/ 
MsOffice-Word-HTML-Writer-0.07
River stage zero No dependents
/ MsOffice::Word::HTML::Writer

NAME

MsOffice::Word::HTML::Writer - Writing documents for MsWord in HTML format

SYNOPSIS

use MsOffice::Word::HTML::Writer;
my $doc = MsOffice::Word::HTML::Writer->new(
  title        => "My new doc",
  WordDocument => {View => 'Print'},
);

$doc->write("<p>hello, world</p>", 
            $doc->page_break, 
            "<p>hello from another page</p>");

$doc->create_section(
  page => {size   => "21.0cm 29.7cm",
           margin => "1.2cm 2.4cm 2.3cm 2.4cm"},
  header => sprintf("Section 2, page %s of %s", 
                                $doc->field('PAGE'), 
                                $doc->field('NUMPAGES')),
  footer => sprintf("printed at %s", 
                                $doc->field('PRINTDATE')),
  new_page => 1,
);
$doc->write("this is the second section, look at header/footer");

$doc->attach("my_image.gif", $path_to_my_image);
$doc->write("<img src='files/my_image.gif'>");

$doc->save_as("/path/to/some/file");

DESCRIPTION

The present module is one way to programatically generate documents targeted for Microsoft Word (MsWord). It doesn't need MsWord to be installed, and doesn't even require a Win32 machine (which is why it is not in the Win32 namespace).

MsWord can read documents encoded in native binary format, in Rich Text Format (RTF), in WordML (an XML dialect), or -- maybe this is less known -- in HTML, with some special markup for pagination and other MsWord-specific features. Such HTML documents are often in several parts, because attachments like images or headers/footers need to be in separate files; however, since it is more convenient to carry all data in a single file, MsWord also supports the "MHTML" format (or "MHT" for short), i.e. an encapsulation of a whole HTML tree into a single file encoded in MIME multipart format. This format can be generated interactively from MsWord by calling the "SaveAs" menu and choosing the .mht extension.

Documents saved with a .mht extension will not directly reopen in MsWord : when clicking on such documents, Windows chooses Internet Explorer as the default display program. However, these documents can be simply renamed with a .doc extension, and will then open directly in MsWord. By the way, the same can be done with WordML or RTF documents. That is to say, MsWord is able to recognize the internal format of a file, without any dependency on the filename.

MsOffice::Word::HTML::Writer helps you to programatically generate MsWord documents in MHT format. The advantage of this technique is that one can rely on standard HTML mechanisms for layout control, such as styles, tables, divs, etc. Of course this markup can be produced using your favorite HTML templating module; the added value of MsOffice::Word::HTML::Writer is to help building the MIME multipart file, and provide some abstractions for representing MsWord-specific features (headers, footers, fields, etc.).

The MHT format is probably the most convenient way for programmatic document generation, because

  • unlike Excel, MsWord native binary format is unpublished and therefore cannot be generated without the MsWord executable.

  • remote control of the MsWord program through an OLE connection, as in Win32::Word::Writer, requires a local installation of Microsoft Office, and is not well suited for servers because the MsWord program might hang or might open dialog boxes that require user input.

  • generation of documents in RTF is possible, but requires deep knowledge of the RTF structure --- see RTF::Writer.

  • generation of documents in "WordML" also requires deep knowledge of WordML structure.

By contrast, MsOffice::Word::HTML::Writer allows you to produce documents even with little knowledge of MsWord. One word of warning, however : opening MHT documents in MsWord is slower than native binary or RTF documents, because MsWord needs to parse the HTML, compute the layout and convert it into its internal representation. Therefore MHT format is not recommended for large documents.

Note : this first release of MsOffice::Word::HTML::Writer is still in an exploratory phase; the programming interface may change in future versions.

METHODS

General convention : method names that start with a verb may change the internal state of the writer object (for example "write", "create_section"); method names that are nouns return data without modifying the internal state (for example "field", "content", page_break).

new

my $doc = MsOffice::Word::HTML::Writer->new(%params);

Creates a new writer object. Optional parameters are :

title

document title

any HTML declarations you may want to include in the head part of the generated document (for example inline CSS styles or links to attached stylesheets).

hf_head

any HTML declarations you may want to include in the head part of the headers and footers HTML document (MsWord requires headers and footers to be specified as divs in a separate HTML document).

WordDocument

a hashref of options to include as an XML island in the HTML head, corresponding to various options in the MsWord "Tools/Options" panel. These will be included in a XML element named <w:WordDocument>, and all children elements will be automatically prefixed by w:. The hashref may contain nested hashrefs, such as

WordDocument => { View => 'Print',
                  Compatibility => {DoNotExpandShiftReturn => "",
                                    BreakWrappedTables     => ""} }

Names and values of options must be found from the Microsoft documentation, or from reverse engineering of HTML files generated by MsWord.

Parameters may also be passed as a hashref instead of a hash.

write

$doc->write("<p>hello, world</p>");

Adds some HTML into the document body.

attach

$doc->attach($localname, $filename);
$doc->attach($localname, "<", \$content);
$doc->attach($localname, "<&", $filehandle);

Adds an attachment into the document; the attachment will be encoded as a MIME part and will be accessible under files/$localname.

