Inspired by https://github.com/layeh/kjv and https://github.com/lukesmithxyz/kjv

To Use

• jpm --no-core quickbin verse-reader.janet build/kjv builds

• build/kjv gen 1:1 to show verse. It shows full books, allows partial names phil, matt etc.

• racket interpreted-bible-parse.rkt gen 1 to validate new data. #ToDo: Move data Cleaning to Janet. #Todo File's currently hardcoded, have both versions accept a file argument and and use that for program name, usage etc.

To Do

• many #ToDo:s littered about in bible-parse.rkt, which should also be ported to Janet.
• add binaries and old books with verses
• change BoM etc. abbreviations to remove spaces: 1 ne -> 1ne

Texts

From layeh:

• Clementine Vulgate
• O.T. Septuagint and SBL Greek New Testament

From Luke Smith:

• KJV + Apocrophya (n.b. I eliminated 1 line in Sirach: "A Prayer to...")

From Mormon Documentation Project:

• Book of Mormon (and soon others) (n.b. Nephi 3 and 4 are corrupted in almost all online sources; my initial attempts to scrape the whole book failed here too. I was able to easily fix the .csv here though!)

Performance

With hyperfine:

1.11.24 with time:

• current: time ./kjvr gen 1
• reference in bash/awk:time PAGER=cat kjv gen 1
• interpreted with file: time racket interpreted-bible-parse.rkt gen 1
• Ryan's line version note, extremely fast fast compilation!
• Ryan's comp parse version, 3-4s compilation and great execution. Now the base.
• my old comp time attempt without #:prefab

For Racket

• raco exe -o kjv bible-parse.rkt builds

• ./kjv gen 1:1 to show verse. It shows full books, allows partial names phil, matt etc.

• racket interpreted-bible-parse.rkt gen 1 to validate new data. #ToDo: File's currently hardcoded, have both versions accept a file argument and and use that for program name, usage etc.

min-test.rkt (basically a hello world) gets Time (mean ± σ): 86.8 ms ± 2.8 ms [User: 61.5 ms, System: 25.1 ms] so the current's 220ms is only 3x slower than Racket's fastest. My various refactors had identical performance, although inline-test.rkt compiles faster to an only 51mb static binary. Racket packs everything into an ELF container, which apparently takes 80ms to initialize, I don't yet understand why Typed Racket has a slower initialization.