├── books.py
└── pysword.py


/books.py:
--------------------------------------------------------------------------------
  1 | def find_book(name):
  2 |     name = name.lower()
  3 |     for testament, books in testaments.iteritems():
  4 |         for num, book in enumerate(books):
  5 |             if book.name_matches(name):
  6 |                 return book
  7 |     return None
  8 | 
  9 | def ref_to_index(book, chapter, verse):
 10 |     '''
 11 |     Computes the index corresponding to a reference.
 12 | 
 13 |     book: a full name of a book
 14 |     chapter, verse: 1-based numbers.
 15 |     '''
 16 |     book = find_book(book)
 17 |     return book.testament, book.get_index_for_ref(chapter, verse)
 18 |     
 19 | 
 20 | 
 21 | class Book(object):
 22 |     def __init__(self, name, osis_name, preferred_abbreviation, chapter_lengths):
 23 |         self.name = name
 24 |         self.osis_name = osis_name
 25 |         self.preferred_abbreviation = preferred_abbreviation
 26 |         self.chapter_lengths = chapter_lengths
 27 |         self.num_chapters = len(chapter_lengths)
 28 | 
 29 |     def __repr__(self): return u'Book(%s)' % self.name
 30 |     
 31 |     def name_matches(self, name):
 32 |         name = name.lower()
 33 |         return name in [self.name.lower(), self.osis_name.lower(), self.preferred_abbreviation.lower()]
 34 | 
 35 |     def get_index_for_ref(self, chapter, verse):
 36 |         # FIXME: it seems that it should be 'verse - 1', but that isn't right.
 37 |         # Maybe I missed something about the offset calculation.
 38 |         return self.offset + self.chapter_offsets[chapter-1] + verse
 39 | 
 40 | 
 41 | # This data came from canon.h in the SWORD repository. Code to transform it to
 42 | # this form can be found in the repository history.
 43 | testaments = {
 44 | 'ot': [
 45 | Book('Genesis', 'Gen', 'Gen', [31, 25, 24, 26, 32, 22, 24, 22, 29, 32, 32, 20, 18, 24, 21, 16, 27, 33, 38, 18, 34, 24, 20, 67, 34, 35, 46, 22, 35, 43, 55, 32, 20, 31, 29, 43, 36, 30, 23, 23, 57, 38, 34, 34, 28, 34, 31, 22, 33, 26]),
 46 | Book('Exodus', 'Exod', 'Exod', [22, 25, 22, 31, 23, 30, 25, 32, 35, 29, 10, 51, 22, 31, 27, 36, 16, 27, 25, 26, 36, 31, 33, 18, 40, 37, 21, 43, 46, 38, 18, 35, 23, 35, 35, 38, 29, 31, 43, 38]),
 47 | Book('Leviticus', 'Lev', 'Lev', [17, 16, 17, 35, 19, 30, 38, 36, 24, 20, 47, 8, 59, 57, 33, 34, 16, 30, 37, 27, 24, 33, 44, 23, 55, 46, 34]),
 48 | Book('Numbers', 'Num', 'Num', [54, 34, 51, 49, 31, 27, 89, 26, 23, 36, 35, 16, 33, 45, 41, 50, 13, 32, 22, 29, 35, 41, 30, 25, 18, 65, 23, 31, 40, 16, 54, 42, 56, 29, 34, 13]),
 49 | Book('Deuteronomy', 'Deut', 'Deut', [46, 37, 29, 49, 33, 25, 26, 20, 29, 22, 32, 32, 18, 29, 23, 22, 20, 22, 21, 20, 23, 30, 25, 22, 19, 19, 26, 68, 29, 20, 30, 52, 29, 12]),
