EasySyntaxTree

A Visual Studio Code usage environment for rsyntaxtree.

Installation

NOTE: Everything in this repository assumes you are on Mac or Linux. I don't believe anything in this repository will work correctly on Windows, other than some simple Python scripts.

1. Install Visual Studio Code.
2. Once installed, click Command+Shift+P, then paste Shell Command: Install 'code' command in PATH into the text field, then click enter.
3. Open up a terminal on your Mac (if you don't know how, just type Command+Spacebar, then type in 'terminal' and the app will be the first result).
4. In the terminal, type git --version, then click enter. If you simply get an output like git version 2.30.0, then continue to step 5. I assume that you will get an installer popup though, so go through the steps for that installation.
5. Now that git is installed, type git clone git://github.com/camball/EasySyntaxTrees.git in the terminal, and click enter.
6. In the terminal, type bash rst-install.sh, then click enter (this may take a couple minutes; just let it run).

NOTE: These instructions are for those lacking technical knowledge/experience working in a Unix Environment. If you know what you are doing, feel free to pick apart the installation script and manually install things. The script assumes you have a fresh install of vscode on your computer, and may do unwanted actions to your settings.json if you already have vscode installed.

Usage

1. With an stree file open in vscode, click "Run Code". This will produce a png file of the syntax tree for the current stree file. The png will have the same name as the current stree file as well.

Motivation & Explanation

I found rsyntaxtree to be a great tool, but it was lacking two major features that were absolutely necessary for me to use it:

1. A way to provide file input
2. A way to name output files

Without file input, this tool was essentially useless to me. I wouldn't be able to save and manage my syntax trees without this. The way rsyntaxtree is set up is entire syntax tree needs to be provided as command line arguments, which I find incredibly counterintuitive. It also killed me that the tool would only output files called syntree.*, depending on the specified output file type.

So, to fix these problems, I thought to set up vscode as a remedy. My methodology was the following:

• Define a custom text file type to store syntax trees in. This way vscode knows when we are dealing with a syntax tree file. I decided on the file extension .stree.
• Customise Code Runner so that when I click 'Run Code' from an stree file in vscode, the png image is produced without any hassle.
• Rename the output png file from rsyntaxtree to have the same name as the stree file. The user can set up a system for organising their syntax trees this way.

I also wanted to be able to use developer tools that I would use for regular software development, like syntax highlighting and bracket matching. For this, I made sure to configure vscode to do this for the stree file type.

I found that it was still cumbersome to type lots of syntax trees, though. So to help, I created a few Python scripts to handle some repetitive tasks:

• Take in a list of sentences from a text file and put them into their own stree files.
• Surround each word in each stree file in a directory with brackets. This cuts down the amount of boilerplate work the user has to do.
• A combination of the above tasks.

There are many more helpful things I would like to add, but they probably aren't worth my time. I can think of all sorts of helpful things, though, if someone else wants to develop them. Namely...

• A way for vscode to syntax check your stree files, like it would for most programming languages.
• A way to define a "formal, natural language", meaning a set of phrase structure rules for a specific natural language (like English or French) that vscode could use to perform a semantic check of your stree files. The power of this would be "extended" syntax checking, where vscode could tell you when your tree is in violation of the rules of a natural language (e.g., you put a preposition after a noun phrase in a prepositional phrase (assuming English language), even if your tree is syntactically correct).
• Auto-complete for constituents. This would make typing bracket-diagrams for long sentences much easier. For example, the user could type "VP" then use the up/down arrow keys to select options containing pre-formatted dominated constituents according to the phrase structure rules for the natural language they are working in.