eval.md

arrai eval and run

Arr.ai's eval and run commands evaluate argument as an arr.ai expression and, by default, print the result to stdout.

The eval command evaluates its argument as an arr.ai expression.

$ arrai eval '2 + 2'
4

The e command is a shortcut and behaves identically to eval.

$ arrai e '2 + 2'
4

The run command evalutes the contents of the source file specified by its argument.

$ cat hi.arrai
$"Hello, ${//os.args(1)}!"
$ arrai run hi.arrai Alice
Hello, Alice!
$ arrai r hi.arrai Alice
Hello, Alice!

Output controls

The eval command supports the --out flag (shorthand -o), which changes arr.ai's output behaviour as follows:

option	description
--out=file:<path>	Outputs the evaluated result, which must be a string or byte array, to the file specified by <path>. file may be abbreviated to f.
--out=dir:<path>	Outputs the evaluated result to the directory specified by <path>. The result to be output must be a recursively nested dict of dicts, with strings and/or byte arrays at the leaves. Each dict represents the contents of a directory at the location given by its key while each string or byte array correspondingly represents the contents of a file. dir may be abbreviated to d.

For both --out=file and --out=dir, strings are UTF-8 encoded when written to the corresponding file.

Examples

$ arrai e --out=file:example.txt '$"Hello, ${//os.args(1)}!\n"' Bob
$ cat example.txt
Hello, Bob!
$ arrai e -o f:example.txt '"hello\n"'
$ cat example.txt
hello
$ arrai e -o :example.txt '"hello\n"'
$ cat example.txt
hello
$ arrai e -o example.txt '"hello\n"'
$ cat example.txt
hello

Note that --out=file offers several shortenings. The final example above only works if the path itself doesn't have a : in it.

$ arrai e --out=dir:out 'let [_, name, ...] = //os.args;
{
    "foo.txt": $"Hello, ${name}.\n",
    "bar": {"baz.txt": <<"Goodbye, ", name, "!", 10>>}
}' Bob
$ tree out
out
├── bar
│   └── baz.txt
└── foo.txt

1 directory, 2 files
$ cat out/foo.txt
Hello, Bob.
$ cat out/bar/baz.txt
Goodbye, Bob!