 50 | Book('Joshua', 'Josh', 'Josh', [18, 24, 17, 24, 15, 27, 26, 35, 27, 43, 23, 24, 33, 15, 63, 10, 18, 28, 51, 9, 45, 34, 16, 33]),
 51 | Book('Judges', 'Judg', 'Judg', [36, 23, 31, 24, 31, 40, 25, 35, 57, 18, 40, 15, 25, 20, 20, 31, 13, 31, 30, 48, 25]),
 52 | Book('Ruth', 'Ruth', 'Ruth', [22, 23, 18, 22]),
 53 | Book('I Samuel', '1Sam', '1Sam', [28, 36, 21, 22, 12, 21, 17, 22, 27, 27, 15, 25, 23, 52, 35, 23, 58, 30, 24, 42, 15, 23, 29, 22, 44, 25, 12, 25, 11, 31, 13]),
 54 | Book('II Samuel', '2Sam', '2Sam', [27, 32, 39, 12, 25, 23, 29, 18, 13, 19, 27, 31, 39, 33, 37, 23, 29, 33, 43, 26, 22, 51, 39, 25]),
 55 | Book('I Kings', '1Kgs', '1Kgs', [53, 46, 28, 34, 18, 38, 51, 66, 28, 29, 43, 33, 34, 31, 34, 34, 24, 46, 21, 43, 29, 53]),
 56 | Book('II Kings', '2Kgs', '2Kgs', [18, 25, 27, 44, 27, 33, 20, 29, 37, 36, 21, 21, 25, 29, 38, 20, 41, 37, 37, 21, 26, 20, 37, 20, 30]),
 57 | Book('I Chronicles', '1Chr', '1Chr', [54, 55, 24, 43, 26, 81, 40, 40, 44, 14, 47, 40, 14, 17, 29, 43, 27, 17, 19, 8, 30, 19, 32, 31, 31, 32, 34, 21, 30]),
 58 | Book('II Chronicles', '2Chr', '2Chr', [17, 18, 17, 22, 14, 42, 22, 18, 31, 19, 23, 16, 22, 15, 19, 14, 19, 34, 11, 37, 20, 12, 21, 27, 28, 23, 9, 27, 36, 27, 21, 33, 25, 33, 27, 23]),
 59 | Book('Ezra', 'Ezra', 'Ezra', [11, 70, 13, 24, 17, 22, 28, 36, 15, 44]),
 60 | Book('Nehemiah', 'Neh', 'Neh', [11, 20, 32, 23, 19, 19, 73, 18, 38, 39, 36, 47, 31]),
 61 | Book('Esther', 'Esth', 'Esth', [22, 23, 15, 17, 14, 14, 10, 17, 32, 3]),
 62 | Book('Job', 'Job', 'Job', [22, 13, 26, 21, 27, 30, 21, 22, 35, 22, 20, 25, 28, 22, 35, 22, 16, 21, 29, 29, 34, 30, 17, 25, 6, 14, 23, 28, 25, 31, 40, 22, 33, 37, 16, 33, 24, 41, 30, 24, 34, 17]),
 63 | Book('Psalms', 'Ps', 'Ps', [6, 12, 8, 8, 12, 10, 17, 9, 20, 18, 7, 8, 6, 7, 5, 11, 15, 50, 14, 9, 13, 31, 6, 10, 22, 12, 14, 9, 11, 12, 24, 11, 22, 22, 28, 12, 40, 22, 13, 17, 13, 11, 5, 26, 17, 11, 9, 14, 20, 23, 19, 9, 6, 7, 23, 13, 11, 11, 17, 12, 8, 12, 11, 10, 13, 20, 7, 35, 36, 5, 24, 20, 28, 23, 10, 12, 20, 72, 13, 19, 16, 8, 18, 12, 13, 17, 7, 18, 52, 17, 16, 15, 5, 23, 11, 13, 12, 9, 9, 5, 8, 28, 22, 35, 45, 48, 43, 13, 31, 7, 10, 10, 9, 8, 18, 19, 2, 29, 176, 7, 8, 9, 4, 8, 5, 6, 5, 6, 8, 8, 3, 18, 3, 3, 21, 26, 9, 8, 24, 13, 10, 7, 12, 15, 21, 10, 20, 14, 9, 6]),
 64 | Book('Proverbs', 'Prov', 'Prov', [33, 22, 35, 27, 23, 35, 27, 36, 18, 32, 31, 28, 25, 35, 33, 33, 28, 24, 29, 30, 31, 29, 35, 34, 28, 28, 27, 28, 27, 33, 31]),
 65 | Book('Ecclesiastes', 'Eccl', 'Eccl', [18, 26, 22, 16, 20, 12, 29, 17, 18, 20, 10, 14]),
 66 | Book('Song of Solomon', 'Song', 'Song', [17, 17, 11, 16, 16, 13, 13, 14]),
 67 | Book('Isaiah', 'Isa', 'Isa', [31, 22, 26, 6, 30, 13, 25, 22, 21, 34, 16, 6, 22, 32, 9, 14, 14, 7, 25, 6, 17, 25, 18, 23, 12, 21, 13, 29, 24, 33, 9, 20, 24, 17, 10, 22, 38, 22, 8, 31, 29, 25, 28, 28, 25, 13, 15, 22, 26, 11, 23, 15, 12, 17, 13, 12, 21, 14, 21, 22, 11, 12, 19, 12, 25, 24]),
