xbars-py

Currently scrapes Wiktionary to return dictionary entries, and stores data regarding parts of speech and verb inflections. The goal is to ultimately use this data to produce an automatic syntax tree generator.

Currently, there is one main module:

• dictsearch.py

Requirements

This project was tested and developed on Python 3.10.2. The following libraries must be installed:

• Beautiful Soup 4 - used by the scraper in dictsearch.py
• (tentative) Natural Language Toolkit (NTLK) - used by psr.py and parser.py.

Unfortunately, I can't make any promises regarding which versions of Python 3.x this project targets. (tentative): I haven't tested thoroughly on Python 3.7, but it is the minimum Python version supported by NTLK.

Modules

dictsearch.py - a basic Wiktionary scraper

Aims to provide an extremely basic interface to scrape English language dictionary entries from Wiktionary.

The class constructor automatically does the scraping in the background, so the interface consists solely of the constructor and a few instance variables. As such, you'll want to make a new instance of the class when you want to search a new word, instead of modifying the instance variables.

Interface

DictSearch(word)

Searches for word on Wiktionary, scrapes the data if an English entry exists, and stores a dictionary entry and part of speech and verb inflection data.

• If a page for word does not exist on Wiktionary, a PageNotFound exception is raised.
• If a page for word exists, but an English entry for it does not exist, a NoEnglishEntry exception is raised.

DictSearch.POSList

• set class attribute that holds the parts of speech relevant for forming syntax trees. If a word has a part of speech that does not belong to DictSearch.POSList but is included in DictSearch._POSList, .entry will include the definition, but this library does not intend to automatically generate phrase structure rules for DictSearch instances with an empty .POS.

.word

• str instance variable that holds the word used to initialize this instance.

.entry

• str instance variable that holds the dictionary entry on Wiktionary for the word.

.POS

• set instance variable that holds the parts of speech associated with .word in English.

.isAux =False

• bool instance variable that is True if .word is an auxiliary/modal verb and is False otherwise.

.isPlural =False

• bool instance variable that is True if .word is a plural noun and is False otherwise.

.verbRoot =""

• str instance variable that holds the root form of .word if it is a verb inflected for tense/aspect and is False otherwise.

.nounRoot =""

• str instance variable that holds the root form of .word if it is a noun inflected for plurality and is False otherwise.

.verbInflections =set()

• set instance variable that holds any verb inflections associated with .word in English and is an empty set if .word is not an inflected verb.

---

psr.py - [WORK IN PROGRESS]

(tentative): Holds the phrase structure rules later imported by parser.py. This module expands the functionality offered by NTLK's implementation of context-free grammars in nltk.grammar.CFG to demonstrate phrasal and head movement.

My implementation may be hacky, and from a (computational) linguistics perspective, there's a number of potential complaints about my implementation. For one, I violate the recursive spirit of Chomsky's generative syntax approach by defining multiple specialized phrase structure rules for, say, subject DPs/TPs/CP. Second of all, to the best of my knowledge, CFGs are unable to account for phrasal/head movement and subject/object control structures, so there may be fundamental concerns with using CFGs in the first place here. Unfortunately, due to the limits of my theoretical knowledge, this is a case of the ends taking precedence over the means.

---

parser.py - [WORK IN PROGRESS]

Given a target configuration, aims to output potential parse trees for an input phrase/sentence.

Restrictions - [WORK IN PROGRESS]

I'm wondering if it'll be possible to show raising/control structures if I have a list of words that trigger this behavior; until I manage to actually get the basic functionality working, I simply can't really say if this'll end up being the case or not.