Merge branches

.gitignore

@@ -1 +1,2 @@
+node_modules/
 target/

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "tree-sitter-dust"]
+	path = tree-sitter-dust
+	url = ssh://git@git.jeffa.io:22022/jeff/tree-sitter-dust.git

Cargo.toml

@@ -0,0 +1,51 @@
+[package]
+name = "dust-lang"
+version = "0.1.1"
+description = "Data-oriented programming language and interactive shell."
+authors = ["jeff <dev@jeffa.io.com>"]
+repository = "https://git.jeffa.io/jeff/dust.git"
+homepage = "https://git.jeffa.io/jeff/dust"
+readme = "README.md"
+license = "MIT"
+edition = "2018"
+default-run = "dust"
+
+[[bin]]
+name = "dust"
+
+[[bin]]
+name = "gui"
+
+[lib]
+name = "dust_lib"
+path = "src/lib.rs"
+
+[build-dependencies]
+cc = "*"
+
+[dependencies]
+tree-sitter = "0.20.10"
+tree-sitter-rust = "0.20.3"
+rand = "0.8.5"
+chrono = "0.4.26"
+trash = "3.0.3"
+rayon = "1.7.0"
+serde = { version = "1.0.171", features = ["derive"] }
+sysinfo = "0.29.6"
+toml = "0.7.6"
+toml_edit = "0.19.14"
+comfy-table = "7.0.1"
+clap = { version = "4.3.19", features = ["derive"] }
+git2 = "0.17.2"
+csv = "1.2.2"
+json = "0.12.4"
+reqwest = { version = "0.11.18", features = ["blocking", "json"] }
+serde_json = "1.0.104"
+egui_extras = "0.22.0"
+rustyline = { version = "12.0.0", features = ["with-file-history", "derive"] }
+ansi_term = "0.12.1"
+iced = "0.10.0"
+egui = "0.22.0"
+eframe = "0.22.0"
+env_logger = "0.10.0"
+once_cell = "1.18.0"

README.md

@@ -1,30 +1,43 @@
 # Dust
 
-Dust is a data-oriented programming language and interactive shell. Dust can be used as a replacement for a traditional command line shell, as a scripting language and as a tool create or manage data. Dust is expression-based, has first-class functions, lexical scope and lightweight syntax.
+Dust is a data-oriented programming language and interactive shell. Dust can be used as a replacement for a traditional command line shell, as a scripting language and as a data format. Dust is expression-based, has first-class functions, lexical scope and lightweight syntax. Dust's grammar is formally defined in code and its minimalism is in large part due to its tree sitter parser, which is lightning-fast, accurate and thoroughly tested.
 
 A basic dust program:
 
 ```dust
-output "Hello world!"
+output { "Hello world!" }
 ```
 
 Dust can do two (or more) things at the same time with effortless concurrency:
 
 ```dust
-run(
-    'output "will this one finish first?"',
-    'output "or will this one?"'
-)
+run <
+    function { output 'will this one finish first?' }
+    function { output 'or will this one?' }
+>
 ```
 
 Dust can do amazing things with data. To load CSV data, isolate a column and render it as a line plot in a GUI window:
 
 ```dust
 read_file("examples/assets/faithful.csv")
-    -> from_csv(input)
-    -> rows(input)
-    -> transform(input, 'input.1')
-    -> plot(input)
+    -> from_csv
+    -> rows
+    -> transform <{item.1}>
+    -> plot
+```
+
+Dust is also a minimal, obvious data format. It is easier to write than JSON and easier to read than TOML and YAML. However, because it is a programming language, it is able to self-reference, perform calculations or load external data.
+
+```dust
+foo = "bar"
+numbers = [1 2 3 4]
+truths = {
+    dust = "the best thing ever"
+    favorite_number = numbers.3
+    another_number = numbers.0 + numbers.1
+}
+old_faithful_data = read_file { "faithful.csv" }
 ```
 
