Create Dockerfile

Dockerfile

@@ -0,0 +1,165 @@
+# LibreTranslate Dockerfile with all 22 Indian Regional Languages
+FROM libretranslate/libretranslate:latest
+
+# Set environment variables for Indian languages
+ENV LT_HOST=0.0.0.0
+ENV LT_PORT=7860
+ENV LT_THREADS=6
+ENV LT_CHAR_LIMIT=20000
+ENV LT_LOAD_ONLY=en,hi,bn,te,mr,ta,ur,gu,kn,ml,or,pa,as,mai,mag,bh,sa,ne,ks,sd,kok,mni,sat,doi,bo
+
+# Create necessary directories
+USER root
+RUN mkdir -p /app/db /home/libretranslate/.local/share/argos-translate/packages
+RUN chown -R libretranslate:libretranslate /app/db /home/libretranslate/.local/share/argos-translate/packages
+
+# Switch back to libretranslate user
+USER libretranslate
+
+# Pre-install all 22 Indian regional languages plus English
+RUN python -c "
+import argostranslate.package
+import argostranslate.translate
+
+# Download and install language packages
+available_packages = argostranslate.package.get_available_packages()
+
+# All 22 official Indian languages (ISO 639-1/639-3 codes)
+indian_languages = [
+    'hi',  # Hindi
+    'bn',  # Bengali
+    'te',  # Telugu
+    'mr',  # Marathi
+    'ta',  # Tamil
+    'ur',  # Urdu
+    'gu',  # Gujarati
+    'kn',  # Kannada
+    'ml',  # Malayalam
+    'or',  # Odia
+    'pa',  # Punjabi
+    'as',  # Assamese
+    'mai', # Maithili
+    'mag', # Magahi
+    'bh',  # Bhojpuri
+    'sa',  # Sanskrit
+    'ne',  # Nepali
+    'ks',  # Kashmiri
+    'sd',  # Sindhi
+    'kok', # Konkani
+    'mni', # Manipuri
+    'sat', # Santali
+    'doi', # Dogri
+    'bo',  # Bodo
+]
+
+# Create comprehensive language pairs
+language_pairs = []
+
+# Primary focus: English to/from all Indian languages (PRIORITY PAIRS)
+print('Adding English <-> Indian language pairs...')
+for lang in indian_languages:
+    language_pairs.extend([
+        ('en', lang),  # English to Indian language - HIGH PRIORITY
+        (lang, 'en'),  # Indian language to English - HIGH PRIORITY
+    ])
+
+# Add ALL possible inter-Indian language combinations
+print('Adding inter-Indian language pairs...')
+for i, lang1 in enumerate(indian_languages):
+    for j, lang2 in enumerate(indian_languages):
+        if i != j:  # Don't add same language pairs
+            language_pairs.append((lang1, lang2))
+
+# Additional international language pairs with Indian languages
+international_languages = ['es', 'fr', 'de', 'it', 'pt', 'ru', 'zh', 'ja', 'ar', 'ko']
+print('Adding international language pairs with Indian languages...')
+for indian_lang in indian_languages:
+    for intl_lang in international_languages:
+        language_pairs.extend([
+            (indian_lang, intl_lang),  # Indian to International
+            (intl_lang, indian_lang),  # International to Indian
+        ])
+
+# Install available packages with priority system
+installed_count = 0
+failed_count = 0
+priority_installed = 0
+
+print('Starting installation of language models...')
+print(f'Total language pairs to attempt: {len(language_pairs)}')
+print('Priority: English <-> Indian languages first')
+
+# Separate English-Indian pairs for priority installation
+english_indian_pairs = [(f, t) for f, t in language_pairs if 'en' in (f, t)]
+other_pairs = [(f, t) for f, t in language_pairs if 'en' not in (f, t)]
+
+print(f'English-Indian priority pairs: {len(english_indian_pairs)}')
+print(f'Other language pairs: {len(other_pairs)}')
+
+# Install English-Indian pairs first (PRIORITY)
+print('\\n=== Installing Priority Pairs (English <-> Indian) ===')
+for from_code, to_code in english_indian_pairs:
+    package_to_install = next(
+        (pkg for pkg in available_packages 
+         if pkg.from_code == from_code and pkg.to_code == to_code), 
+        None
+    )
+    if package_to_install:
+        try:
+            argostranslate.package.install_from_path(package_to_install.download())
+            print(f'✓ PRIORITY: Successfully installed {from_code}->{to_code}')
+            installed_count += 1
+            priority_installed += 1
+        except Exception as e:
+            print(f'✗ PRIORITY: Failed to install {from_code}->{to_code}: {e}')
+            failed_count += 1
+    else:
+        print(f'⚠ PRIORITY: Package not available: {from_code}->{to_code}')
+        failed_count += 1
+
+# Install other pairs (inter-Indian and international)
+print('\\n=== Installing Additional Pairs (Inter-Indian & International) ===')
+for from_code, to_code in other_pairs:
+    package_to_install = next(
+        (pkg for pkg in available_packages 
+         if pkg.from_code == from_code and pkg.to_code == to_code), 
+        None
+    )
+    if package_to_install:
+        try:
+            argostranslate.package.install_from_path(package_to_install.download())
+            print(f'✓ Successfully installed {from_code}->{to_code}')
+            installed_count += 1
+        except Exception as e:
+            print(f'✗ Failed to install {from_code}->{to_code}: {e}')
+            failed_count += 1
+    else:
+        # Don't print warnings for unavailable inter-language pairs to reduce noise
+        failed_count += 1
+
+print(f'\\n=== Final Installation Summary ===')
+print(f'Priority pairs (English <-> Indian): {priority_installed} installed')
+print(f'Total successfully installed: {installed_count} language pairs')
+print(f'Failed or unavailable: {failed_count} language pairs')
+print(f'\\n🎯 FOCUS: English to Indian Regional Languages Translation Server Ready!')
+print(f'📋 Supported: All available English <-> Indian language pairs')
+print(f'🌐 Bonus: Inter-Indian and international language pairs where available')
+print(f'\\n⚠️  Note: Some regional languages may not be available in Argos Translate yet.')
+print(f'✅ Available languages will be loaded automatically on server start.')
+print(f'🚀 Server optimized for English to Indian regional language translation!')
+"
+
+# Expose port
+EXPOSE 5000
+
+# Health check optimized for Indian languages
+HEALTHCHECK --interval=30s --timeout=15s --start-period=60s --retries=3 \
+    CMD curl -f http://localhost:5000/health || exit 1
+
+# Default command with Indian language optimization
+CMD ["libretranslate", \
+     "--host", "0.0.0.0", \
+     "--port", "7860", \
+     "--threads", "6", \
+     "--char-limit", "20000", \
+     "--load-only", "en,hi,bn,te,mr,ta,ur,gu,kn,ml,or,pa,as,ne,sa"]