 68 | Book('Jeremiah', 'Jer', 'Jer', [19, 37, 25, 31, 31, 30, 34, 22, 26, 25, 23, 17, 27, 22, 21, 21, 27, 23, 15, 18, 14, 30, 40, 10, 38, 24, 22, 17, 32, 24, 40, 44, 26, 22, 19, 32, 21, 28, 18, 16, 18, 22, 13, 30, 5, 28, 7, 47, 39, 46, 64, 34]),
 69 | Book('Lamentations', 'Lam', 'Lam', [22, 22, 66, 22, 22]),
 70 | Book('Ezekiel', 'Ezek', 'Ezek', [28, 10, 27, 17, 17, 14, 27, 18, 11, 22, 25, 28, 23, 23, 8, 63, 24, 32, 14, 49, 32, 31, 49, 27, 17, 21, 36, 26, 21, 26, 18, 32, 33, 31, 15, 38, 28, 23, 29, 49, 26, 20, 27, 31, 25, 24, 23, 35]),
 71 | Book('Daniel', 'Dan', 'Dan', [21, 49, 30, 37, 31, 28, 28, 27, 27, 21, 45, 13]),
 72 | Book('Hosea', 'Hos', 'Hos', [11, 23, 5, 19, 15, 11, 16, 14, 17, 15, 12, 14, 16, 9]),
 73 | Book('Joel', 'Joel', 'Joel', [20, 32, 21]),
 74 | Book('Amos', 'Amos', 'Amos', [15, 16, 15, 13, 27, 14, 17, 14, 15]),
 75 | Book('Obadiah', 'Obad', 'Obad', [21]),
 76 | Book('Jonah', 'Jonah', 'Jonah', [17, 10, 10, 11]),
 77 | Book('Micah', 'Mic', 'Mic', [16, 13, 12, 13, 15, 16, 20]),
 78 | Book('Nahum', 'Nah', 'Nah', [15, 13, 19]),
 79 | Book('Habakkuk', 'Hab', 'Hab', [17, 20, 19]),
 80 | Book('Zephaniah', 'Zeph', 'Zeph', [18, 15, 20]),
 81 | Book('Haggai', 'Hag', 'Hag', [15, 23]),
 82 | Book('Zechariah', 'Zech', 'Zech', [21, 13, 10, 14, 11, 15, 14, 23, 17, 12, 17, 14, 9, 21]),
 83 | Book('Malachi', 'Mal', 'Mal', [14, 17, 18, 6]),
 84 | ],
 85 | 'nt': [
 86 | Book('Matthew', 'Matt', 'Matt', [25, 23, 17, 25, 48, 34, 29, 34, 38, 42, 30, 50, 58, 36, 39, 28, 27, 35, 30, 34, 46, 46, 39, 51, 46, 75, 66, 20]),
 87 | Book('Mark', 'Mark', 'Mark', [45, 28, 35, 41, 43, 56, 37, 38, 50, 52, 33, 44, 37, 72, 47, 20]),
 88 | Book('Luke', 'Luke', 'Luke', [80, 52, 38, 44, 39, 49, 50, 56, 62, 42, 54, 59, 35, 35, 32, 31, 37, 43, 48, 47, 38, 71, 56, 53]),
 89 | Book('John', 'John', 'John', [51, 25, 36, 54, 47, 71, 53, 59, 41, 42, 57, 50, 38, 31, 27, 33, 26, 40, 42, 31, 25]),
 90 | Book('Acts', 'Acts', 'Acts', [26, 47, 26, 37, 42, 15, 60, 40, 43, 48, 30, 25, 52, 28, 41, 40, 34, 28, 41, 38, 40, 30, 35, 27, 27, 32, 44, 31]),
 91 | Book('Romans', 'Rom', 'Rom', [32, 29, 31, 25, 21, 23, 25, 39, 33, 21, 36, 21, 14, 23, 33, 27]),
 92 | Book('I Corinthians', '1Cor', '1Cor', [31, 16, 23, 21, 13, 20, 40, 13, 27, 33, 34, 31, 13, 40, 58, 24]),
 93 | Book('II Corinthians', '2Cor', '2Cor', [24, 17, 18, 18, 21, 18, 16, 24, 15, 18, 33, 21, 14]),
 94 | Book('Galatians', 'Gal', 'Gal', [24, 21, 29, 31, 26, 18]),
 95 | Book('Ephesians', 'Eph', 'Eph', [23, 22, 21, 32, 33, 24]),
 96 | Book('Philippians', 'Phil', 'Phil', [30, 30, 21, 23]),
 97 | Book('Colossians', 'Col', 'Col', [29, 23, 25, 18]),
 98 | Book('I Thessalonians', '1Thess', '1Thess', [10, 20, 13, 18, 28]),