The remaining arguments to attach specify the source of the attachment; they are directly passed to "open" in perlfunc and therefore have the same API flexibility : you can specify a filename, a reference to a memory variable, a reference to another filehandle, etc.

create_section

$doc->create_section(
  page => {size   => "21.0cm 29.7cm",
           margin => "1.2cm 2.4cm 2.3cm 2.4cm"},
  header => sprintf("Section 2, page %s of %s", 
                                $doc->field('PAGE'), 
                                $doc->field('NUMPAGES')),
  footer => sprintf("printed at %s", 
                                $doc->field('PRINTDATE')),
  new_page => 1,
);

Opens a new section within the document (or, if this is called before any "write", setups pagination parameters for the first section). Subsequent calls to the "write" method will add content to that section, until the next "create_section" call.

Pagination parameters are all optional and may be given either as a hash or as a hashref; accepted parameters are :

page

Hashref of CSS page styles, such as :

size

Paper size (for example 21cm 29.7cm)

margin

Margins (top right bottom left).

header_margin

Margin for header

Margin for footer

Header content (in HTML)

first_header

Header content for the first page of that section.

Footer content (in HTML).

Footer content for the first page.

new_page

If true, a page break will be inserted before the new section.

save_as

$doc->save_as("/path/to/some/file");

Generates the MIME document and saves it at the given location. If no extension is present, file extension .doc will be added by default to the filename.

content

Returns the whole MIME-encoded document as a single string; this is used internally by the "save_as" method. Direct call is useful if you don't want to save the document into a file, but want to do something else like embedding it in a message or a ZIP file, or returning it as an HTTP response.

page_break

Returns HTML markup for encoding a page break.

tab

my $html = $doc->tab($n_tabs);

Returns HTML markup for encoding one or several tabs. If $n_tab is omitted, it defaults to 1.

field

my $html = $doc->field($fieldname, $args, $content);

Returns HTML markup for a MsWord field.

Optional $args is a string with arguments or flags for the field. See MsWord help documentation for the list of field names and their associated arguments or flags.

Optional $content is the initial displayed content for the field (because unfortunately MsWord does not immediately compute the field content when opening the document; users will have to explicitly request to update all fields, by selecting the whole document and then hitting the F9 key).

Here are some examples :

my $header = sprintf "%s of %s", $doc->field('PAGE'), 
                                 $doc->field('NUMPAGES');
my $footer = sprintf "created at %s, printed at %s", 
               doc->field(CREATEDATE => '\\@ "d MM yyyy"'),
               doc->field(PRINTDATE  => '\\@ "dddd d MMMM yyyy" \\* Upper');
my $quoted = $doc->field('QUOTE', '"hello, world"', 'hello, world');

quote

my $html = $doc->quote($text);

Shortcut to produce a QUOTE field (see last field example just above).

AUTHORING MHT DOCUMENTS

HTML for MsWord

MsWord does not support the full HTML and CSS standard, so authoring MHT documents requires some trial and error. Basic divs, spans, paragraphs and tables, are reasonably supported, together with their common CSS properties; but fancier features like floats, absolute positioning, etc. may yield some surprises.

To specify widths and heights, you will get better results by using CSS properties rather than attributes of the HTML table model.

In case of difficulties for implementing specific features, try to see what MsWord does with that feature when saving a document in HTML format (plain HTM, not MHT!). The generated HTML is quite verbose, but after eliminating unnecessary tags one can sometimes figure out which are the key tags (they start with o: or w:) or the key attributes (they start with mso-) which correspond to the desired functionality.

Collaboration with the Template Toolkit

The Template Toolkit (TT for short) is a very helpful tool for generating the HTML. Below are some hints about collaboration between the two modules.

Client code calls both TT and Word::HTML::Writer

The first mode is to use the Template Toolkit for generating various document parts, and then assemble them into MsOffice::Word::HTML::Writer.

use Template;
my $tmpl_app = Template->new(%options);
$tmpl_app->process("doctmpl/html_head.tt", \%data, \my $html_head);
$tmpl_app->process("doctmpl/body.tt",      \%data, \my $body);
$tmpl_app->process("doctmpl/header.tt",    \%data, \my $header);
$tmpl_app->process("doctmpl/footer.tt",    \%data, \my $footer);

use MsOffice::Word::HTML::Writer;
my $doc = MsOffice::Word::HTML::Writer->new(
  title  => $data{title},
  head   => $html_head,
);
$doc->create_section(
  header => $header,
  footer => $footer,
);
$doc->write($body);
$doc->save_as("/path/to/some/file");

This architecture is straightforward, but various document parts are split into several templates, which might be inconvenient when maintaining a large body of document templates.

HTML parts as blocks in a single template

Document parts might also be encoded as blocks within one single template :

[% BLOCK html_head %]
<style>...CSS...</style>
[% END; # BLOCK html_head %]

[% BLOCK body %]
  Hello, world
[% END; # BLOCK body %]

etc.