 <!--toc:start-->
@@ -32,26 +45,26 @@ read_file("examples/assets/faithful.csv")
   - [Features](#features)
   - [Usage](#usage)
   - [Installation](#installation)
+  - [Implementation](#implementation)
   - [Contributing](#contributing)
   - [The Dust Programming Language](#the-dust-programming-language)
-    - [Variables and Data Types](#variables-and-data-types)
-    - [Tools](#tools)
+    - [Declaring Variables](#declaring-variables)
+    - [Integers and Floats](#integers-and-floats)
     - [Lists](#lists)
     - [Maps](#maps)
     - [Tables](#tables)
-    - [The Yield Operator](#the-yield-operator)
     - [Functions](#functions)
-    - [Time](#time)
+    - [Empty Values](#empty-values)
 <!--toc:end-->
 
 ## Features
 
-- Data visualization: GUI (not TUI) plots, graphs and charts are available from directly within dust. No external tools are needed.
-- Powerful tooling: Built-in commands reduce complex tasks to plain, simple code. You can even partition disks or install software.
+- Simplicity: Dust is designed to be easy to learn and powerful to use, without compromising either.
+- Speed: Dust is built on [Tree Sitter] and [Rust] to prioritize performance and correctness.
+- Data format: Dust is data-oriented, so first and foremost it makes a great language for defining data.
 - Pipelines: Like a pipe in bash, dust features the yield `->` operator.
 - Format conversion: Effortlessly convert between dust and formats like JSON, CSV and TOML.
 - Structured data: Dust can represent data with more than just strings. Lists, maps and tables are easy to make and manage.
-- Developer tools: Dust has a complete tree sitter grammar, allowing syntax highlighting and completion in most code editors.
 
 ## Usage
 
@@ -60,10 +73,10 @@ Dust is an experimental project under active development. At this stage, feature
 To get help with the shell you can use the "help" tool.
 
 ```dust
-help()         # Returns a table will all tool info.
-help("random") # Returns a table with info on tools in the specified group.
+help            # Returns a table will all tool info.
+help {"random"} # Returns a table with info on tools in the specified group.
 # The above is simply a shorthand for this:
-help() -> where(input, 'tool == "random"')    
+help -> where { input, function <tool> { tool == "random" } }
 ```
 
 ## Installation
@@ -72,19 +85,23 @@ You must have the default rust toolchain installed and up-to-date. Install [rust
 
 To build from source, clone the repository and run `cargo run` to start the shell. To see other command line options, use `cargo run -- --help`.
 
+## Implementation
+
+Dust is formally defined as a Tree Sitter grammar in the tree-sitter-dust module. Tree sitter generates a parser, written in C, from a set of rules defined in JavaScript. Dust itself is a rust binary that calls the C parser using FFI. Dust does not use Javascript at runtime.
+
+Tree Sitter generates a concrete syntax tree, which the Rust code maps to an abstract syntax tree by traversing each node once. Tree sitter is fast enough to be updated on every keystroke which is perfect for a data-oriented language like Dust because it allows only the relevant sections to be re-evaluated and the result displayed instantly.
+
 ## Contributing
 
-Please submit any thoughts or suggestions for this project. To contribute a new command, see the library documentation. Implementation tests are written in dust and are run by a corresponding rust test so dust tests will be run when `cargo test` is called.
+Please submit any thoughts or suggestions for this project. For instructions on the internal API, see the library documentation. Implementation tests are written in dust and are run by a corresponding rust test so dust tests will be run when `cargo test` is called.
 
 ## The Dust Programming Language
 
-Dust is a hard fork of [evalexpr]; a simple expression language. Dust's core language features maintain this simplicity. But it can manage large, complex sets of data and perform complicated tasks through commands. It should not take long for a new user to learn the language, especially with the assistance of the shell.
+It should not take long for a new user to learn the language, especially with the assistance of the shell. If your editor supports tree sitter, you can use [tree-sitter-dust] for syntax highlighting and completion support. Aside from this guide, the best way to learn dust is to read the examples and tests to get a better idea of what dust can do.
 