 99 | Book('II Thessalonians', '2Thess', '2Thess', [12, 17, 18]),
100 | Book('I Timothy', '1Tim', '1Tim', [20, 15, 16, 16, 25, 21]),
101 | Book('II Timothy', '2Tim', '2Tim', [18, 26, 17, 22]),
102 | Book('Titus', 'Titus', 'Titus', [16, 15, 15]),
103 | Book('Philemon', 'Phlm', 'Phlm', [25]),
104 | Book('Hebrews', 'Heb', 'Heb', [14, 18, 19, 16, 14, 20, 28, 13, 28, 39, 40, 29, 25]),
105 | Book('James', 'Jas', 'Jas', [27, 26, 18, 17, 20]),
106 | Book('I Peter', '1Pet', '1Pet', [25, 25, 22, 19, 14]),
107 | Book('II Peter', '2Pet', '2Pet', [21, 22, 18]),
108 | Book('I John', '1John', '1John', [10, 29, 24, 21, 21]),
109 | Book('II John', '2John', '2John', [13]),
110 | Book('III John', '3John', '3John', [14]),
111 | Book('Jude', 'Jude', 'Jude', [25]),
112 | Book('Revelation of John', 'Rev', 'Rev', [20, 29, 22, 11, 14, 17, 17, 13, 21, 11, 19, 17, 18, 20, 8, 21, 18, 24, 21, 15, 27, 21]),
113 | ],
114 | }
115 | 
116 | # Compute index offsets and add other data
117 | # FIXME: this is still a little hairy.
118 | for testament, books in testaments.iteritems():
119 |     idx = 1 # start after the testament heading
120 |     for book in books:
121 |         book.testament = testament
122 |         book.offset = idx
123 |         offset = 1 # start after the book heading
124 |         book.chapter_offsets = []
125 |         for chapter_len in book.chapter_lengths:
126 |             offset += 1 # skip the chapter heading
127 |             book.chapter_offsets.append(offset)
128 |             offset += chapter_len
129 |         idx += offset
130 | 
131 | 


--------------------------------------------------------------------------------
/pysword.py:
--------------------------------------------------------------------------------
  1 | #!/usr/bin/env python
  2 | 
  3 | # A native Python implementation of the SWORD Project Bible Reader
  4 | # Currently only ztext Bible modules are implemented.
  5 | 
  6 | # * ztext format documentation
  7 | # I'll use Python's struct module's format strings.
  8 | # See http://docs.python.org/lib/module-struct.html
  9 | # Take the Old Testament (OT) for example. Three files:
 10 | #
 11 | #  - ot.bzv: Maps verses to character ranges in compressed buffers.
 12 | #    10 bytes ('<IIH') for each verse in the Bible:
 13 | #    - buffer_num (I): which compressed buffer the verse is located in
 14 | #    - verse_start (I): the location in the uncompressed buffer where the verse begins
 15 | #    - verse_len (H): length of the verse, in uncompressed characters
 16 | #    These 10-byte records are densely packed, indexed by VerseKey 'Indicies' (docs later).
 17 | #    So the record for the verse with index x starts at byte 10*x.
 18 | #
 19 | #  - ot.bzs: Tells where the compressed buffers start and end.
