Adana

Scripting programming language, repl and namespaced aliases for commands.

Table of Contents

1. Introduction
2. Installation
3. Programming language
   • Getting started
   • Comments
   • Multiline
   • F-Strings
   • Operators and constants
   • Variable definition
   • Memory Management
   • Plugins
   • Standard Library
   • Loops
   • Ranges
   • Conditions
   • Types
   • Structs
   • Manipulate arrays
   • Functions
   • Include a script file
   • Builtin functions
4. Namespaced aliases
   • Introduction
   • Try it
   • Available commands
   • Shortcuts
   • Environment variables
   • Arguments

Introduction

This project started as a way to put into practice what I learned while reading the Rust programming language book.

It includes features that I find useful, such as a REPL, a calculator, a scripting language, and a way to store, execute, and load command-line aliases based on the project I'm working on.

Best practices and performance optimization were not a priority, so the code may not be the cleanest or most optimized.

If you would like to contribute, your pull request would be welcome.

Where the name comes from?

My favorite dish 😋

Installation

1. Docker
   • From the docker hub:
     • docker run -it nbittich/adana # latest from master
     • docker run -it nbittich/adana:v0.18.5 # latest release
   • Manually:
     • clone the repo
     • build the docker image: docker build -t adana .
     • docker run -it adana
2. Cargo
   • From crate.io:
     • cargo install adana
     • adana
   • Manually:
     • cargo build --release
     • ./target/release/adana
3. WASM Playground
   • Try it out at https://nbittich.github.io/adana

Programming language

Getting Started

First, we start with the traditional hello world:

 println("hello world!") # prints hello world

In the repl, you could also simply write:

 "hello world!" # prints hello world

Comments

Comments are defined like in python, starting with #. You can put them after the last statement or before any useful code, for example:

 # to go to the next line in the repl, press CTRL+x

 # this will be ignored by the repl

 println("hello world!") # this is also ok

Multiline

Semicolons are not needed, if you need multiline statements, you can use a multiline block:

fancy_string = multiline {
    "this string\n" +
    "\tis\n" +
    "\t\ton several\n" +
    "lines\n"
}

Multiline is useful when you want to process different instructions in several lines:

complex_math_stuff = multiline {
    1 *2
    + 5 *sqrt(2) / 2.
    + 300 / 3
    * 400 % 2 - (1 * 10^3)
}

F-Strings

For more complex strings, you can use string blocks / F-Strings. You can define them using the java syntax:

block_string= """Hello world
I hope you are well.
This is a string block. you can use stuff like "string"
there, nothing will stop you"""

Like in javascript, you can embed expressions to an f-string:

person = struct {
            name  : "nordine",
            wasup : (age) => {
                if (age > 30) {
                    "you are old!"
                } else {
                    "you are young!"
                }
            },
            age  : 34
        }

s1 = """Hello ${person.name}!
You are ${person.age} years old.
${person.wasup(person.age)}"""

Operators and constants

There are 22 operators & 3 constants:

Example:

 5 + 5 # 10
 5 + 5.5 # 10.5
 5 / 5 # 1
 5 / 6 # 0
 5 / 6. # 0.8333333333333334 -- we force it to make a float division by adding "."
 5 % 6 # 5 -- modulo on int
 5 % 4.1 # 0.9000000000000004 modulo on double
 5 ^ 5 # 3125
 5 * 5 # 25
 5 * 5.1 # 25.5
 5 * (5+ 1/ (3.1 ^2) * 9) ^3. # 1046.084549281999
 2² # 4
 2³ # 8

You can apply an operator before re-assigning a variable, like:

x =2
x+=1 # 3
x-=2 # 1
x*=4 # 4
x%=3 # 1
x/=0.5 # 2

It is legal in some circumstances to use the multiply operator implicitly. It will only work when there is no space between a number (int, decimal) and a variable name.

Example:

x=2
3x²+2x== x*(3x+2) # true
y=0.5x # 1

Variable definition

To define a variable, simply type the name of the variable followed by "=". Variable must always start with a letter and can have numerics or "_" in it. Add and assign(+=), subtract and assign (-=), etc are also supported.

 vat = 1.21 # 1.21
 sub_total1 = 10000
 total = vat * sub_total1 # 12100
 sub_total2 = 500 * vat # 605
 sub_total1 = sub_total1 + sub_total2 # 10605

It could be simplified as such:

 vat = 1.21 # 1.21
 sub_total1 = 10000
 total = vat * sub_total1 # 12100
 sub_total2 = 500 * vat # 605
 sub_total1 += sub_total2 # 10605

It's also possible to use the special variable name _ to notify the language that this value is not used and doesn't have to be stored in context:

_ = 1

for _, n in 1..3 {
   println(n)
}

_ = struct {
   _: "I will not be stored!",
   x: 39
}

Memory Management

By default, everything is cloned. As a hobby project, this was a simple way to achieve more and have fun.

Now that enough features have been built, an attempt to implement automatic-ish reference counting has started.

