react-native-executorch 0.10.6 → 0.10.7

package/third-party/common/phonemis/src/phonemis/lang/it/constants.h

@@ -1,77 +1,157 @@
 #pragma once
 
+#include <phonemis/base/processor/layers/num2word/types.h>
+#include <phonemis/base/processor/common/conditions.h>
+#include <phonemis/base/processor/layers/num2word/notation.h>
+#include <phonemis/base/processor/layers/term2word/types.h>
 #include <phonemis/base/tokenizer/types.h>
 
+#include <array>
 #include <cstdint>
 #include <map>
 #include <string>
+#include <string_view>
 #include <unordered_map>
 #include <unordered_set>
+#include <utility>
 #include <vector>
 
 namespace phonemis::it::constants {
 
 // --- Number to Word Normalization ---
 namespace num2word {
-  // Base cardinals: number -> word. Includes units, teens, and round tens
-  // 20-90. All other values are composed: tens+units fuse into a single word
-  // with elision rules, hundreds/thousands concatenate, millions+ take a space.
-  inline const std::unordered_map<int32_t, std::u32string> kCardinals = {
-    {0,  U"zero"},       {1,  U"uno"},         {2,  U"due"},          {3,  U"tre"},         {4,  U"quattro"},
-    {5,  U"cinque"},     {6,  U"sei"},         {7,  U"sette"},        {8,  U"otto"},        {9,  U"nove"},
-    {10, U"dieci"},      {11, U"undici"},      {12, U"dodici"},       {13, U"tredici"},     {14, U"quattordici"},
-    {15, U"quindici"},   {16, U"sedici"},      {17, U"diciassette"},  {18, U"diciotto"},    {19, U"diciannove"},
-    {20, U"venti"},      {30, U"trenta"},      {40, U"quaranta"},     {50, U"cinquanta"},   {60, U"sessanta"},
-    {70, U"settanta"},   {80, U"ottanta"},     {90, U"novanta"}
+  using ::phonemis::processor::num2word::DateFormat;
+  using ::phonemis::processor::num2word::Forms;
+  using ::phonemis::processor::num2word::Joiners;
+  using ::phonemis::processor::num2word::Keys;
+  using ::phonemis::processor::num2word::keys_of;
+  using ::phonemis::processor::num2word::Notation;
+  using ::phonemis::processor::num2word::PhoneFormat;
+  using ::phonemis::processor::num2word::SuffixFormat;
+
+  inline constexpr std::array<std::u32string_view, 20> kOnes = {
+    U"zero",   U"uno",    U"due",         U"tre",      U"quattro",    U"cinque",      U"sei",
+    U"sette",  U"otto",   U"nove",        U"dieci",    U"undici",     U"dodici",      U"tredici",
+    U"quattordici", U"quindici", U"sedici", U"diciassette", U"diciotto", U"diciannove"
   };
 
-  // Base ordinals in masculine singular form. Feminine/plural variants are
-  // derived at verbalization time from the provided suffix.
-  inline const std::unordered_map<int32_t, std::u32string> kOrdinals = {
-    {1,  U"primo"},    {2,  U"secondo"},  {3,  U"terzo"},    {4,  U"quarto"},    {5,  U"quinto"},
-    {6,  U"sesto"},    {7,  U"settimo"},  {8,  U"ottavo"},   {9,  U"nono"},      {10, U"decimo"}
+  inline constexpr std::array<std::u32string_view, 10> kTens = {
+    U"", U"", U"venti", U"trenta", U"quaranta", U"cinquanta", U"sessanta", U"settanta", U"ottanta", U"novanta"
   };
 
-  // Large-scale names (singular forms).
-  inline const std::map<std::int64_t, std::u32string> kLargeCardinals = {
-    {100LL,           U"cento"},
-    {1000LL,          U"mille"},
-    {1000000LL,       U"milione"},
-    {1000000000LL,    U"miliardo"}
+  // Ordinals up to ten; later ones are the cardinal with -esimo
+  inline constexpr std::array<std::u32string_view, 11> kOrdinals = {
+    U"", U"primo", U"secondo", U"terzo", U"quarto", U"quinto", U"sesto", U"settimo", U"ottavo", U"nono", U"decimo"
   };
 
-  // Plural forms for scales. "cento" is invariable as a multiplier prefix;
-  // "mille" becomes "mila"; "milione"/"miliardo" pluralize as regular nouns.
-  inline const std::unordered_map<std::int64_t, std::u32string> kLargeCardinalsPlural = {
-    {100LL,           U"cento"},
-    {1000LL,          U"mila"},
-    {1000000LL,       U"milioni"},
-    {1000000000LL,    U"miliardi"}
+  inline const std::vector<std::u32string> kMonths = {
+    U"gennaio", U"febbraio", U"marzo",     U"aprile",  U"maggio",   U"giugno",
+    U"luglio",  U"agosto",   U"settembre", U"ottobre", U"novembre", U"dicembre"
   };
 
