muss/mps-interpreter

mps-interpreter

All necessary components to interpret and run a MPS script.

MpsInterpretor uses a non-standard Iterator implementation, so it is recommended to use MpsRunner to execute a script. Since MPS is centered around iterators, script execution is also done by iterating.

MpsInterpretor is misspelt to emphasise that it behaves strangely: after every MPS statement, a None item is returned even when the script is not complete. MpsRunner wraps MpsInterpretor so that this behaviour is hidden when iterating.

use std::io::Cursor;
use mps_interpreter::*;

let cursor = Cursor::new(
"files(folder=`~/Music/`, recursive=true)" // retrieve all files from Music folder
);

let interpreter = MpsRunner::with_stream(cursor);

// warning: my library has ~3800 songs, so this outputs too much information to be useful.
for result in interpreter {
    match result {
        Ok(item) => println!("Got song `{}` (file: `{}`)", item.title, item.filename),
        Err(e) => panic!("Got error while executing: {}", e),
    }
}

Standard Vocabulary

By default, the standard vocabulary is used to parse the stream when iterating the interpreter. The standard vocabulary defines the valid statement syntax for MPS and parses syntax into special Rust Iterators which can be used to execute the statement. To declare your own vocabulary, use MpsRunner::with_settings to provide a MpsInterpretor with a custom vocabulary (I'm not sure why you would, but I'm not going to stop you).

Oddities

The MPS standard syntax does a few things that most other languages don't, because I wanted it to.

` can be used in place of " -- To make it easier to write SQL, string literals may be surrounded by backticks instead of quotation marks.

; -- The REPL will automatically place semicolons when Enter is pressed and it's not inside of brackets or a literal. MPS requires semicolons at the end of every statement, though, so MPS files must use semicolons.

Filters

Operations to reduce the items in an iterable: iterable.(filter);. Filters are statements of the format something.(predicate), where "something" is a variable name or another valid statement, and "predicate" is a valid filter predicate (see below). E.g. files(folder="~/Music/", recursive=true).(title == "Romantic Traffic"); is valid filter syntax to filter all songs in the Music folder for songs named "Romantic Traffic" (probably just one song).

field == something

field like something

field matches some_regex

field != something

field >= something

field > something

field <= something

field < something -- e.g. iterable.(title == "Romantic Traffic");

Compare all items, keeping only those that match the condition. Valid field names are those of the MpsMusicItem (title, artist, album, genre, track, etc.), though this will change when proper object support is added. Optionally, a ? or ! can be added to the end of the field name to skip items whose field is missing/incomparable, or keep all items whose field is missing/incomparable (respectively).

start..end -- e.g. iterable.(0..42);

Keep only the items that are at the start index up to the end index. Start and/or end may be omitted to start/stop at the iterable's existing start/end (respectively). This stops once the end condition is met, leaving the rest of the iterator unconsumed.

start..=end -- e.g. iterable.(0..=42);

Keep only the items that are at the start index up to and including the end index. Start may be omitted to start at the iterable's existing start. This stops once the end condition is met, leaving the rest of the iterator unconsumed.

index -- e.g. iterable.(4);

Keep only the item at the given index. This stops once the index is reached, leaving the rest of the iterator unconsumed.

predicate1 || predicate2 -- e.g. iterable.(4 || 5);

Keep only the items that meet the criteria of predicate1 or predicate2. This will always consume the full iterator.

[empty] -- e.g. iterable.();

Matches all items

if filter: operation1 else operation2 -- e.g. iterable.(if title == "Romantic Traffic": repeat(item, 2) else item.());

Replace items matching the filter with operation1 and replace items not matching the filter with operation2. The else operation2 part may be omitted to preserve items not matching the filter. To perform operations with the current item, use the special variable item. The replacement filter may not contain || -- instead, use multiple filters chained together.

Functions

Similar to most other languages: function_name(param1, param2, etc.);. These always return an iterable which can be manipulated. Functions are statements of the format function_name(params), where "function_name" is one of the function names (below) and params is a valid parameter input for the function. Each function is responsible for parsing it's own parameters when the statement is parsed, so this is very flexible. E.g. files(folder="~/Music/", recursive=true); is valid function syntax to execute the files function with parameters folder="~/Music/", recursive=true.

sql_init(generate = true|false, folder = "path/to/music");

Initialize the SQLite database connection using the provided parameters. This must be performed before any other database operation (otherwise the database will already be connected with default settings).

sql("SQL query here");

Perform a raw SQLite query on the database which MPS auto-generates. An iterator of the results is returned.

song("something");

Retrieve all songs in the database with a title like something.

album("something");

Retrieve all songs in the database with an album title like something.

artist("something");

Retrieve all songs in the database with an artist name like something.

genre("something");

Retrieve all songs in the database with a genre title like something.

repeat(iterable, count);