 20 | #    12 bytes ('<III') for each compressed buffer:
 21 | #    - offset (I): where the compressed buffer starts in the file
 22 | #    - size (I): the length of the compressed data, in bytes
 23 | #    - uc_size (I): the length of the uncompressed data, in bytes (unused)
 24 | #    These 12-byte records are densely packed, indexed by buffer_num (see previous).
 25 | #    So the record for compressed buffer buffer_num starts at byte 12*buffer_num.
 26 | #
 27 | #  - ot.bzz: Contains the compressed text. Read 'size' bytes starting at 'offset'.
 28 | #
 29 | #  NT is analogous.
 30 | #
 31 | # Example usage:
 32 | #  python pysword.py esv 1pet 2 9
 33 | 
 34 | import os
 35 | modules_path = os.environ["HOME"]+"/.sword/modules/texts/ztext"
 36 | from books import ref_to_index, testaments, find_book, Book
 37 | 
 38 | import struct, zlib
 39 | from os.path import join as path_join
 40 | 
 41 | class ZModule(object):
 42 |     def __init__(self, module):
 43 |         self.module = module
 44 |         self.files = {
 45 |             'ot': self.get_files('ot'),
 46 |             'nt': self.get_files('nt')
 47 |             }
 48 |    
 49 |     def get_files(self, testament):
 50 |         '''Given a testament ('ot' or 'nt'), returns a tuple of files
 51 |         (verse_to_buf, buf_to_loc, text)
 52 |         '''
 53 |         base = path_join(modules_path, self.module)
 54 |         v2b_name, b2l_name, text_name = [path_join(base, '%s.bz%s' % (testament, code))
 55 |                                          for code in ('v', 's', 'z')]
 56 |         return [open(name, 'rb') for name in (v2b_name, b2l_name, text_name)]
 57 | 
 58 |     def text_for_index(self, testament, index):
 59 |         '''Get the text for a given index.'''
 60 |         verse_to_buf, buf_to_loc, text = self.files[testament]
 61 | 
 62 |         # Read the verse record.
 63 |         verse_to_buf.seek(10*index)
 64 |         buf_num, verse_start, verse_len = struct.unpack('<IIH', verse_to_buf.read(10))
 65 |        
 66 |         uncompressed_text = self.uncompressed_text(testament, buf_num)
 67 |         return uncompressed_text[verse_start:verse_start+verse_len]
 68 | 
 69 |     def uncompressed_text(self, testament, buf_num):
 70 |         verse_to_buf, buf_to_loc, text = self.files[testament]
 71 | 
 72 |         # Determine where the compressed data starts and ends.
 73 |         buf_to_loc.seek(buf_num*12)
 74 |         offset, size, uc_size = struct.unpack('<III', buf_to_loc.read(12))
 75 | 
 76 |         # Get the compressed data.
 77 |         text.seek(offset)
 78 |         compressed_data = text.read(size)
 79 |         return zlib.decompress(compressed_data)
 80 | 
 81 |     def text_for_ref(self, book, chapter, verse):
 82 |         '''Get the text for a given reference'''
 83 |         chapter, verse = int(chapter), int(verse)
 84 |         testament, idx = ref_to_index(book, chapter, verse)
 85 |         return self.text_for_index(testament, idx)
 86 | 
 87 |     def all_verses_in_testament(self, testament):
 88 |         books = testaments[testament]
 89 |         for book in books:
 90 |             for verse in self.all_verses_in_book(book):
 91 |                 yield verse
 92 | 
 93 |     def all_verses_in_book(self, book):
 94 |         if not isinstance(book, Book): book = find_book(book)
 95 |         for chapter, verses in enumerate(book.chapter_lengths):
 96 |             chapter = chapter + 1
 97 |             for verse in range(verses):
 98 |                 verse = verse+1
 99 |                 yield (book, chapter, verse,
100 |                        self.text_for_index(book.testament,
101 |                                            book.get_index_for_ref(chapter, verse)))
102 |         
103 | 
104 | if __name__=='__main__':
105 |     import sys
106 |     mod_name, book, chapter, verse = sys.argv[1:]
107 |    
108 |     module = ZModule(mod_name)
109 |     print module.text_for_ref(book, chapter, verse)
110 |     
111 |     
112 | 


--------------------------------------------------------------------------------