-If your editor supports tree sitter, you can use [tree-sitter-dust] for syntax highlighting and completion support. Aside from this guide, the best way to learn dust is to read the examples and tests to get a better idea of what dust can do.
+### Declaring Variables
 
-### Variables and Data Types
-
-Variables have two parts: a key and a value. The key is always a text string. The value can be any of the following data types:
+Variables have two parts: a key and a value. The key is always a string. The value can be any of the following data types:
 
 - string
 - integer
@@ -94,165 +111,111 @@ Variables have two parts: a key and a value. The key is always a text string. Th
 - map
 - table
 - function
-- time
-- empty
 
 Here are some examples of variables in dust.
 
 ```dust
-string = "The answer is 42.";
-integer = 42;
-float = 42.42;
-list = (1, 2, string, integer, float);
-map.key = "value";
-empty = ();
+string = "The answer is 42."
+integer = 42
+float = 42.42
+list = [1 2 string integer float] # Commas are optional when writing lists.
+map = {
+    key = `value`
+}
 ```
 
-### Tools
+Note that strings can be wrapped with any kind of quote: single, double or backticks. Numbers are always integers by default. And commas are optional in lists.
 
-**Tools** are dust's built-in functions. Some of them can reconfigure your whole system while others do very little. They may accept different inputs, or none at all. For example, commands in the `random` group can be run without input, but the `random_integer` command can optionally take two numbers as in inclusive range.
+### Integers and Floats
 
-```dust
-die_roll = random_integer(1, 6);
-d20_roll = random_integer(1, 20);
-coin_flip = random_boolean();
-```
-
-```dust
-message = "I hate dust.";
-replace(message, "hate", "love")
-```
+Integer and floating point values are dust's numeric types. Any whole number (i.e. without a decimal) is an integer. Floats are declared by adding a single decimal to or number. If you divide integers or do any kind of math with a float, you will create a float value.
 
 ### Lists
 
-Lists are sequential collections. They can be built by grouping values with parentheses and separating them with commas. Values can be indexed by their position to access their contents. Lists are used to represent rows in tables and most commands take a list as an argument. Their contents can be indexed using dot notation with an integer.
+Lists are sequential collections. They can be built by grouping values with square brackets. Commas are optional. Values can be indexed by their position to access their contents. Their contents can be indexed using dot notation with an integer. Dust lists are zero-indexed.
 
 ```dust
-list = (true, 41, "Ok");
+list = [true 41 "Ok"]
 
-assert_equal(list.0, true);
+assert_equal { list.0 true }
 
-the_answer = list.1 + 1;
+the_answer = list.1 + 1
 
-assert_equal(the_answer, 42);
+assert_equal { the_answer, 42 }
 ```
 
 ### Maps
 
-Maps are flexible collections with arbitrary key-value pairs, similar to JSON objects. Under the hood, all of dust's runtime variables are stored in a map, so, as with variables, the key is always a string.
+Maps are flexible collections with arbitrary key-value pairs, similar to JSON objects. Under the hood, all of dust's runtime variables are stored in a map, so, as with variables, the key is always a string. A map is created with a pair of curly braces and its entries and just variables declared inside those braces. Map contents can be accessed using dot notation and a value's key.
 
 ```dust
-reminder.message = "Buy milk";
-reminder.tags = ("groceries", "home");
+reminder = {
+    message = "Buy milk"
+    tags = ["groceries", "home"]
+}
 
-json = to_json(reminder);
-append(json, "info.txt");
+output { reminder.message }
 ```
 
 ### Tables
 
-Tables are strict collections, each row must have a value for each column. Empty cells must be explicitly set to an empty value.
+Tables are strict collections, each row must have a value for each column. If a value is "missing" it should be set to an appropriate value for that type. For example, a string can be empty and a number can be set to zero. Dust table declarations consist of a list of column names, which are identifiers enclosed in pointed braces. The column names are followed by a pair of curly braces filled with list values. Each list will become a row in the new table.
 
 ```dust
-animals = create_table (
-    ("name", "species", "age"),
-    (
-        ("rover", "cat", 14),
-        ("spot", "snake", 9),
-        ("bob", "giraffe", 2)
-    )
-);
+animals = table <name species age> {
+    ["rover" "cat" 14]
+    ["spot" "snake" 9]
+    ["bob" "giraffe" 2]
+}
 ```
 
 Querying a table is similar to SQL.
-  
+
 ```dust
-names = select(animals, "name");
-youngins = where(animals, 'age < 5');
+names = select name from animals
+youngins = select species from animals where age <= 10
 ```
 
