Benker’s documentation 0.5

The Benker library can be used to convert tables from one format to another.

Yes, it only converts the tables, not the whole document, but it tries to do it well. The document itself is not changed, and the paragraphs inside the cells, neither. It’s your responsibility to do this part of the work.

General conversion process

Available formats

The Benker library works on XML documents. Currently, it can handle:

✱ OOXML is an XML-based format for office documents, including word processing documents, spreadsheets, presentations, as well as charts, diagrams, shapes, and other graphical material. This is the XML format used by Microsoft Word documents: *.docx.

_images/logo-ooxml.png

✱ CALS table model is a standard for representing tables in SGML/XML. Developed as part of the CALS Department of Defence initiative. The DTD of the CALS table model is available in the OASIS web site.

_images/logo-oasis.png

✱ Formex describes the format for the exchange of data between the Publication Office and its contractors. In particular, it defines the logical markup for documents which are published in the different series of the Official Journal of the European Union. Formex v4 is based on the international standard XML.

_images/logo-publications-office.png

Conversion stages

To convert a document, Benker uses several stages:

  1. Parse the source document and construct a nodes tree,

  2. Search for table elements and construct the table objects,

  3. Build the target nodes tree by replacing table nodes,

  4. Serialise the target document.

General sequence diagram

Converters: Parsers + Builders

The decoupling between parsing, building and final serialization allows a simplified and modular implementation. This decoupling also allows to multiply the combinations: it is easy to change a builder to another one, and to develop its own parser…

The advantage of this approach is that we avoid having a specific document conversion for each format pair (input, output). Instead, you can build a converter by choosing a parser and a builder, as you assemble the pieces of a puzzle.

Converter = Parser + Builder

The following table show you the available converters which groups parser and builders by pairs.

Available converters

_images/puzzle-ooxml-builder-40.png _images/puzzle-html-builder-40.png _images/puzzle-cals-builder-40.png _images/puzzle-formex4-builder-40.png
_images/puzzle-ooxml-parser-40.png

(unavailable)

convert_ooxml2cals()

convert_ooxml2formex()

_images/puzzle-html-parser-40.png

(unavailable)

(unavailable)

(unavailable)

_images/puzzle-cals-parser-40.png

(unavailable)

(unavailable)

convert_cals2formex()

_images/puzzle-formex4-parser-40.png

(unavailable)

(unavailable)

convert_formex2cals()

You can create your own converter by inheriting the available base classes:

General class diagram (BaseConverter, BaseParser and BaseBuilder)
  • BaseConverter: inherit this class to create your own converter. Set your own parser class to the parser_cls class attribute, and your own builder class to the builder_cls class attribute.

  • BaseParser: inherit this class to create your own parser. The method transform_tables() is an abstract method, so you need to implement it in your subclass: it must call the method generate_table_tree() each time a table node is found and converted to a Table object.

  • BaseBuilder: inherit this class to create your own builder. The method generate_table_tree() is an abstract method, so you need to implement it in your subclass: it must convert the Table object into a target XML node (the resulting table format). You can also implement the method finalize_tree() to do any post-processing to the resulting XML tree.

Hint

Contribution is welcome!

Usage

For example, to convert the tables of a .docx document to Formex4 format, you can process as follow:

import os
import zipfile

from benker.converters.ooxml2formex import convert_ooxml2formex

# - Unzip the ``.docx`` in a temporary directory
src_zip = "/path/to/demo.docx"
tmp_dir = "/path/to/tmp/dir/"
with zipfile.ZipFile(src_zip) as zf:
    zf.extractall(tmp_dir)

# - Source paths
src_xml = os.path.join(tmp_dir, "word/document.xml")
styles_xml = os.path.join(tmp_dir, "word/styles.xml")

# - Destination path
dst_xml = "/path/to/demo.xml"

# - Create some options and convert tables
options = {
    'encoding': 'utf-8',
    'styles_path': styles_xml,
}
convert_ooxml2formex(src_xml, dst_xml, **options)

This code produces a table like that:

<TBL COLS="7" NO.SEQ="0001">
  <CORPUS>
    <ROW>
      <CELL COL="1" ROWSPAN="2">
        <w:p w:rsidR="00EF2ECA" w:rsidRDefault="00EF2ECA"><w:r><w:t>A</w:t></w:r></w:p>
      </CELL>
      <CELL COL="2" COLSPAN="2">
        <w:p w:rsidR="00EF2ECA" w:rsidRDefault="00EF2ECA"><w:r><w:t>B</w:t></w:r></w:p>
      </CELL>
      <CELL COL="4">
        <IE/>
      </CELL>
      <CELL COL="5">
        <IE/>
      </CELL>
      <CELL COL="6">
        <IE/>
      </CELL>
      <CELL COL="7">
        <IE/>
      </CELL>
    </ROW>
    <ROW>
      ...
    </ROW>
  </CORPUS>
</TBL>

The content of the cells still contains OOXML fragments. It’s your own responsibility to convert them to the target format.

Indices and tables