Recent changes to feature-requests

Pluggable fact readers and result exporters?

2013-10-15T11:35:17.176000Z

It would be nice if one could import facts directly from a SQL database, CSV file or JSON file. Likewise it would be useful to be able to export the results in various forms.

Simple tutorial with examples needed

2013-10-15T11:13:31.347000Z

Currently, the documentation, while a good reference, is a bit convoluted and hard to approach for new users of PyKE. A tutorial written in layman's terms using some simple (and short) examples would be helpful and make the project more approachable and easier to start using.

Also, it would be good to illustrate more the different purpose of forward vs. backward chaining rules and their relative merits as it seems they can to some extent be used interchangeably?

Move to github?

2013-10-15T11:07:23.877000Z

Would you care to move this project to github? I bet there would be a lot more community involvement. At least it would be a lot easier. SourceForge is so lagging behind.

Function returning list of specific facts

2013-09-12T09:13:02.783000Z

Hello,

it would be nice to have a function similar to fact_base.dump_specific_facts() which returns the list of specific facts rather than just printing them. As for now I'm using this to retrieve the list of facts after forward chaining:

...
engine.activate('fc')
mylist = engine.knowledge_bases['ontology'].entity_lists['triple'].case_specific_facts

Is it right? If yes, it would be nice to have a function doing just this (and the same for universal facts).

Thanks,

Nicola

Automatically create fact bases mentioned in fc rules

2010-04-23T21:51:23Z

Pyke currently throws an error if a forward-chaining rule references a fact base that has no facts in it. This happens before any rules are run to be able to add facts, which would create the fact base.

This suggestion is to create these fact bases automatically, rather than raising an exception.

Create top-level pyke class

2008-01-07T01:12:26Z

Create a top-level pyke class to support multiple instances of the knowledge engine.

This would also be useful to be able to run the pyke compiler in a separate pyke instance without having to launch another copy of the python interpretor.

Automatically generate plan specs

2008-01-04T15:56:44Z

It would be nice if pyke could automatically generate "$$()" plan specs for *when* premises that return a plan but don't have a plan spec specified. This might generate a "return $$()" for the last premise if there is also no *with* clause.