Repeat the iterable count times, or infinite times if count is omitted.

files(folder = "path/to/music", recursive = true|false, regex = "pattern");

Retrieve all files from a folder, matching a regex pattern.

reset(iterable);

Explicitly reset an iterable. This useful for reusing an iterable variable.

interlace(iterable1, iterable2, ...);

Combine multiple iterables in an interleaved pattern. This is a variant of union(...) where the first item in iterable1, then iterable2, ... is returned, then the second item, etc. until all iterables are depleted. There is no limit on the amount of iterables which can be provided as parameters.

union(iterable1, iterable2, ...);

Combine multiple iterables in a sequential pattern. All items in iterable1 are returned, then all items in iterable2, ... until all provided iterables are depleted. There is no limit on the amount of iterables which can be provided as parameters.

intersection(iterable1, iterable2, ...);

Combine multiple iterables such that only items that exist in iterable1 and iterable2 and ... are returned. The order of items from iterable1 is maintained. There is no limit on the amount of iterables which can be provided as parameters.

empty();

Empty iterator. Useful for deleting items using replacement filters.

Sorters

Operations to sort the items in an iterable: iterable~(sorter) OR iterable.sort(sorter)

field -- e.g. iterable~(filename)

Sort by a MpsItem field. Valid field names change depending on what information is available when the MpsItem is populated, but usually title, artist, album, genre, track, filename are valid fields. Items with a missing/incomparable fields will be sorted to the end.

shuffle

random shuffle -- e.g. iterable~(shuffle)

Shuffle the songs in the iterator. This is random for up to 2^16 items, and then the randomness degrades (but at that point you won't notice). The more verbose syntax is allowed in preparation for future randomisation strategies.

advanced bliss_first -- e.g. iterable~(advanced bliss_first)

Sort by the distance (similarity) from the first song in the iterator. Songs which are more similar (lower distance) to the first song in the iterator will be placed closer to the first song, while less similar songs will be sorted to the end. This uses the bliss music analyser, which is a very slow operation and can cause music playback interruptions for large iterators. This requires the advanced feature to be enabled (without the feature enabled this is still valid syntax but doesn't change the order).

advanced bliss_next -- e.g. iterable~(advanced bliss_next)

Sort by the distance (similarity) between the last played song in the iterator. Similar to bliss_first. The song which is the most similar (lower distance) to the previous song in the iterator will be placed next to it, then the process is repeated. This uses the bliss music analyser, which is a very slow operation and can cause music playback interruptions for large iterators. This requires the advanced feature to be enabled (without the feature enabled this is still valid syntax but doesn't change the order).

Procedures

Operations to apply to each item in an iterable: iterable.{step1, step2, ...}

Comma-separated procedure steps will be executed sequentially (like a for loop in regular programming languages). The variable item contains the current item of the iterable.

let variable = something -- e.g. let my_var = 42

Declare the variable and (optionally) set the initial value to something. The assignment will only be performed when the variable has not yet been declared. When the initial value (and equals sign) is omitted, the variable is initialized as empty().

variable = something -- e.g. my_var = 42

Assign something to the variable. The variable must have already been declared.

empty() -- e.g. empty()

The empty or null constant.

if condition { something } else { something_else } -- e.g.

if item.title == `Romantic Traffic` {

   } else {
       remove item
   }

Branch based on a boolean condition. Multiple comma-separated procedure steps may be supplied in the if and else branches. This does not currently support if else chains, but they can be nested to accomplish similar behaviour.

something1 == something2

something1 != something2

something1 >= something2

something1 > something2

something1 <= something2

something1 < something2 -- e.g. item.filename != item.title

Compare something1 to something2. The result is a boolean which is useful for branch conditions.

op iterable_operation -- e.g. op files().(0..=42)~(shuffle)

An iterable operation inside of the procedure. When assigned to item, this can be used to replace item with multiple others. Note that iterable operations are never executed inside the procedure; when item is iterable, it will be executed immediately after the end of the procedure for the current item.

(something1)

-something1

something1 - something2

something1 + something2

something1 || something2

something1 && something2 -- e.g. 42 + (128 - 64)

Various algebraic operations: brackets (order of operations), negation, subtraction, addition, logical OR, logical AND; respectively.

Item(field1 = something1, field2 = something2, ...) - e.g. item = Item(title = item.title, filename = /dev/null)

Constructor for a new item. Each function parameter defines a new field and it's value.

~string_format something -- e.g. ~{filename} item

Format a value into a string. This behaves differently depending on the value's type: When the value is an Item, the item's corresponding field will replace all {field} instances in the format string. When the value is a primitive type (String, Int, Bool, etc.), the value's text equivalent will replace all {} instances in the format string. When the value is an iterable operation (Op), the operation's script equivalent will replace all {} instances in the format string.

License: LGPL-2.1-only OR GPL-3.0-only