constants.py

low_level_subfamilies = {
    # Germanic
    "Afrikaans": "West Germanic",
    "Bavarian": "Austro-Bavarian",
    "Danish": "North Germanic",
    "Dutch": "West Germanic",
    "English": "West Germanic",
    "German": "West Germanic",
    "Icelandic": "North Germanic",
    "Low Saxon": "West Germanic",
    "Luxembourgish": "West Germanic",
    "Norwegian (Bokmal)": "North Germanic",
    "Norwegian (Nynorsk)": "North Germanic",
    "Scots": "West Germanic",
    "Swedish": "North Germanic",
    "West Frisian": "West Germanic",

    # Romance
    "Aragonese": "Ibero-Romance",
    "Asturian":  "Ibero-Romance",
    "Catalan":   "Occitano-Romance",
    "French":    "Gallo-Romance",
    "Galician":  "Ibero-Romance",
    "Italian":   "Italo-Dalmatian",
    "Lombard":   "Gallo-Italic",
    "Occitan":   "Occitano-Romance",
    "Piedmontese":"Gallo-Italic",
    "Portuguese":"Ibero-Romance",
    "Romanian":  "Eastern Romance",
    "Sicilian":  "Italo-Dalmatian",
    "Spanish":   "Ibero-Romance",

    # Slavic
    "Belarusian":"East Slavic",
    "Bosnian":   "South Slavic",
    "Bulgarian": "South Slavic",
    "Croatian":  "South Slavic",
    "Czech":     "West Slavic",
    "Macedonian":"South Slavic",
    "Polish":    "West Slavic",
    "Russian":   "East Slavic",
    "Serbian":   "South Slavic",
    "Serbo-Croatian":"South Slavic",
    "Slovak":    "West Slavic",
    "Slovenian": "South Slavic",
    "Ukrainian": "East Slavic",

    # Celtic
    "Breton": "Brythonic",
    "Welsh":  "Brythonic",
    "Irish":  "Goidelic",

    # Baltic
    "Latvian": "Baltic",
    "Lithuanian": "Baltic",

    # Indo-Aryan (Indic)
    "Bengali":  "Eastern Indo-Aryan",
    "Bishnupriya Manipuri": "Eastern Indo-Aryan",
    "Gujarati": "Western Indo-Aryan",
    "Hindi":    "Central Indo-Aryan",
    "Marathi":  "Southern Indo-Aryan",
    "Nepali":   "Northern Indo-Aryan",
    "Punjabi":  "North-western Indo-Aryan",
    "Urdu":     "Central Indo-Aryan",
    "Western Punjabi": "North-western Indo-Aryan",

    # Iranian
    "Persian (Farsi)": "South-western Iranian",
    "Tajik":          "South-western Iranian",

    # Turkic
    "Azerbaijani": "Oghuz",
    "South Azerbaijani": "Oghuz",
    "Turkish":     "Oghuz",
    "Bashkir":     "Kipchak",
    "Kazakh":      "Kipchak",
    "Kirghiz":     "Kipchak",
    "Tatar":       "Kipchak",
    "Crimean Tatar":"Kipchak",
    "Chuvash":     "Oghur",
    "Uzbek":       "Karluk",

    # Uralic
    "Estonian": "Finnic",
    "Finnish":  "Finnic",
    "Hungarian":"Ugric",

    # Dravidian
    "Kannada":  "Southern Dravidian",
    "Malayalam":"Southern Dravidian",
    "Tamil":    "Southern Dravidian",
    "Telugu":   "South-Central Dravidian",

    # Sinitic (Chinese)
    "Chinese (Simplified)":   "Mandarin",
    "Chinese (Traditional)":  "Mandarin",
    "Min Nan Chinese":        "Southern Min",

    # Other Sino-Tibetan
    "Burmese": "Burmish",
    "Newar":   "Newaric",

    # Japonic / Koreanic
    "Japanese": "Japonic",
    "Korean":   "Koreanic",

    # Caucasian & Kartvelian
    "Chechen":  "Nakh",
    "Georgian": "Kartvelian",

    # Austronesian
    "Cebuano":     "Central Philippine",
    "Tagalog":     "Central Philippine",
    "Waray-Waray": "Central Philippine",
    "Indonesian":  "Malayic",
    "Malay":       "Malayic",
    "Minangkabau": "Malayic",
    "Javanese":    "Javanese",
    "Sundanese":   "Sundic",
    "Malagasy":    "East Barito",

    # Philippine & Oceanic already covered above

    # Vietic & MSEA
    "Vietnamese": "Vietic",

    # Altaic hypotheses excluded; Kipchak/Oghuz etc already above

    # Afro-Asiatic
    "Arabic":          "Central Semitic",
    "Egyptian Arabic": "Central Semitic",
    "Hebrew":          "North-west Semitic",

