Crate fltrs

source · []
Expand description

Fltrs

Easy to define filters for querying lists. Fltrs has no dependencies!

Overview

Fltrs want to support creating easy, fast and expandable filters for iterable things (like Vec, Array, Map, Set, …) in rust. A filter is created based on an input string (query). This has particular advantages if the filter is created at runtime, i.e. in a GUI or command line tool (CLI).

Extensions:

It is possible, to expand the filter/query to your own needs:

  • create your own crate::operator
  • create a converter for the filter Value (e.g.: conversion of units).

You can find examples on the Query builder page.

Examples:

use fltrs::query;

let result: Vec<_> = [3, 2, 1, 4, 5, 7, 5, 4, 3]
        .into_iter()
        .filter(query("> 1 and < 5").unwrap())
        .collect();

assert_eq!(vec![3, 2, 4, 4, 3], result);
use fltrs::query;

let result: Vec<_> = ["Inge", "Petra", "Paul", "Egon", "Peter"]
        .into_iter()
        .filter(query("contains 'e'").unwrap())
        .collect();

assert_eq!(vec!["Inge", "Petra", "Peter"], result);

Option queries:

use fltrs::query;

let result: Vec<Option<char>> = [None, Some('a'), None, Some('b'), Some('c'), Some('a')]
        .into_iter()
        .filter(query(" != 'a' and not = none ").unwrap())
        .collect();

assert_eq!(vec![Some('b'), Some('c')], result);

Nested and Not queries:

use fltrs::query;

let result: Vec<_> = [3, 2, 1, 4, 5, 7, 5, 4, 3]
        .into_iter()
        .filter(query("(= 1 or = 5) and > 1").unwrap())
        .collect();

assert_eq!(vec![5, 5], result);
 use fltrs::query;

 let result: Vec<_> = [3, 2, 1, 4, 5, 7, 5, 4, 3]
         .into_iter()
         .filter(query("not( (= 1 or = 5) and > 1)").unwrap())
         .collect();

 assert_eq!(vec![3, 2, 1, 4, 7, 4, 3], result);

Fltrs supported queries on structs too.

This is possible, if the struct implement the trait: PathResolver.

use fltrs::{PathResolver, Filterable, query};

#[derive(PartialEq, Debug)]
struct Point {
    name: &'static str,
    x:    i32,
    y:    i32,
}

impl PathResolver for Point {
    fn path_to_index(path: &str) -> Option<usize> {
        match path {
            "name"  => Some(0),
            "x"     => Some(1),
            "y"     => Some(2),
            _ => None,
        }
    }

    fn value(&self, idx: usize) -> &dyn Filterable {
        match idx {
            0 => &self.name,
            1 => &self.x,
            _ => &self.y,
        }
    }
}

let result: Vec<Point> =
    [
      Point { name: "Point_1_3", x: 1, y: 3},
      Point { name: "Point_3_3", x: 3, y: 3},
      Point { name: "Point_2_6", x: 2, y: 6},
    ]
     .into_iter()
     .filter(query(r#"x one_of [3, 7]"#).unwrap())
     .collect();

assert_eq!(vec![Point { name: "Point_3_3", x: 3, y: 3}], result);

Re-exports

pub use crate::error::FltrError;

Modules

The error mod contains two error types: FltrError and ParseError.

The operator mod contains operators, from which the predicates are defined.

The value mod contains different value types, which are the result of the parse process.

Structs

The Query is an builder to configure the query(). It is possible, to extend the Operators in the modul: crate::operator.

Traits

Is a replacement for the std::fmt::Display trait. It is not possible to implement Display for the enum std::option::Option.

Filterable means, the given value can be compared to Value and implement the trait core::fmt::Display.

PathResolver is a possibility to get the value from a field of an given struct.

Functions

The query function create a Predicate respectively core::ops::Fn with which you can execute a filter on a given slice.

Type Definitions

A Predicate is an boxed core::ops::Fn.

The default core::result::Result with the error: FltrError.