This is highly experimental and a partial or full rewrite of the ast / parser may be needed to implement it properly. To keep the fun working on this, it is not a priority for now.

You can define a reference as such:

x = 100
y = &x # y points to x, no clone
p = 0
for _ in 0..&x {
  p = p+1
}

x = 99
y = &x
x = 100 # now y == 100

Plugins

It is possible to load plugins written in Rust dynamically. Because Rust doesn't have a stable ABI yet, the plugin must be built with the same version that was used to build adana.

The Rust version is specified when running the repl.

To load a library dynamically, you can either specify a relative path, or an absolute path. In case of a relative path, it should be relative to the shared lib path (by default: $HOME/.local/share/adana/db).

You can override this by providing a path when starting the repl (e.g: adana -slp /tmp).

If the path is a directory, it will try to build it using cargo, so you need to have Rust installed on your machine.

If it is an .so file, it will automatically load it.

An example of plugin can be found in this repo (dynamic_lib/example_lib_src).

For example:

• Copy the .so file in tmp: cp dynamic_lib/libplugin_example.so /tmp/
• Run and override the lib path: adana -slp /tmp
• Execute the following:

     lib = require("libplugin_example.so")
     text = lib.hello("Nordine", "la", "forme?")

Or in one line:

   text = require("libplugin_example.so").hello("Nordine", "la", "forme?")

Standard Library

A basic standard library exists here.

You can use it in this way:

fs = require("@std/fs")
fs.api_description() # description of the api

If it is not installed yet, you will see instructions on how to install it, e.g:

[rust~/toyprograms/adana(master)] fs = require("@std/fs")
std lib doesn't exist: "/home/nbittich/.local/share/adana/lib/adana-std/fs.so".

Try to install it like so:
    - wget -P /tmp https://github.com/nbittich/adana-std/releases/download/0.18.5/adana-std.tar.gz
    - mkdir -p /home/nbittich/.local/share/adana/lib/adana-std && tar xvzf /tmp/adana-std.tar.gz \
            -C /home/nbittich/.local/share/adana/lib/adana-std

Loops

There are two loops, the while loop and the for-each loop. The while loop looks like the one in C, while the for-each loop is a little bit more modern.

For-each loop & while loop don't require parenthesizes. You can only iterate over structs, strings and arrays.

count = 0

while(count < 10) {
    println(count)
    count = count + 1
}

# also valid
while count < 10 {
    println(count)
    count = count + 1
}

for n in [1,2,3] {
   println(n)
}

You have access to the current index in a for-each:

for index, n in [1, 2, 3] {
    println("index: " + index + " value: " + n)
}

It is also possible to use the for-each loop with a string:

for i, letter in "hello" {
  println(i)
}

In the case of a struct, the variable will be an entry (a struct with key/value)

s = struct {
    name: "nordine",
    age: 34,
    members: ["natalie", "roger","fred"]
}

for  id, entry in s {
     println("Id: "+id +" Key: "+entry.key + " Value: " + to_string(entry.value))
}

Parenthesizes are optional for for-each:

arr = [1,2,3,4]
total = 0
idx_total = 0
for (index,a in arr) {
 total = total + a
 idx_total = idx_total + index
}

You can break if you match a certain condition within a while:

while count < 10 {
     println(count)
     count = count + 1
     if(count % 3 ==0) {
        break
     }
}

Ranges

It is possible to define a range like so "start..end" (end exclusive), or "start..=end" (end inclusive):

x = 0..10 # [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
x = 0..=10 # [0,1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

for i in 1..=10 {
  println("print 10 times this message")
}

Conditions

Same as C but parenthesizes are optional:

if age > 12  {
    println("age > 12")
} else if age < 9 {
    println("age < 9")
} else {
    println("dunno")
}

Types

There is no type-checking in the language. You can add a string to an array, nothing will stop you!

In some cases though, you might get an error.

Below, is a list of types and how you declare them. You can also define your structure.

Structs

You can define structs. Structs are a way of grouping related variables or functions together. You can define function variables within a struct, but you cannot update the members of the function from within the struct (there is no self or this).

The comma is required to separate each member, but not for the latest one.

Example of defining a struct:

person = struct {
    name: "hello",
    age: 20,
    headers: struct {
      "Content-Type": "application/json"
    }
}

person_service = struct {
    say_hi: (person) => { println("hi " + person.name) },
    check_age: (person) => {
             if (person.age < 18) {
                 println("you are too young")
             } else {
                 println("you are too old")
             }
    }
}

person_service.check_age(person)

You can access a struct in two ways:

name = person["name"] # name contains "hello"
println(person["age"])

age=person.age # age contains "age"
println(person.age)

Manipulate arrays

Arrays are declared like in javascript but are "immutable". After declaration, you cannot (yet) push new data in them. to do that, you have to concatenate them with another array using the "+" operator.

 arr = [] # declare an empty array
 arr[0] = "kl" # Err: index out of range
 arr = arr + ["kl"] # arr now is ["kl"]

