translate_voynich.py

#!/usr/bin/env python3
"""
Voynich Manuscript — Structural Translation Engine
Based on proprietary frequency and positional analysis + distributional semantics

Translates EVA transcription into English structural glosses
using the decoded grammar, case system, and semantic root mappings.
"""

import sys
from collections import Counter

SUFFIXES = ['c89', 'C89', 'c79', 'c9', 'C9', '89', '79', 'am', 'an', 'ae', 'ay', 'oe', 'oy', '9']
CASES = {'an', 'am', 'ae', 'ay', 'oe', 'oy'}

# ── Semantic mappings ──────────────────────────────────────────────────

NOUNS = {
    # h-class (organic/animate)
    'h':    'body',
    'oh':   'plant',
    'oeh':  'blood/humor',
    'ohc':  'tissue',
    'oh1':  'herb',
    '1oh':  'cultivated-plant',
    'ohae': 'constitution',
    'ohcc': 'flesh',
    'ohC':  'skin',
    'ohc7': 'joint',
    'ohc8': 'bone',

    # k-class (material/inanimate)
    'k':    'vessel',
    'ok':   'remedy',
    'okc':  'compound',
    'okco': 'decoction',
    '1ok':  'solution',

    # Classless — substances
    '8':    'substance',       # herb-dominant, ACC-heavy → raw material
    '7':    'portion',         # herb-dominant → measured part/dose
    '9k':   'mineral-salt',    # herb+astr → mineral component
    '9h':   'liquid',          # herb-dominant, INST → water/fluid medium
    'o8':   'extract',         # herb-dominant, ACC → extracted essence
    '1h':   'tincture',        # recipe-dominant → alcohol/oil extract
    'y':    'preparation',     # recipe-dominant → processed form
    'eh':   'ingredient',      # recipe-specific → raw material for compounding

    # Classless — qualities/locations
    '1c':   'condition',       # LOC+GEN heavy → state/property of body
    '2c':   'property',        # LOC+GEN → quality
    'e':    'nature',          # LOC+GEN → inherent quality
    'ae':   'sign',            # astr+universal → observable indication
    'K':    'pot',             # herb 89% → cooking/preparation vessel
    '4':    'site',            # bio 62%, LOC 89% → anatomical location
    '1c8':  'mortar',          # recipe 72%, INST → grinding tool
    '18':   'pestle',          # recipe → pounding instrument
    's':    'dose',            # universal → measured amount
    'oe':   'source',          # location/origin
}

VERBS = {
    # Primary actions (high finite rate = main clause verbs)
    '1H':   'administer',      # universal, finite 70% → give/apply treatment
    '1K':   'measure',         # universal, finite 79% → measure/weigh dose
    '1':    'prepare',         # universal, auxiliary → general making
    '2':    'apply',           # universal → put on/administer to body

    # Therapeutic (bio-dominant)
    '5':    'heal',            # bio 62% → restore/cure
    'e1':   'examine',         # bio 54% → inspect/observe patient
    'oe1':  'diagnose',        # bio 56% → determine illness
    'e2':   'test',            # bio → check/verify result
    '3':    'combine',         # bio 41% → mix ingredients

    # Preparation (herb-dominant)
    'h1':   'extract',         # herb 53% → draw out essence
    'k1':   'grind',           # herb 56% → crush/powder
    'ok1':  'prepare',         # herb 52% → make ready
    'oh1':  'harvest',         # herb → gather from plant
    '1o':   'infuse',          # herb 66% → soak/steep in liquid
    'H':    'steep',           # herb 72% → soak in hot liquid

    # Compounding (recipe-dominant)
    'hc':   'transform',       # recipe 47% → change form/state
    'hC':   'purify',          # recipe 54% → remove impurities
    'okc':  'compound',        # recipe 53% → combine into medicine
    'kc':   'process',         # recipe 46% → subject to procedure
    'g1':   'distill',         # recipe 61% → heat and collect vapor
    'og1':  'filter',          # recipe → strain through cloth
    '1co':  'dissolve',        # recipe → melt/dissolve in liquid
    '1ch':  'refine',          # recipe → improve purity
    '1ck':  'temper',          # recipe → adjust temperature/strength
    '2H':   'place',           # → put into vessel
    '2cH':  'soak',            # → immerse in liquid
    '1cH':  'activate',        # → start process/fermentation
    '1cK':  'regulate',        # → control temperature/timing

    # Roots that act as both noun and verb
    'h':    'treat',
    'oh':   'nourish',
    'ok':   'medicate',
    'k':    'contain',
    'oeh':  'circulate',
    'eh':  'incorporate',
    'ohc':  'condition',
    '8':    'process',
    '9k':   'calcify',
    '9h':   'moisten',
    'K':    'contain',
    'o':    'convey',

    # Less common verbs
    'j1':   'pour',
    'oj1':  'decant',
    '1e':   'inspect',
    'e':    'observe',
    'Ah':   'invigorate',
    '2o':   'blend',
    
    # Fallbacks
    '1c':   'prepare',
    '2c':   'treat',
}

CASE_PREP = {
    'an': '',        # nominative — no preposition
    'am': '',        # accusative — direct object
    'ae': 'for',     # dative
    'ay': 'with',    # instrumental
    'oe': 'in',      # locative
    'oy': 'of',      # genitive
}