Then the client code calls each block in turn to gather the various parts :

use Template::Context;
my $tmpl_ctxt = Template::Context->new(%options);
my $tmpl      = $tmpl_ctxt->template("doctmpl/all_blocks.tt");
my $html_head = $tmpl_ctxt->process($tmpl->blocks->{html_head}, \%data);
my $body      = $tmpl_ctxt->process($tmpl->blocks->{body},      \%data);
my $header    = $tmpl_ctxt->process($tmpl->blocks->{header},    \%data);
my $footer    = $tmpl_ctxt->process($tmpl->blocks->{footer},    \%data);

# assemble into MsOffice::Word::HTML::Writer, same as before

Template toolkit calls MsOffice::Word::HTML::Writer

Now let's look at a different architecture: the client code calls the Template toolkit, which in turn calls MsOffice::Word::HTML::Writer.

The most common way to call modules from TT is to use a TT plugin; but since there is currently no TT plugin for MsOffice::Word::HTML::Writer, we will just tell TT that templates can load regular Perl modules, by turning on the LOAD_PERL option.

The client code looks like any other TT application; but the output of the process method is a fully-fledged MHT document, instead of plain HTML.

use Template;
my $tmpl_app = Template->new(LOAD_PERL => 1, %other_options);
$tmpl_app->process("doc_template.tt", \%data, \my $msword_doc);

Within doc_template.tt, we have

[% # main entry point

   # gather various parts
   SET html_head = PROCESS html_head;
   SET header    = PROCESS header;
   SET footer    = PROCESS footer;
   SET body      = PROCESS body;

   # create Word::HTML::Writer object
   USE msword = MsOffice.Word.HTML.Writer(head=html_head);

   # setup section format
   CALL msword.create_section(
      page => {size          => "21.0cm 29.7cm",
               margin        => "1cm 2.5cm 1cm 2.5cm",
               header_margin => "1cm",
               footer_margin => "0cm",},
      header => header,
      footer => footer
    );

    # write the body
   CALL msword.write(body);

   # return the MIME-encoded MsWord document
   msword.content();  %]

[% BLOCK html_head %]
...

Inheritance through TT views

The above architecture can be refined one step further, by using TT views to encapsulate documents. Views have an inheritance mechanism, so it becomes possible to define families of document templates, that inherit properties or methods from common ancestors. Let us start with generic_letter.tt2, a generic letter template :

[% VIEW generic_letter
      title="Generic letter template";

     BLOCK main;
       USE msword = MsOffice.Word.HTML.Writer(
          title => view.title,
          head  => view.html_head(),
       );
       view.write_body();
       msword.content();
     END; # BLOCK main
  
     BLOCK write_body;
       CALL msword.create_section(
          page   => {size          => "21.0cm 29.7cm",
                     margin        => "1cm 2.5cm 1cm 2.5cm"},
          header => view.header(),
          footer => view.footer()
       );
       CALL msword.write(view.body());
     END write_body; %]
  
     BLOCK body;
       view.letter_head();
       view.letter_body();
     END; # BLOCK body
  
     BLOCK letter_body; %]
      Generic letter body; please override BLOCK letter_body in subviews
  [% END; # BLOCK letter_body;

     # ... other blocks for header, footer, letter_head, etc.

   END; # VIEW generic_letter

[% # call main() method if this templated was loaded directly
   letter.main() UNLESS component.caller %]

This is quite similar to an object-oriented class : assignments within the view are like object attributes (i.e. the title variable), and blocks within the view are like methods.

After the end of the view, we call the main method, but only if that view was called directly from client code. If the view is inherited, as displayed below, then the call to main will be from the subview.

Now we can define a specific letter template that inherits from the generic letter and overrides the letter_body block :

[% PROCESS generic_letter.tt2; # loads the parent view 

   VIEW advertisement;

     BLOCK letter_body; %]

       <p>Dear [% receiver.name %],</p>
       <p>You have won a wonderful [% article %].
          Just call us at [% sender.phone %].</p>
       <p>Best regards,</p>
       [% view.signature(name => sender.name ) %]

[%   END; # BLOCK letter_body
   END; # VIEW advertisement

   advertisement.main() UNLESS component.caller %]

TO DO

This module is still exploratory; many features need to be added. For example:

- odd/even pages
- link same header/footers across several sections
- multiple columns
- watermarks (I tried hard to reverse engineer MsWord behaviour, 
  but it still doesn't work ... couldn't figure out all details 
  of VML markup)

Contributions welcome!

AUTHOR

Laurent Dami, <laurent DOT dami AT etat DOT geneve DOT ch>

BUGS

Please report any bugs or feature requests to bug-win32-word-html-writer at rt.cpan.org, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=MsOffice-Word-HTML-Writer. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.

SUPPORT

You can find documentation for this module with the perldoc command.

perldoc MsOffice::Word::HTML::Writer

You can also look for information at:

SEE ALSO

Win32::Word::Writer, RTF::Writer, Spreadsheet::WriteExcel.

COPYRIGHT & LICENSE

Copyright 2009 Laurent Dami, all rights reserved.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.