You can update a value in the array with the syntax above, as long as the array is greater than the index provided, e.g:

arr = ["this", "is", "ax", "array", true, 1, 2.3]
arr[2] = "an" #fix the typo
print(arr[2]) # an

To get the length of an array, you can use the built-in function length

 arr_len = length(arr) # 7

Characters within a string can be accessed & updated just like an array:

 s = "this is a strink"
 s[2] # i
 length(s) # 16
 s[15] = "g" # fix the typo
 s # this is a string

Here are some other examples of what you can do with arrays:

count = 9
arr = null
# arr = [1, [2, [3, [4, [5, [6, [7, [8, [9, null]]]]]]]]]
while(count > 0) {
    arr = [count, arr]
    count = count -1
}
# prints 1,2,3,4,5,6,7,8,9,done
while(arr != null) {
    print(arr[0] +",")
    arr=arr[1]
}
print("done")

Functions

The function can be declared inline or as a block. In the case of a function parameter, you either assign the function to a variable or use an anonymous function block.

Parameters cannot be modified within a function. if you want to update something, you have to return it and reassign it.

# no parameters
hello = () => { println("hello, world!") }
hello()

# one parameter
hello_name = (name) => { println("hello "+name) }
hello_name("Bachir")

# takes an array and a function as a parameter
for_each = (arr, consumer) => {
    count = 0
    len = length(arr)
    while(count < len) {
        consumer(arr[count])
        count = count + 1
    }
    return ""  # do not print the count as the repl will print the latest statement
}

for_each(["Mohamed", "Hakim", "Sarah", "Yasmine", "Noah", "Sofia", "Sami"], hello_name)
# or for_each(["Mohamed", "Hakim", "Sarah", "Yasmine", "Noah", "Sofia", "Sami"],
              (name) => { println("hello "+name) }
             )

Parameters cannot be modified within a function. if you want to update something, you have to return it and reassign it. Everything that changes within the scope of a function won't have any effect on the outer scope.

Some other examples of what you can do with functions:

arr  = ["Mohamed", "Hakim", "Sarah", "Yasmine", "Noah", "Sofia", "Sami"]

acc  = (arr, v) => {arr + [v]} # arr is immutable, thus you have to reassign it if you call that function

arr = acc(arr, "Malika")

find_first = (arr, predicate) => {
    len = length(arr)
    count = 0
    while(count < len) {
        temp = arr[count]
        if(predicate(temp)) {
            return temp
        }
        count = count + 1
    }
    return null
}


find_first(arr, (v) => {
    v[0] == "S" || v[0] == "s"
})

# recursive
fact = (n) => {
   if(n>=1) {
    n * fact(n-1)
   }else {
     1
   }
}
fact(10)

Include a script file

You can dynamically load a script written in adana in the repl. Assuming you've cloned the repo and you use docker, here's an example of how to do it.

Note that the extension must be .adana.

• map the example directory as a docker volume:

docker run -v $PWD/file_tests:/scripts -it adana

include("scripts/test_fn.adana") # the built-in function to include
m = map()
m = push_v("nordine", 34, m)
get_v("nordine", m)

Builtin functions

There are several built-in functions available.

You already have seen length to find the length of an array or string, include to include a script inside the repl and println to print something.

Here is a list of built-in functions available:

Matching regexes

There are two built-in functions for matching a regex against a string:

pattern = """(?i)a+(?-i)b+"""
text = "AaAaAbbBBBb"
is_match(text, pattern) # true

# match [["Item1: $100", "Item1", "100"], ["Item2: $200", "Item2", "200"], ["Item3: $300", "Item3", "300"]]
pattern = """(\w+): \$(\d+)"""
text = "Item1: $100, Item2: $200, Item3: $300"
match(text, pattern)

Note that you can use the repl command script_ctx to see what variables are stored in the context.

Namespaced aliases

You can alias useful commands in a separate namespace (e.g: "work", "git", "docker").

You can then run that command through the repl. They will be save in disk so you can backup them, restore them etc.

You can also add any kind of values (e.g, ssh keys) to store them.

There is no possible interaction with the scripting language yet.

Try it

docker run -it -v $PWD/sample.json:/adanadb.json nbittich/adana -im

restore

use misc

ds

printenv

Available commands

Shortcuts

CTRL + x => new line in the repl
CTRL + d => quit
CTRL + c => undo
CTRL + l => clear screen
CTRL + r => history search
CTRL + p => π

Environment variables

RUST_LOG=adana=debug adana

Arguments

TODO NOT EXHAUSTIVE

Run a script without entering in the repl

# using file
adana -sp /path/to/script.adana

# using file with --daemon feature
adana -sp /path/to/script.adana -d

# using code
adana -e 1+1

# open an in memory db

adana --inmemory

# override db path & history path + no fallback in memory in case of an error (default to false)
# path must exist! file doesn't have to.

adana --dbpath /tmp/mydb.db --historypath /tmp/myhistory.txt --nofb

# specify shared lib path
adana -slp /tmp/shared