iLexicon is an “intelligent” dictionary that can be used to build Natural Language applications. I have two implementations, one in Lisp and another in Prolog. Both implementations are memory-based, in order to speed up performance. I have written several articles referencing it, for example see this.
LiteDB is a NoSQL database for .NET. I wrote about it a few weeks ago. Since that time, I have been working on implementing ILexicon using LiteDB as the back-end. That is almost complete. The primary reason for moving iLexicon to LiteDB is that I can now access the lexicon from both Lisp and Prolog through the respective COM/Automation modules.
In today’s article, I will show the primary COM Server interface and how it is used from Sicstus Prolog. iLexicon is quite feature-rich and it is difficult to demonstrate its capabilities in a single article. You may want to see my other posts related to this.
Here is the COM Server layer, written in C#.
COM Server
The above exposes two functions:
- hasPOS(Word, POS) checks if the given word has the indicated POS (remember, a word can have many parts of speech)
- hasLemma(Word, Lemma) checks if the given word has the indicated Lemma
The second functionality is often used when implementing an ATN-based (semantic) parser for English.
Here is the Prolog code that interfaces with the lexicon:
The Prolog Interface
I think the mapping is self-explanatory.
The following commands show how this can be used:
Sample Commands
To take this further, here is a toy grammar for English written in Definite Clause Grammar (DCG) notation:
A Toy English Grammar
The has_pos predicate is used to check the part-of-speech of the in-coming word. Here are some example sentences parsed using the grammar (I am skipping tokenization, structure building, etc.):
Parsing Using the Grammar
I used Visual Studio 2019 Release 16.9.2 for building the 64-bit COM server. In an earlier article, I have explained how to build the COM server in C#.
Have a nice weekend!
Recent Comments