Bible Reference Formatter

This project formats OSIS-style Bible references (Matt.1.2-Matt.1.3) as:

1. Human-readable references in English (Matthew 1:2-3).
2. Paratext-style Bible references (MAT 1:2-3).

For the reverse, to convert human-readable passage references to OSIS references, see Bible Passage Reference Parser.

Node Usage

$3

en.js exports exactly one function (below as osisToEn) that takes two required string arguments and one optional string argument:

1. Output format type (described below).
2. OSIS string. A comma-separated list of references: e.g., Matt.1.2-Matt.1.3,Matt.1.4.
3. Optional context OSIS string. You can provide a single OSIS string (not a comma-separated list) to give contextual verses for the second argument:

``javascript
const osisToEn = require("./es6/en")
osisToEn("niv-long", "Matt.1.2-Matt.1.3,Matt.1.4") // "Matthew 1:2–3,4"
osisToEn("niv-short", "Matt.1.2-Matt.1.3,Matt.1.4") // "Matt 1:2–3,4"
osisToEn("niv-shortest", "Matt.1.2-Matt.1.3,Matt.1.4") // "Mt 1:2-3, 4"

osisToEn("niv-long", "Matt.1.2-Matt.1.3,Matt.1.4", "Matt.1") // "vv. 2–3,4"
osisToEn("niv-short", "Matt.1.2-Matt.1.3,Matt.1.4", "Matt.1") // "vv. 2–3,4"
osisToEn("niv-shortest", "Matt.1.2-Matt.1.3,Matt.1.4", "Matt.1") // "ver 2-3, 4"
`

#### Output Format Type

The first argument indicates the desired preset output format: you may want long or short book names, for example, or a particular name for a book.

There are three basic formats: ESV, NIV, and NLT, and up to three lengths in each format.

These styles attempt to mimic the style guide of a particular print study Bible. The NIV styles are drawn from the NIV Zondervan Study Bible. The NLT styles are drawn from the NLT Study Bible. The ESV styles are drawn from the ESV Study Bible.

The two longer NIV styles don't include spaces between verses. The short ESV style is generally longer and more formal than the other "short" styles.

$3

osisToParatext.js exports exactly one function that takes a single string argument: the OSIS reference(s) to convert to Paratext. You can include multiple references by separating them with commas.

`javascript
const osisToParatext = require("./es6/osisToParatext")
osisToParatext("Matt.1.2-Matt.1.3,Matt.1.4") // "MAT 1:2-3,MAT 1:4"
`

$3

paratextToOsis.js exports exactly one function that takes a single string argument: the Paratext reference(s) to convert to OSIS. You can include multiple references by separating them with commas. It expects a book name after every comma.

`javascript
const paratextToOsis = require("./es6/paratextToOsis")
paratextToOsis("MAT 1:2-3,MAT 1:4") // "Matt.1.2-Matt.1.3,Matt.1.4"
`

Browser Usage

The browser versions of these scripts (in js/) are compatible with modern browsers and IE10 and later.

`html


`

`html


`

`html


`

`html


`

Build Your Own Output Style

osisFormatter.js provides the foundation to build a variety of output styles. This section explains how to use it.

`javascript
const OsisFormatter = require("./osisFormatter")
const osisFormatter = new OsisFormatter

osisFormatter.setBooks({"Ps": ["Psalm", "Psalms"]})
osisFormatter.setOptions({"^v": "$verses "})
osisFormatter.format("Ps.1.1") // "Psalm 1:1"
osisFormatter.format("Ps.2-Ps.3") // "Psalms 2-3"
osisFormatter.format("Ps.2.2,Ps.2.3", "Ps.2") // "vv 2, 3"
`

In this code, osisFormatter is an OsisFormatter object that exports three functions:

1. setBooks({}) defines the book names to use in the output string.
2. setOptions({}) defines the options to use when formatting the output.
3. format("", "") returns a string formatted according to the books and options you specified.

Above, the first call to .format() returns a single verse (including the default ":" chapter/verse separator). The second call returns a chapter range (using the default "-" range separator). The third call returns a sequence of verses, with the second argument providing context; the vv is the default value for multiple $verses, where $verses is a variable that knows when it's followed by a single verse or multiple verses and adjusts its output accordingly.

$3

.setBooks() takes a single object where each key is an OSIS book abbreviation, and each value is an array of one, two, or three strings. Every book that you want to translate to a readable format should have a key, which means that you probably want to set at least 66 books in this object.