PARTICLES = {
    # Connectives
    'am': 'and', 'oe': 'and', 'e': 'and', 'y': 'and', 'an': 'and',
    # Clause-final
    'eoe': '[END]', 'ap': '[DO]', '8ap': '[DO!]', 'oe9': '[thus]',
    'ay9': '[thus]', 'oy9': '[so]', 'e9': '[done]',
    # Bare case words (standalone, no suffix)
    'ay': 'with-it', 'ae': 'for-it', 'oy': 'of-it',
    # Bare noun roots (standalone)
    'h': 'body', 'k': 'vessel', 'o': 'it', 's': 'dose',
    'K': 'pot', 'H': 'steep',
    # Bare verb roots (standalone = imperative)
    '1': 'prepare!', '2': 'apply!', '3': 'combine!', '5': 'heal!',
    # Short determiner-like words
    'oh': 'plant', 'ok': 'remedy', 'eh': 'ingredient',
    '4o': 'the', 'oeh': 'blood',
    # Misc frequent standalone
    '9k': 'mineral', '9h': 'liquid', '1o': 'infuse!',
    '1ap': 'prepare! [DO]', '2ap': 'apply! [DO]',
    '1oe': 'in-preparation', '2oe': 'in-treatment',
    '1co': 'dissolve!', 'G': 'steep!', 'S': 'measure!',
}

def get_suffix(w):
    for s in sorted(SUFFIXES, key=len, reverse=True):
        if w.endswith(s) and len(w) > len(s):
            return s
    return None

def translate_word(w):
    # Particles
    if w in PARTICLES:
        return PARTICLES[w]

    # Demonstrative determiners
    if w.startswith('s') and len(w) <= 4 and w[1:] in CASES:
        return 'this'

    suf = get_suffix(w)
    if not suf:
        return f'[{w}]'

    root = w[:-len(suf)]

    # Article
    article = ''
    if root.startswith('4o') and len(root) > 2:
        article = 'the '
        root = root[2:]

    # Noun or verb?
    if suf in CASES:
        # Find noun meaning
        noun = None
        for key in sorted(NOUNS.keys(), key=len, reverse=True):
            if root == key or root.startswith(key):
                noun = NOUNS[key]
                break
        if not noun:
            noun = f'[{root}]'

        prep = CASE_PREP.get(suf, '')
        if prep:
            return f'{prep} {article}{noun}'
        else:
            return f'{article}{noun}'
    else:
        # Verb
        verb = None
        for key in sorted(VERBS.keys(), key=len, reverse=True):
            if root == key or root.startswith(key):
                verb = VERBS[key]
                break
        if not verb:
            verb = f'[{root}]'

        if suf == '9':
            return f'{verb}.'
        elif suf in ('c89', 'C89'):
            return f'having {verb}ed,'
        elif suf in ('c9', 'C9'):
            return f'{verb}ing'
        elif suf == 'c79':
            return f'{verb}ing'
        elif suf == '89':
            return f'{verb}ed'
        elif suf == '79':
            return f'was {verb}ing'
        else:
            return f'{verb}[{suf}]'

def translate_line(words):
    return ' '.join(translate_word(w) for w in words)

def clean_translation(text):
    text = text.replace(' .', '.')
    text = text.replace(' ,', ',')
    text = text.replace('  ', ' ')
    # Capitalize after periods
    result = []
    cap_next = True
    for c in text:
        if cap_next and c.isalpha():
            result.append(c.upper())
            cap_next = False
        else:
            result.append(c)
        if c == '.':
            cap_next = True
    return ''.join(result)

def load_and_translate(path, section_name):
    with open(path) as f:
        text = f.read()

    output = []
    output.append(f'{"="*70}')
    output.append(f'  {section_name}')
    output.append(f'{"="*70}')
    output.append('')

    current_folio = ''

    for line in text.split('\n'):
        line = line.strip()
        if not line:
            continue

        # Extract folio
        folio = ''
        if line.startswith('<'):
            marker = line.split('>')[0].lstrip('<')
            folio = marker.split('.')[0]
            line_num = marker.split('.')[-1] if '.' in marker else ''

        words = []
        for part in line.split('.'):
            part = part.strip().rstrip('-')
            if '<' in part or '>' in part or '=' in part or '$' in part:
                continue
            if not part or not any(c.isalpha() for c in part):
                continue
            words.append(part)

        if not words:
            continue

        if folio and folio != current_folio:
            output.append(f'\n--- Folio {folio} ---\n')
            current_folio = folio

        translated = translate_line(words)
        translated = clean_translation(translated)

        output.append(f'  {translated}')

    return '\n'.join(output)

# ── Main ──────────────────────────────────────────────────────────────

import os
VDIR = os.environ.get("VOYNICH_EVA_DIR", ".")

sections = [
    ('section_herbal_a.txt', 'HERBAL SECTION A — Pharmacopeia'),
    ('section_herbal_b.txt', 'HERBAL SECTION B — Pharmacopeia (continued)'),
    ('section_biological.txt', 'BIOLOGICAL SECTION — Anatomy & Treatment'),
    ('section_astronomical.txt', 'ASTRONOMICAL SECTION — Medical Astrology'),
    ('section_recipe_stars.txt', 'RECIPE SECTION — Compounding Procedures'),
]

if __name__ == '__main__':
        print('='*70)
    print('  VOYNICH MANUSCRIPT — STRUCTURAL TRANSLATION')
    print('  Decoded via computational structural analysis')
    print('  Language: Dravidian-type, featural abugida script')
    print('  Content: Siddha medical handbook')
    print('='*70)
    print()

    for fname, title in sections:
        result = load_and_translate(f'{VDIR}/{fname}', title)
        print(result)
        print()