    # Niger-Congo
    "Swahili": "Sabaki",
    "Yoruba":  "Yoruboid",

    # Isolates & special groups
    "Albanian": "Albanian",
    "Armenian": "Armenian",
    "Basque":   "Language Isolate",
    "Greek":    "Hellenic",
    "Latin":    "Latino-Faliscan",
    "Japanese": "Japonic",
    "Esperanto":"Constructed",
    "Ido":      "Constructed",
    "Volapük":  "Constructed",
    "Haitian":  "French-based Creole",
}


# ── 1. high-level genealogical families ─────────────────────────────────────────
high_level_families = {
    "Afrikaans": "Indo-European",
    "Albanian": "Indo-European",
    "Arabic": "Afroasiatic",
    "Egyptian Arabic": "Afroasiatic",
    "Aragonese": "Indo-European",
    "Armenian": "Indo-European",
    "Asturian": "Indo-European",
    "Azerbaijani": "Turkic",
    "Bashkir": "Turkic",
    "Basque": "Language Isolate",
    "Bavarian": "Indo-European",
    "Belarusian": "Indo-European",
    "Bengali": "Indo-European",
    "Bishnupriya Manipuri": "Indo-European",
    "Bosnian": "Indo-European",
    "Breton": "Indo-European",
    "Bulgarian": "Indo-European",
    "Burmese": "Sino-Tibetan",
    "Catalan": "Indo-European",
    "Cebuano": "Austronesian",
    "Chechen": "Northeast Caucasian",
    "Chinese (Simplified)": "Sino-Tibetan",
    "Chinese (Traditional)": "Sino-Tibetan",
    "Min Nan Chinese": "Sino-Tibetan",
    "Chuvash": "Turkic",
    "Croatian": "Indo-European",
    "Czech": "Indo-European",
    "Danish": "Indo-European",
    "Dutch": "Indo-European",
    "English": "Indo-European",
    "Estonian": "Uralic",
    "Finnish": "Uralic",
    "French": "Indo-European",
    "Galician": "Indo-European",
    "Georgian": "Kartvelian",
    "German": "Indo-European",
    "Greek": "Indo-European",
    "Gujarati": "Indo-European",
    "Haitian": "Creole",
    "Hebrew": "Afroasiatic",
    "Hindi": "Indo-European",
    "Hungarian": "Uralic",
    "Icelandic": "Indo-European",
    "Ido": "Constructed",
    "Indonesian": "Austronesian",
    "Irish": "Indo-European",
    "Italian": "Indo-European",
    "Japanese": "Japonic",
    "Javanese": "Austronesian",
    "Kannada": "Dravidian",
    "Kazakh": "Turkic",
    "Kirghiz": "Turkic",
    "Korean": "Koreanic",
    "Latin": "Indo-European",
    "Latvian": "Indo-European",
    "Lithuanian": "Indo-European",
    "Lombard": "Indo-European",
    "Low Saxon": "Indo-European",
    "Luxembourgish": "Indo-European",
    "Macedonian": "Indo-European",
    "Malagasy": "Austronesian",
    "Malay": "Austronesian",
    "Malayalam": "Dravidian",
    "Marathi": "Indo-European",
    "Minangkabau": "Austronesian",
    "Nepali": "Indo-European",
    "Newar": "Sino-Tibetan",
    "Norwegian (Bokmal)": "Indo-European",
    "Norwegian (Nynorsk)": "Indo-European",
    "Occitan": "Indo-European",
    "Persian (Farsi)": "Indo-European",
    "Piedmontese": "Indo-European",
    "Polish": "Indo-European",
    "Portuguese": "Indo-European",
    "Punjabi": "Indo-European",
    "Romanian": "Indo-European",
    "Russian": "Indo-European",
    "Scots": "Indo-European",
    "Serbian": "Indo-European",
    "Serbo-Croatian": "Indo-European",
    "Sicilian": "Indo-European",
    "Slovak": "Indo-European",
    "Slovenian": "Indo-European",
    "South Azerbaijani": "Turkic",
    "Spanish": "Indo-European",
    "Sundanese": "Austronesian",
    "Swahili": "Niger-Congo",
    "Swedish": "Indo-European",
    "Tagalog": "Austronesian",
    "Tajik": "Indo-European",
    "Tamil": "Dravidian",
    "Tatar": "Turkic",
    "Telugu": "Dravidian",
    "Turkish": "Turkic",
    "Ukrainian": "Indo-European",
    "Urdu": "Indo-European",
    "Uzbek": "Turkic",
    "Vietnamese": "Austroasiatic",
    "Volapük": "Constructed",
    "Waray-Waray": "Austronesian",
    "Welsh": "Indo-European",
    "West Frisian": "Indo-European",
    "Western Punjabi": "Indo-European",
    "Yoruba": "Niger-Congo",
    "Esperanto": "Constructed",
    "Crimean Tatar": "Turkic",
}