For example:

`javascript
osisFormatter.setBooks({
"Ps": ["Ps.", "Pss.", "Psalms"],
"Gen": ["Gen."]
})
`

The first string is the book name; in English, every book except Psalms will probably only have one element in the array.

The second, optional, string is the book name to use when the OSIS refers to more than one chapter (e.g., Pss. 3-4 or Pss. 5:6-7:8) or when there is a sequence of multiple chapters (e.g., Pss. 3, 4 or Pss. 5:6, 7:8).

The third, optional, string is the book name to use when referring to the complete book on its own:

`javascript
osisFormatter.format("Ps,Gen") // "Psalms; Gen."
`

#### Chapters

In English, for Psalms, you never want to say, "Psalms chapter 23" or "chapter 23"--it's always Psalm 23.

You can achieve this effect by setting an OSIS.$chapters key in the setBooks() object:

`javascript
osisFormatter.setBooks({
"Ps": ["Ps.", "Pss."],
"Ps.$chapters": ["Ps.", "Pss."]
})
osisFormatter.setOptions({
"^c": "$chapters "
})

osisFormatter.format("Ps.1", "Ps") // "Ps. 1"
osisFormatter.format("Ps.1-Ps.2", "Ps") // "Pss. 1-2"
`

By default, these two cases would show ch 1 and chs 1-2. The Ps.$chapters key tells the script to use the values in the array (as usual, the first value is singular, and the second is plural) rather than the default $chapters value.

You can set this value for any book, but in practice you probably only need it for Psalms.

#### Special Ranges and Sequences

You may want to handle certain book ranges and sequences uniquely. For example:

`javascript
osisFormatter.setBooks({
"1Kgs": ["1 Kings"],
"2Kgs": ["2 Kings"],
"1John": ["1 John"],
"2John": ["2 John"],
"3John": ["3 John"],
"1Kgs-2Kgs": ["1-2 Kings"],
"1John,2John": ["1 and 2 John"],
"1John,2John,3John": ["1, 2, and 3 John"]
})
osisFormatter.format("1Kgs-2Kgs,1John,2John,3John") // "1-2 Kings, 1, 2, and 3 John"
osisFormatter.format("1John,2John,3John.1.5") // "1 and 2 John, 3 John 5"
`

This special formatting only occurs when the OSIS specifies no chapters or verses. In the second example, the specificity of 3John.1.5 prevents it from triggering the special formatting.

These special ranges and sequences always use the first item in the array for each key.

$3

.setOptions() takes a single object where each key is the option to set, and the value is the value to set it to. Most options are strings.

The complete list of options appears in OptionsType.

#### Variables

You can set two variables to include in other outputs: $chapters and $verses.

Both $chapters and $verses take a one- or two-element array. The script uses the first element in an array when the chapter or verse appears on its own ("verse 1"), and it uses the second when there is a range or sequence ("verses 1-2", "verses 1, 2").

`javascript
osisFormatter.setBooks({
"Ps": ["Ps.", "Pss."],
"Gen": ["Gen."],
"Ps.$chapters": ["Ps.", "Pss."]
})
osisFormatter.setOptions({
"$chapters": ["chapter", "chapters"],
"$verses": ["verse", "verses"],
"^c": "$chapters "
})

osisFormatter.format("Gen.1", "Gen") // "chapter 1"
osisFormatter.format("Gen.1-Gen.2", "Gen") // "chapters 1-2"
osisFormatter.format("Ps.1", "Ps") // "Ps. 1"
osisFormatter.format("Ps.1-Ps.2", "Ps") // "Pss. 1-2"
`

In this case, the ^c option formats a string that starts with a chapter because it has a starting context. The script then inserts the appropriate value.

The Ps.$chapters array in the setBooks() call works similarly. In this case, we don't want to refer to a Psalm as chapter, but as Ps.

Punctuation rules (separators, ranges, and sequences) can use the variables $b, $c, and $v in their values, which stand for the book, chapter, and verse (respectively) that precede the punctuation. If there is no preceding chapter or verse, then $c and $v return empty strings, respectively.

#### Single-Chapter Books

You can control how to format single-chapter books with the singleChapterFormat key. It takes one of three values:

1. b tries never to use a chapter number.
2. bv (the default) uses a chapter number only when referring to the chapter without a verse.
3. bcv always uses a chapter number when there's a chapter or verse.

`javscript
osisFormatter.setBooks({
"Phlm": ["Philemon"]
})
`

There is also a singleChapterBooks key that takes an array of OSIS book names that only contain one chapter. The default value should be comprehensive; you shouldn't need to change it.

