zhangtaihao / simplerdf Goto Github PK

In the API documentation for URI formatters, the point needs to be made that if a formatter produces resolvable URIs, RDF triples where the object URI is made with the formatter enable semantic crawlers to follow through and find the RDF metadata of related objects.

A similar message needs also to be added in the custom URI pattern form.

Postponed. Moved from #8.

Taxonomy fields, possibly needs to work in tandem with #14.

Need to add data type mapping. This possibly involves stitching on an additional (optional) layer of plugins to provide certain data types (and subsequently data formatting) for field types. At this point, how this will influence the "fields" present for term and user mapping I have no idea.

Add weights to node URI generators.

When saving a basic custom field, I get the following error:

Warning: array_filter() [function.array-filter]: The first argument should be an array in _simplerdf_custom_fields_admin_edit_basic_form_validate() (line 259 of /simplerdf/modules/simplerdf_custom_fields/simplerdf_custom_fields.admin.inc).

On reediting the field, the "Applicable object type" is unchecked.

Expand prefixes when constructing ARC2 index.

Also track semsol/arc2#1.

Content mappers currently provide only for one or more modules. They should be able to specify, for example, different variants of "number" without switching field by themselves. This will block on #5.

As in title.

Rewrite simplerdf_sparql_index() to make it possible for implementors to re-index using the Batch API.

Most of this should be done for the update hooks to be written. Still, need to check that the variables are actually created and cleaned up properly.

Minimal information needs to be published for user. Specifically:

• Name
• Email
• Homepage

This task is to generalize the mapper concept into a "meta-mapper". This will necessarily mean strapping on a simplistic entity-like subsystem to base node, term, and user mappers upon.

When view path is node/%/rdf and RDF is turned on, show a warning in mappings and settings forms.

The dashboard is sprawled out at the moment. It needs to be collapsed into individual tabs, possibly via jQuery.

Currently, simplerdf_admin_mappings_form_inject() loads values from the database. The function needs to do that initially and then repopulate values from $form_state should the form containing intermediate values be rebuilt.

See if we can't integrate simplerdf into the RDF module.

Two aspects to consider:

• Granularity
• Start/end (possibly repeat)

Need to find out whether or not the namespace is actually defined or merely not shown.

Generally, vocabularies and terms need to have basic metadata. Specifically:

• Vocabulary/term mappings:
  • Class
  • Name
  • Description
• Term-only mappings:
  • Parent
  • Related
  • Synonym

These settings should be per vocabulary and stored as variables (so as to be easily exportable via Strongarm).

Extract the RDF display mechanisms (e.g. rendering, auto-discovery links) and place them into an optional module to make Simple RDF play nice with RDF. This will block on #20.

Add a CRUD-triggered SPARQL endpoint that indexes the site for Simple RDF-built triples.

Write up new entries in the API documentation to reflect the latest plugin architecture.

In Simple RDF SPARQL, when mappings are changed, cached RDF data become stale.

This task consists of two parts:

• An option in the SPARQL settings form for specifying whether or not to automatically re-index RDF on mappings change (still need to consider whether to split option by type, though probably not).
• If the option is checked, on saving mappings the admin will be directed to a batch execution screen for indexing. Otherwise, a warning message is shown advising the admin to re-index manually.

Write test cases for the module.

Add ordering to the mapping page.

• Add a new entry in mapper plugin to specify whether it wants to allow reordering.
• If reordering is enabled, add callback to revert ordering to default (for example, reverting to CCK weights).
• If a feature default ordering is present, use this ordering as default instead of the regular defaults.

It now makes more sense to specify "field modules" to provide for in a content mapper. Originally, it was envisioned that mappers providing for field modules can detect specific field metadata on their own. However, we need to account for a possibility to provide for field modules as well as (or alternatively to) field types.

At the very least, provide what RDF CCK provides:

• Link
• Number
• FileField
• ImageField (UPDATE: 2010-12-02; moved to #13)
• Node reference (?)
• User reference (?)

The question marks above refer to the fact that links are local and may therefore be useless.

On RDF-provided paths, the displayed RDF document contains unexpanded prefixes.

At the moment, the SPARQL endpoint is directly from ARC2. The visual styles are not consistent with the rest of the site. Ideally, this should be under the site's theme.

This will most likely involve extending ARC2_StoreEndpoint.