-The commands `create_table` and `insert` make sure that all of the memory used to hold the rows is allocated at once, so it is good practice to group your rows together instead of using a call for each row.
+The keywords `table` and `insert` make sure that all of the memory used to hold the rows is allocated at once, so it is good practice to group your rows together instead of using a call for each row.
 
 ```dust
-insert(
-    animals,
-    (
-        ("eliza", "ostrich", 4),
-        ("pat", "white rhino", 7),
-        ("jim", "walrus", 9)
-    )
-);
+insert into animals {
+    ["eliza" "ostrich" 4]
+    ["pat" "white rhino" 7]
+    ["jim" "walrus" 9]
+}
 
-assert_equal(count(animals.all), 6);
-
-sorted = sort(animals);
-```
-
-### The Yield Operator
-
-Like a pipe in bash, zsh or fish, the yield operator evaluates the expression on the left and passes it as input to the expression on the right. That input is always assigned to the **`input` variable** for that context. These expressions may simply contain a value or they can call a command or function that returns a value.
-
-```dust
-"Hello dust!" -> output(input)
-```
-
-This can be useful when working on the command line but to make a script easier to read or to avoid fetching the same resource multiple times, we can also declare variables. You should use `->` and variables together to write efficient, elegant scripts.
-
-```dust
-json = download("https://api.sampleapis.com/futurama/characters");
-from_json(json)
-    -> select(input, "name");
-    -> input.4
+assert_equal { count { animals }, 6 };
 ```
 
 ### Functions
 
-Functions are first-class values in dust, so they are assigned to variables like any other value. The function body is wrapped in single parentheses. To call a function, it's just like calling a command: simply pass it an argument or use an empty set of parentheses to pass an empty value.
-
-In the function bod, the **`input` variable** represents whatever value is passed to the function when called.
+Functions are first-class values in dust, so they are assigned to variables like any other value. The function body is wrapped in single parentheses. To create a function, use the "function" keyword. The function's arguments are identifiers inside of a set of pointed braces and the function body is enclosed in curly braces. To call a fuction, invoke its variable name and use a set of curly braces to pass arguments (or leave them empty to pass nothing). You don't need commas when listing arguments and you don't need to add whitespace inside the function body but doing so may make your code easier to read. Use your best judgement, the parser will disambiguate any valid syntax.
 
 ```dust
-say_hi = 'output "hi"';
-add_one = 'input + 1';
+say_hi = function <> {
+    output {"hi"}
+}
 
-say_hi();
-assert_equal(add_one(3), 4);
+add_one = function <number> {
+    number + 1
+}
+
+say_hi {}
+assert_equal { add_one{3}, 4 }
 ```
 
 This function simply passes the input to the shell's standard output.
 
 ```dust
-print = 'output(input)';
+print = function <input> {
+    output { input }
+}
 ```
 
-Because functions are stored in variables, we can use collections like maps to
-organize them.
+### Empty Values
 
-```dust
-math.add = 'input.0 + input.1';
-math.subtract = 'input.0 - input.1';
-
-assert_equal(math.add(2, 2), 4);
-assert_equal(math.subtract(100, 1), 99);
-```
-
-### Time
-
-Dust can record, parse and convert time values. Dust can parse TOML datetime
-values or can create time values using commands.
-
-```dust
-dob = from_toml("1979-05-27T07:32:00-08:00")
-
-output "Date of birth = " + local(dob);
-```
-
-```dust
-time = now();
-
-output "Universal time is " + utc(time);
-output "Local time is " + local(time);
-```
+Dust does not have a null type. Instead, it uses the "empty" type to represent a lack of any other value. There is no syntax to create this value: it is only used by the interpreter. Note that Dust does have the NaN value, which is a floating point value that must exist in order for floats to work as intended. Integers will never be NaN and no value will ever be null or undefined.
 
 [dnf]: https://dnf.readthedocs.io/en/latest/index.html
 [evalexpr]: https://github.com/ISibboI/evalexpr

examples/fibonacci.ds

@@ -0,0 +1,11 @@
+fibonacci = function <number> {
+	if number <= 0 then 0
+	else if number == 1 then 1
+	else
+		first = fibonacci { number - 2 }
+		second = fibonacci { number - 1 }
+
+		first + second
+}
+
+