#### Psalms

Psalms have a couple of special cases.

Ps151Format indicates whether to treat the osis Ps151 (i.e., AddPs) as a book (b) or as a chapter in Psalms (bc). The latter, which is the default, generally leads to more comprehensible output:

`javascript
osisFormatter.setBooks({
"Ps": ["Ps.", "Pss."],
"Ps151": ["Psalm 151"],
})
osisFormatter.setOptions({"Ps151Format": "b"})
osisFormatter.format("Ps151.1.5") // "Psalm 151 5"

osisFormatter.setOptions({"Ps151Format": "b", "singleChapterFormat": "bcv"})
osisFormatter.format("Ps151.1.5") // "Psalm 151 1:5"

osisFormatter.setOptions({"Ps151Format": "bc"})
osisFormatter.format("Ps151.1.5") // "Ps. 151:5"
`

You can also set maxPs (default 150) to indicate the number of Psalms, a value that is used in some unusual plural situations. You shouldn't need to change it.

#### Separators

Separators separate individual components of an OSIS: the two . in Gen.1.2. They come in several different types:

1. . is the fallback separator when nothing else is defined. It defaults to a space ( ), which means that all the below separators except c.v also default to a space.
2. c.v separates chapter numbers from verse numbers. It defaults to a colon (:); for example, the colon in Genesis 1:2.
3. b.c separates a preceding book from a chapter number; for example, the space in Genesis 1:2. It defaults to a space.
4. b1.c separates a single-chapter book from the chapter number; for example, the space in Philemon 1:2. It defaults to a space, though you also need to set "singleChapterFormat": "bc" for it to appear; otherwise, you will see a b.v.
5. b.v separates a single-chapter book from a verse number; for example, the space in Philemon 2. It defaults to a space.
6. .c separates a preceding book (b or b1) from a chapter number. It defaults to a space.
7. .v works out to be an alias for b.v if b.v isn't set. Because c.v is set to a colon by default, c.v overrides the value of .v.

A separator with higher specificity overrides a separator with lower specificity. For example, b.v overrides .v, which in turn overrides ..

#### Ranges

Ranges separate start and end OSIS references: the - in Gen.1.2-Gen.3.4. Range options always include the - character and, if they have any other characters, include an optional start type and a required end type.

In these options, a b indicates a book name; a c indicates a chapter number; and a v indicates a verse number.

The basic idea is that you change the range character(s) depending on what the range is joining. You might want a v-v range (between verses in the same chapter) to use a different character from a c-c range (between chapters in the same book).

As with separators, higher-specificity (longer) rules override lower-specificity (shroter) rules; bcv-bcv takes precedence over v-b.

Let's look at the niv-long range rules:

`javascript
osisFormatter.setOptions({
// Default to an em-dash when no other rules match.
"-": "\u2014",
// Gen-Gen.2. An unusual situation, where there's only a start book, and the end part of the range is a chapter in the same book. Without this rule, this situation would lead to Genesis—2, which kind of makes sense, but by adding $chapters after the em-dash, we get the somewhat more-intelligible Genesis—ch. 2. This rule also affects Gen-Gen.1.2, and outputs Genesis—ch. 1:2; if we wanted to do something different in that situation, we could set the b-cv option.
"b-c": "\u2014$chapters ",
// Phlm-Phlm.1.2. Another unusual situation. It results in Philemon—Philemon 2 since we never want to use $chapters for a single-chapter book.
"b-v": "\u2014$b ",
// Gen.1-Gen.1.2. Another unusual situation. Without this rule, the script would output the misleading Genesis 1—2. This rule leads to Genesis 1—1:2.
"c-v": "\u2014$c:",
// Gen.1.2-Gen.3. Without this rule, we would see Genesis 1:2—3. With this rule, we see Genesis 1:2—ch. 3.
"v-c": "\u2014$chapters ",
// Gen.1.2-Gen.3.4. Because the above v-c rule also matches v-cv (resulting in Genesis 1:2—ch. 3:4), we override the v-c rule here so that it uses the standard em-dash: Genesis 1:2—3:4.
"v-cv": "\u2014",
// Gen.1.2-Gen.1.3. Use an en-dash rather than an em-dash for verse ranges within the same chapter: Genesis 1:2–3
"v-v": "\u2013"
})
`

There are over 50 possible range rules in total, but in practice you probably only want to change a few of them. The full list is in OptionsType.

