Segment Sort

Create string sort algorithm from segment config.

Usage

npm i segment-sort

import segmentSorter from 'segment-sort';

// Case sensitive, upper < lower < _
const comparator1 = segmentSorter(['AZ', 'az', '_']);
['a', 'b', 'A', 'B', '_'].sort(comparator1); // ['A', 'B', a', 'b', '_']

// Case insensitive, lower < upper < _
const comparator2 = segmentSorter(['aA', '_']);
['a', 'b', 'A', 'B', '_'].sort(comparator2); // ['a', 'A', 'b', 'B', '_']

For more info, please check the APIs.

APIs Documentation

Type Aliases

Comparator

Ƭ Comparator: (a: string | undefined, b: string | undefined) => number

Type declaration

▸ (a, b): number

Type of a function to compare two strings.

Parameters

Name	Type
a	string | undefined
b	string | undefined

Returns

number

0 if a === b; or negative if a < b; or positive if a > b

Defined in

index.ts:13

---

CompareRule

Ƭ CompareRule: SegSymbol[]

String comparison rule.

Defined in

types.ts:16

---

SegSymbol

Ƭ SegSymbol: "az" | "AZ" | "aA" | "aZ" | "Aa" | "Az" | "_"

Symbols for char segments:

• az - Lower case letters, i.e. [a-z].
• AZ - Upper case letters, i.e. [A-Z].
• aA or aZ - Both case letters and lower case first, i.e. [a-zA-Z] and 'a' < 'A' < 'b' < 'B' < ...
• Aa or Az - Both case letters and upper case first, i.e. [a-zA-Z] and 'A' < 'a' < 'B' < 'b' < ...
• _ - Chars with ASCII from 91 to 96, i.e. [, \, ], ^, _ , `(backtick).

Defined in

types.ts:11

Functions

default

▸ default(rule): undefined | Comparator

Generate a string comparison function based on the given rule.

Parameters

Name	Type	Description
rule	CompareRule	Comparison rule

Returns

undefined | Comparator

A string comparison function; or undefined if rule is invalid

Defined in

index.ts:22

Algorithm

The key of segment-sort is to define a CompareRule so it creates a custom string sort algorithm for you.

Compare Rule

A CompareRule is an array of segments.

Segment

A segment is a collection of characters with a sorting rule. Currently, there are 5 predefined segments:

• "az": Lower-case letters ([a-z]) sorted alphabetically.
• "AZ": Upper-case letters([A-Z]) sorted alphabetically.
• "aA" or "aZ": Both case letters ([a-zA-Z]) sorted case-insensitively and lower case first in case of a tie ('a' < 'A' < 'b' < 'B' < ...).
• "Aa" or "Az": Both case letters ([a-zA-Z]) sorted case-insensitively and upper case first in case of a tie ('A' < 'a' < 'B' < 'b' < ...).
• "_" - Chars of ASCII from 91 to 96, i.e. [, \, ], ^, _, `(backtick), sorted alphabetically.

Case Sensitivity

The segments used in CompareRule implicitly decide whether to compare strings case-sensitively or -insensitively:

• "az" or "AZ": Compare strings case-sensitively;
• "aA", "Aa", "aZ" or "Az": Compare strings case-insensitively;

Some Examples

["_", "aA"] or ["_", "aZ"]

• Strings are compared case-insensitively, and lower case goes first in case of a tie.
• [, \, ], ^, _, `(backtick) are in front of letters ([a-zA-Z]).

A sorted example is ['_', 'a', 'A', 'b', 'B'].

["Aa", "_"] or ["Az", "_"]

• Strings are compared case-insensitively, and upper case goes first in case of a tie.
• [, \, ], ^, _, `(backtick) are after letters ([a-zA-Z]).

This is widely used, e.g. as the default option ("case-insensitive") in TSLint Rule: ordered-imports. A sorted example is ['A', 'a', 'B', 'b', '_'].

["az", "_", "AZ"]

• Strings are compared case-sensitively, and lower-case letters ([a-z]) are in front of upper-case letters ([A-Z]).
• [, \, ], ^, _, `(backtick) are behind lower-case letters and before upper-case letters.

This corresponds to "lowercase-first" in TSLint Rule: ordered-imports. A sorted example is ['a', 'b', '_', 'A', 'B'].

["AZ", "_", "az"]

• Strings are compared case-sensitively, and upper-case letters ([A-Z]) are in front of lower-case letters ([a-z]).
• [, \, ], ^, _, `(backtick) are behind upper-case letters and before lower-case letters.

This corresponds to "lowercase-last" in TSLint Rule: ordered-imports.

A sorted example is ['A', 'B', '_', 'a', 'b'].

Incomplete Rules

The algorithm in this package is smart enough to complete CompareRule by appending missing segments in the end.

For example, ["az", "_"] will be padded with "AZ", and equals to ["az", "_", "AZ"].

But it will give up if there is uncertainty. For example, ["az"] can't be completed as the order between "_" and "AZ" is unknown, hence undefined is returned.

Here are some incomplete but meaningful rules:

• ["az", "_"] => ["az", "_", "AZ"]
• ["AZ", "_"] => ["AZ", "_", "az"]
• ["Aa"] => ["Aa", "_"]
• ["aA"] => ["aA", "_"]

Overlapped Rules

When segments overlap with each other, the one that appears first takes effect.

For example, ["aA", "az"] is equal to ["aA"] because "az" is covered by previous "aA".

["az", "aA"] is equal to ["az", "AZ"] because the lower-case part of "aA" is overlapped, but not the upper-case part.

The algorithm tolerates overlapped rules for better usability but you should treat them as potential mistakes.

License

MIT © Zhao DAI [e-mail]