src/evaluator.rs

@@ -1,57 +1,3 @@
-//! The top level of Dust's API with functions to interpret Dust code.
-//!
-//! You can use this library externally by calling either of the "eval"
-//! functions or by constructing your own Evaluator.
-use std::fmt::{self, Debug, Formatter};
-
-use serde::{Deserialize, Serialize};
-use tree_sitter::{Node, Parser, Tree as TSTree, TreeCursor};
-
-use crate::{language, Error, Result, Value, VariableMap};
-
-/// Evaluate the given source code.
-///
-/// Returns a vector of results from evaluating the source code. Each comment
-/// and statemtent will have its own result.
-///
-/// # Examples
-///
-/// ```rust
-/// # use dust_lib::*;
-/// assert_eq!(eval("1 + 2 + 3"), vec![Ok(Value::from(6))]);
-/// ```
-pub fn eval(source: &str) -> Vec<Result<Value>> {
-    let mut context = VariableMap::new();
-
-    eval_with_context(source, &mut context)
-}
-
-/// Evaluate the given source code with the given context.
-///
-/// # Examples
-///
-/// ```rust
-/// # use dust_lib::*;
-/// let mut context = VariableMap::new();
-///
-/// context.set_value("one".into(), 1.into());
-/// context.set_value("two".into(), 2.into());
-/// context.set_value("three".into(), 3.into());
-///
-/// let dust_code = "four = 4 one + two + three + four";
-///
-/// assert_eq!(
-///     eval_with_context(dust_code, &mut context),
-///     vec![Ok(Value::Primitive(Primitive::Empty)), Ok(Value::from(10))]
-/// );
-/// ```
-pub fn eval_with_context(source: &str, context: &mut VariableMap) -> Vec<Result<Value>> {
-    let mut parser = Parser::new();
-    parser.set_language(language()).unwrap();
-
-    Evaluator::new(parser, context, source).run()
-}
-
 /// This trait is implemented by the Evaluator's internal types.
 pub trait EvaluatorTree: Sized {
     /// Interpret the syntax tree at the given node and return the abstraction.

src/lib.rs

@@ -14,6 +14,7 @@ pub use crate::{
 };
 
 mod error;
+mod evaluator;
 mod interface;
 mod value;
 
@@ -35,6 +36,45 @@ pub fn language() -> Language {
 /// [`node-types.json`]: https://tree-sitter.github.io/tree-sitter/using-parsers#static-node-types
 pub const NODE_TYPES: &'static str = include_str!("../../../src/node-types.json");
 
+/// Evaluate the given source code.
+///
+/// Returns a vector of results from evaluating the source code. Each comment
+/// and statemtent will have its own result.
+///
+/// ```rust
+/// # use dust_lib::*;
+/// assert_eq!(eval("1 + 2 + 3"), vec![Ok(Value::from(6))]);
+/// ```
+pub fn eval(source: &str) -> Vec<Result<Value>> {
+    let mut context = VariableMap::new();
+
+    eval_with_context(source, &mut context)
+}
+
+/// Evaluate the given source code with the given context.
+///
+/// ```rust
+/// # use dust_lib::*;
+/// let mut context = VariableMap::new();
+///
+/// context.set_value("one".into(), 1.into());
+/// context.set_value("two".into(), 2.into());
+/// context.set_value("three".into(), 3.into());
+///
+/// let dust_code = "one + two + three";
+///
+/// assert_eq!(
+///     eval_with_context(dust_code, &mut context),
+///     vec![Ok(Value::Empty), Ok(Value::from(6))]
+/// );
+/// ```
+pub fn eval_with_context(source: &str, context: &mut VariableMap) -> Vec<Result<Value>> {
+    let mut parser = Parser::new();
+    parser.set_language(language()).unwrap();
+
+    Evaluator::new(parser, context, source).run()
+}
+
 // Uncomment these to include any queries that this grammar contains
 
 // pub const HIGHLIGHTS_QUERY: &'static str = include_str!("../../queries/highlights.scm");
@@ -45,7 +85,7 @@ pub const NODE_TYPES: &'static str = include_str!("../../../src/node-types.json"
 #[cfg(test)]
 mod tests {
     #[test]
-    fn test_can_load_grammar() {
+    fn load_grammar() {
         let mut parser = tree_sitter::Parser::new();
         parser
             .set_language(super::language())

tree-sitter-dust

@@ -0,0 +1 @@
+Subproject commit 916b59b4b6ff3a9ece271a292932202200df04b8