You can omit the first part of the range: -b, -c, and -v are all valid options, for example.

#### Sequences

Sequence separators separate multiple OSIS references or ranges: the , in Gen.1.2,Gen.3.4. Sequence options always include the , character and, if they have any other characters, include an optional start type and a required end type.

In these options, a b indicates a book name; a c indicates a chapter number; and a v indicates a verse number.

The basic idea is that you change the sequence character(s) depending on what is on either side of the sequence. You might want a v,v sequence (between verses in the same chapter) to use a different character from a c,c sequence (between chapters in the same book.)

Let's look at the niv-long sequence rules:

`javascript
osisFormatter.setOptions({
// Default to using a semicolon and a space to separate references in a sequence.
",": "; ",
// Gen,Gen.2. An unusual situation, where the first reference is only a start book, and the second reference is a chapter in the same book. Without this rule, this situation would lead to Genesis; 2.Adding $chapters after the semicolon results in Genesis; ch. 2. This rule also affects Gen,Gen.1.2, and outputs Genesis; ch. 1:2; if we wanted to do something different in that situation, we could set the b,cv option.
"b,c": "; $chapters ",
// Phlm,Phlm.1.2. Another unusual situation. It results in Philemon; Philemon 2 since the default (Philemon; 2) is a little confusing.
"b,v": "; $b ",
// Gen.1,Gen.1.2. Without this rule, we would see Genesis 1; 2. With this rule, we see Genesis 1; 1:2. The $c variable tells the script to insert the current chapter.
"c,v": "; $c:",
// Gen.1.2,Gen.3. Without this rule, we would see Genesis 1:2; 3. With this rule, we see Genesis 1:2; ch. 3.
"v,c": "; $chapters ",
// Gen.1.2,Gen.3.4. Because the above v,c rule also matches v,cv (resulting in Genesis 1:2; ch. 3:4), we override the v,c rule here so that it uses the standard semicolon and space: Genesis 1:2; 3:4.
"v,cv": "; ",
// Gen.1.2,Gen.1.3. When two verses in the same chapter appear in a sequence, separate them with a comma. The NIV style omits a space: Genesis 1:2,3
"v,v": ","
})
`

#### Ending Sequences

You may want to separate the last item in a sequence differently; for example, you might want to format Gen.1.2,Gen.1.3,Gen.1.4 as Genesis 1:2, 3, and 4 rather than the default Genesis 1:2, 3, 4. Ending sequences affect only the last item in the full sequence string provided to .format().

Ending sequences come in two varieties: & for a two-item sequence (Gen.1,Gen.2), and ,& for a three-or-more item sequence (Gen.1,Gen.2,Gen.3,Gen.4).

You enable these ending sequences by setting a & or ,& key in options. If you've set any specific , options, you probably want to recreate similar rules with & and/or ,& variants to avoid inconsistencies.

For example, let's adapt the niv-long options to include & and ,& options:

`javascript
osisFormatter.setOptions({
// Subset of niv-long rules.
",": "; ",
"v,c": "; $chapters ",
"v,cv": "; ",
"v,v": ",",
"$chapters": ["ch.", "chs."],

// Add & variants for two-item sequences. Default to the word "and" surrounded by spaces.
"&": " and ",
// Include the value of $chapters when a chapter follows a verse.
"v&c": " and $chapters ",
// Override the v&c value with the default in this case.
"v&cv": " and ",
// We don't need a v,v option because the default and works fine.

// Add ,& variants for three-or-more item sequences. Default a semicolon, followed by the word "and" surrounded by spaces.
",&": "; and ",
// Include the value of $chapters when a chapter follows a verse.
"v,&c": "; and $chapters ",
// Override the v,&c value with the default in this case.
"v,&cv": " and ",
// Here we do need to include v,v because we want to use a comma rather than a semicolon.
"v,&v": ", and "
})

osisFormatter.format("Gen.1,Gen.3") // "Genesis 1 and 3"
osisFormatter.format("Gen.1.2,Gen.3") // "Genesis 1:2 and ch. 3"
osisFormatter.format("Gen.1.2,Gen.1.4") // "Genesis 1:2 and 4"
osisFormatter.format("Gen.1,Gen.3,Gen.5") // "Genesis 1; 3; and 5"
osisFormatter.format("Gen.1.2,Gen.1.4,Gen.3") // "Genesis 1:2,4; and ch. 3"
osisFormatter.format("Gen.1.2,Gen.1.4,Gen.1.6") // "Genesis 1:2,4, and 6"
osisFormatter.format("Gen,1John,2John,3John") // "Genesis and 1, 2, and 3 John"
osisFormatter.format("Gen,Phlm,1John,2John,3John") // "Genesis; Philemon; and 1, 2, and 3 John"
`

As you can see in the final two examples, a special book sequence (1John,2John,3John becoming 1, 2, and 3 John) counts as a single "book" for the purposes of determining the location of the last sequence separator.