# ── 2. primary branches (first subdivision inside each family) ────────────────
primary_families_branches = {
    # Indo-European
    "Afrikaans": "Germanic",
    "Albanian": "Albanian",
    "Aragonese": "Romance",
    "Armenian": "Armenian",
    "Asturian": "Romance",
    "Bavarian": "Germanic",
    "Belarusian": "Slavic",
    "Bengali": "Indo-Aryan",
    "Bishnupriya Manipuri": "Indo-Aryan",
    "Bosnian": "Slavic",
    "Breton": "Celtic",
    "Bulgarian": "Slavic",
    "Catalan": "Romance",
    "Croatian": "Slavic",
    "Czech": "Slavic",
    "Danish": "Germanic",
    "Dutch": "Germanic",
    "English": "Germanic",
    "French": "Romance",
    "Galician": "Romance",
    "German": "Germanic",
    "Greek": "Hellenic",
    "Gujarati": "Indo-Aryan",
    "Hindi": "Indo-Aryan",
    "Icelandic": "Germanic",
    "Irish": "Celtic",
    "Italian": "Romance",
    "Latin": "Italic",
    "Latvian": "Baltic",
    "Lithuanian": "Baltic",
    "Lombard": "Romance",
    "Low Saxon": "Germanic",
    "Luxembourgish": "Germanic",
    "Macedonian": "Slavic",
    "Marathi": "Indo-Aryan",
    "Nepali": "Indo-Aryan",
    "Norwegian (Bokmal)": "Germanic",
    "Norwegian (Nynorsk)": "Germanic",
    "Occitan": "Romance",
    "Persian (Farsi)": "Iranian",
    "Piedmontese": "Romance",
    "Polish": "Slavic",
    "Portuguese": "Romance",
    "Punjabi": "Indo-Aryan",
    "Romanian": "Romance",
    "Russian": "Slavic",
    "Scots": "Germanic",
    "Serbian": "Slavic",
    "Serbo-Croatian": "Slavic",
    "Sicilian": "Romance",
    "Slovak": "Slavic",
    "Slovenian": "Slavic",
    "Spanish": "Romance",
    "Swedish": "Germanic",
    "Tajik": "Iranian",
    "Ukrainian": "Slavic",
    "Urdu": "Indo-Aryan",
    "West Frisian": "Germanic",
    "Western Punjabi": "Indo-Aryan",
    "Welsh": "Celtic",

    # Afroasiatic
    "Arabic": "Semitic",
    "Egyptian Arabic": "Semitic",
    "Hebrew": "Semitic",

    # Turkic
    "Azerbaijani": "Oghuz",
    "South Azerbaijani": "Oghuz",
    "Turkish": "Oghuz",
    "Bashkir": "Kipchak",
    "Kazakh": "Kipchak",
    "Kirghiz": "Kipchak",
    "Tatar": "Kipchak",
    "Crimean Tatar": "Kipchak",
    "Chuvash": "Oghur",
    "Uzbek": "Karluk",

    # Uralic
    "Estonian": "Finnic",
    "Finnish": "Finnic",
    "Hungarian": "Ugric",

    # Sino-Tibetan
    "Chinese (Simplified)": "Sinitic",
    "Chinese (Traditional)": "Sinitic",
    "Min Nan Chinese": "Sinitic",
    "Burmese": "Tibeto-Burman",
    "Newar": "Tibeto-Burman",

    # Austronesian
    "Cebuano": "Malayo-Polynesian",
    "Indonesian": "Malayo-Polynesian",
    "Javanese": "Malayo-Polynesian",
    "Malagasy": "Malayo-Polynesian",
    "Malay": "Malayo-Polynesian",
    "Minangkabau": "Malayo-Polynesian",
    "Sundanese": "Malayo-Polynesian",
    "Tagalog": "Philippine",
    "Waray-Waray": "Philippine",

    # Dravidian
    "Kannada": "South Dravidian",
    "Malayalam": "South Dravidian",
    "Tamil": "South Dravidian",
    "Telugu": "South-Central Dravidian",

    # Niger-Congo
    "Swahili": "Atlantic-Congo",
    "Yoruba": "Atlantic-Congo",

    # Misc. single-branch families
    "Basque": "Language Isolate",
    "Chechen": "Nakh",
    "Georgian": "Kartvelian",
    "Japanese": "Japonic",
    "Korean": "Koreanic",
    "Vietnamese": "Vietic",

    # Creole & Constructed
    "Haitian": "French-based Creole",
    "Esperanto": "Constructed",
    "Ido": "Constructed",
    "Volapük": "Constructed",
}