TM::Serializable::CSV - Topic Maps, trait for parsing (and later dumping) CSV stream


   # 1) bare bones
   my $tm = .....;  # get a map from somewhere (can be empty)
   Class::Trait->apply ($tm, "TM::Serializable::CSV");

   use Perl6::Slurp;
   $tm->deserialize (slurp 'myugly.csv');

   # 2) exploiting the timed sync in/out mechanism
   my $tm = new TM::.... (url => 'file:myugly.csv'); # get a RESOURCEABLE map from somewhere


This trait provides parsing and dumping from CSV formatted text streams.




$tm->deserialize ($text)

This method consumes the text string passed in and interprets it as CSV formatted information. What topic map information is generated, depends on the header line (the first line):

  • If the header line contains a field called association-type, then all rows will be interpreted as assertions. In that the remaining header fields (in that order) are interpreted as roles (role types). For all rows in the CSV stream, the position where the association-type field was is ignored. The other fields (in that order) are affiliated with the corresponding roles.



    Scoping cannot be controlled. Also all players and roles (obviously) are directly interpreted as identifiers. Subject identifiers and locators are not (yet) implemented.

  • If the header line contains a field called id, then all further rows will be interpreted as topic characteristics, with each topic on one line. The column position where the id field in the header is will be interpreted as toplet identifier.

    All further columns will be interpreted according to the following:

    • If the header column is named name, the values will be used as topic names.

    • Otherwise if the value looks like a URI, an occurrence with that URI value will be be added to the topic.

    • Otherwise an occurrence with a string value will be added to the topic.




[Since TM 1.53] This method serializes a fragment of a topic map into CSV. Which fragment can be controlled with the header line and options (see constructor).

header_line (only for serialization)

This string contains a comma separated list (CSV parseable) of headings. If one of the headings is association-type, then the generated CSV content will contain associations only. Nothing else is implemented yet. The other headings control which roles (and in which order) should be included in the CSV content. If a particular role type has more than one player, then all players are included.

NOTE: As this is inconsistent, this will have to change.

type (only for serialization)

If existing, then this controls which association type is to be taken.

baseuri (only for serialization)

If existing and non-zero, the base URI of the map will remain in the identifiers. Otherwise it will be removed.


If existing (and when selecting only associations), this specification will be interpreted in the sense of asserts (see TM).


    $tm->serialize (header_line => 'association-type,location,bio-unit',
                    type        => 'is-born',
                    baseuri     => 0);


TM, TM::Serializable


Copyright 2010 Robert Barta.

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

1 POD Error

The following errors were encountered while parsing the POD:

Around line 297:

=back without =over