#### Start Contexts

You can specify several options for formatting the start of the output string when providing a start context to .format().

For example, maybe you'd like the string to say verses 2, 3 instead of just 2, 3 if .format("Gen.1.2,Gen.1.3", "Gen.1"), which you can do with osisFormatter.setOptions({"^v": "$verses ", "$verses": ["verse", "verses"]}).

The b1^cv option requires "singleChapterFormat": "bcv". The b1^c option does nothing if "singleChapterFormat": "b".

$3

If you want the raw tokens from the tokenization process because you want to do further processing on them, you can call osisFormatter.tokenize(). It takes the same two arguments as .format() (an OSIS string and an optional context string) and returns an object with a tokens key containing an array of TokenType objects.

The structure of each object is a bit complicated and mostly contains artifacts from internal processing.

`javascript
osisFormatter.setBooks({"Gen": ["Gen."]})
osisFormatter.tokenize("Gen")
/*
{"tokens: [{
"osis": "Gen",
"type": "b",
"parts": [{
"type": "b",
"subType": "",
"b": "Gen",
"laters": []
}],
"laters": [],
"format": "Gen."
}]}
*/
`

Here's a quick explanation:

* osis: The input OSIS string for the token.
* type: The output type, which doesn't necessarily match the input type; with the appropriate context, an input bcv could just be a v.
* parts: An array of parts that compose the token: each one is of type b, c, v, ., -, or ,. The subType can have additional processing information (b.v if a . separates a book and verse, for example). b indicates the ending book context for the token; c and v may also appear and similarly establish ending chapter and verse context. laters lists future part types to help establish plural or singular usage of $chapters and $verses.
* laters: Future token types establish plural or singular usage of $chapters and $verses.
* format: The output string for that token. Joining all tokens' formats with "" produces the same result as calling osisFormatter.format() with the same arguments.

A token can also contain:

* bookRange. If there is a valid book range in books (such as 1John-2John), the key to use is here.
* bookSequence. If there is a valid book sequence in books (such as 1John,2John), the key to use is here.
* subTokens. If there is a bookSequence key, there is also a subTokens array containing the tokens that the sequence encompasses. If you're trying to match your input OSIS string to the output array, be sure to look in subTokens if it exists; otherwise, your counts will be off.
* position. A , token has a & or ,& position if it's the last , token in the array. & indicates there are two OSISes in the sequence, while ,& indicates there are three or more OSISes.

Folder Structure

flow contains the raw Flow (Javascript with type annotations) source.

es6 contains the output Javascript (ES6-compatible) files for use by Node.

js contains browser-compatible Javascript (ES5) files for inclusion in HTML.

test contains about 5,000 Jasmine-style tests.

System Requirements

The es6/ files require an ES6-capable Node.js. I developed it on 6.3.0.

The js/ files require any modern browser or IE10 or higher. They use ES5 syntax.

Development Requirements

If you want to build the project using compile-and-run.sh, here are the dependencies:

* Flow
* Babel to transform Flow to straight Javascript.
* Flow Strip Types to handle Flow type annotations
* Flow ES2015 Preset to produce ES5-compatible Javascript that Webpack transforms to browser-compatible Javascript.
* Webpack to produce browser-compatible Javascript.
* Jasmine for running tests.
* Jasmine-Node to run tests from the CLI.
* Istanbul for calculating test coverage.

Limitations

It doesn't validate that the references exist; it happily formats the nonexistent Gen.99 as Genesis 99`.

It follows the OSIS and Paratext formats strictly; if provided an unexpected book name or format, it throws an exception.

Purpose

My Bible verse reference identifier outputs OSIS, but you may want to transform OSIS back into a human-readable format. For example, if someone enters "genesis 1.2" into a Bible search engine, we could transform it internally to "Gen.1.2" and then output it as "Genesis 1:2".

I also wanted to try Flow and ES6.

Future Directions

I don't have specific plans to develop this library further beyond bug fixes. However, if you're interested, here are some possibilities:

1. Add more conversion target formats (more translations and languages).
2. Use a standard Grunt-based build process.
3. Package it for npm.