandRunner()
    runner.run(command, cwd="/tmp")

    if runner.returncode != 0:
        with open(filename + ".log", "w") as logfile:
            print(f"stdout\n{runner.stdout}\n", file=logfile)
            print(f"stderr\n{runner.stderr}\n", file=logfile)
            raise util.ConversionError(
                "{} failed. More info in the log file: {}".format(
                    command[0], filename + ".log"
                )
            )

    return runner.stdout

fix_wv_output()

Fix headers in the docx xhtml output.

Examples of headings:

h1: OVTTASKASOLBMOT

h2: čoahkkáigeassu

Ulbmil ja váldooasit

h3: Geaográfalaš ráddjen 

                </html:p>
            </html:div>
        </html:li>

    </html:ol>
</html:ol>

Iskanjoavku ja sámegielaga definišuvdn a 

    </html:ol>
</html:ol>

h4: Mildosat:

Source code in /home/anders/projects/CorpusTools/corpustools/docconverter.py 
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
def fix_wv_output():
    """Fix headers in the docx xhtml output.

    Examples of headings:

    h1:
    <html:ul>
        <html:li value="2">
            <html:p/>
            <html:div align="left" name="Overskrift 1">
                <html:p>
                    <html:b>
                        <html:span>
                            OVTTASKASOLBMOT
                        </html:span>
                    </html:b>
                </html:p>
            </html:div>
        </html:li>
    </html:ul>

    h2:
    <html:ol type="1">
        <html:li value="1">
            <html:p/>
            <html:div align="left" name="Overskrift 2">
                <html:p>
                    <html:b>
                        <html:span>
                            čoahkkáigeassu
                        </html:span>
                    </html:b>
                </html:p>
            </html:div>
        </html:li>
    </html:ol>

    <html:ol type="1">
        <html:ol type="1">
            <html:li value="2">
                <html:p/>
                <html:div align="left" name="Overskrift 2">
                    <html:p>
                        <html:b>
                            <html:span>
                                Ulbmil ja váldooasit
                            </html:span>
                        </html:b>
                    </html:p>
                </html:div>
            </html:li>
        </html:ol>
    </html:ol>

    h3:
    <html:ol type="1">
        <html:ol type="1">
            <html:ol type="1">
                <html:li value="1">
                    <html:p>
                    </html:p>
                    <html:div align="left" name="Overskrift 3">
                        <html:p>
                            <html:b>
                                <html:span>
                                    Geaográfalaš
                                </html:span>
                            </html:b>
                            <html:b>
                                <html:span>
                                    ráddjen
                                </html:span>
                            </html:b>

                        </html:p>
                    </html:div>
                </html:li>

            </html:ol>
        </html:ol>
    </html:ol>

    <html:ol type="1">
        <html:ol type="1">
            <html:ol type="1">
                <html:li value="1">
                    <html:p>
                    </html:p>
                    <html:div align="left" name="Overskrift 3">
                        <html:p>
                            <html:b>
                            <html:span>Iskanjoavku ja sámegielaga
                            definišuvdn</html:span></html:b>
                            <html:b><html:span>a</html:span></html:b>
                        </html:p>
                    </html:div>
                </html:li>

            </html:ol>
        </html:ol>
    </html:ol>

    h4:
    <html:div align="left" name="Overskrift 4">
        <html:p>
            <html:b>
                <html:i>
                    <html:span>
                        Mildosat:
                    </html:span>
                </html:i>
            </html:b>
        </html:p>
    </html:div>

    """