-  // Currencies (singular and plural). "euro" is invariable; "dollaro" and
-  // "sterlina" follow regular noun pluralization.
-  inline const std::unordered_map<char32_t, std::u32string> kCurrencies = {
-    {U'$', U"dollaro"}, {U'€', U"euro"}, {U'£', U"sterlina"}
+  inline const std::unordered_map<std::u32string, Forms> kCurrencies = {
+    {U"€", {U"euro", U"euro"}}, {U"$", {U"dollaro", U"dollari"}}, {U"£", {U"sterlina", U"sterline"}}
   };
 
-  inline const std::unordered_map<char32_t, std::u32string> kCurrenciesPlural = {
-    {U'$', U"dollari"}, {U'€', U"euro"}, {U'£', U"sterline"}
+  // Most scales are written as the word they are read with: 3 milioni
+  inline const std::unordered_map<std::u32string, Forms> kScales = {
+    {U"mila", {U"mila", U"mila"}},             {U"k", {U"mila", U"mila"}},           {U"K", {U"mila", U"mila"}},
+    {U"milione", {U"milione", U"milione"}},    {U"milioni", {U"milioni", U"milioni"}},
+    {U"miliardo", {U"miliardo", U"miliardo"}}, {U"miliardi", {U"miliardi", U"miliardi"}},
+    {U"mln", {U"milione", U"milioni"}},        {U"mld", {U"miliardo", U"miliardi"}}
   };
 
-  // Months (index 1 = January)
-  inline const std::vector<std::u32string> kMonths = {
-    U"",         U"gennaio",   U"febbraio",  U"marzo",     U"aprile",   U"maggio",   U"giugno",
-    U"luglio",   U"agosto",    U"settembre", U"ottobre",   U"novembre", U"dicembre"
+  inline const std::unordered_map<std::u32string, Forms> kUnits = {
+    {U"km", {U"chilometro", U"chilometri"}},          {U"m", {U"metro", U"metri"}},
+    {U"cm", {U"centimetro", U"centimetri"}},          {U"mm", {U"millimetro", U"millimetri"}},
+    {U"m²", {U"metro quadro", U"metri quadri"}},      {U"m2", {U"metro quadro", U"metri quadri"}},
+    {U"km²", {U"chilometro quadrato", U"chilometri quadrati"}}, {U"m³", {U"metro cubo", U"metri cubi"}},
+    {U"kg", {U"chilo", U"chili"}},                    {U"g", {U"grammo", U"grammi"}},
+    {U"mg", {U"milligrammo", U"milligrammi"}},        {U"l", {U"litro", U"litri"}},
+    {U"ml", {U"millilitro", U"millilitri"}},          {U"TB", {U"terabyte", U"terabyte"}},
+    {U"GB", {U"giga", U"giga"}},                      {U"MB", {U"mega", U"mega"}},
+    {U"Hz", {U"hertz", U"hertz"}},                    {U"kHz", {U"kilohertz", U"kilohertz"}},
+    {U"MHz", {U"megahertz", U"megahertz"}},           {U"GHz", {U"gigahertz", U"gigahertz"}},
+    {U"W", {U"watt", U"watt"}},                       {U"kW", {U"kilowatt", U"kilowatt"}},
+    {U"kWh", {U"kilowattora", U"kilowattora"}},       {U"V", {U"volt", U"volt"}},
+    {U"km/h", {U"chilometro orario", U"chilometri orari"}}, {U"°C", {U"grado", U"gradi"}},
+    {U"°", {U"grado", U"gradi"}}
+  };
+
+  inline const std::map<std::pair<uint64_t, uint64_t>, std::u32string> kFractions = {
+    {{1, 2}, U"un mezzo"}, {{1, 3}, U"un terzo"}, {{2, 3}, U"due terzi"}, {{1, 4}, U"un quarto"},
+    {{3, 4}, U"tre quarti"}
+  };
+
+  // Words before a time with a dot, which make it a time: alle 21.30
+  inline const std::unordered_set<std::u32string> kTimeWords = {
+    U"alle", U"ore", U"dalle", U"le", U"verso", U"entro", U"per"
+  };
+
+  // --- Notation: how numbers are written ---
+  inline const Notation kNotation = {
+    .group_separators = U".", .decimal_separators = U",.", .time_separators = U":.",
+    .currencies = keys_of(kCurrencies), .scales = keys_of(kScales), .units = keys_of(kUnits), .fractions = true
+  };
+
+  // Words between numbers: 3.8.10, 10.0.0.51, 10-15, 1920x1080, 2:1, -5, 2 + 2
+  inline const Joiners kJoiners = {
+    .point = U"punto", .dot = U"punto", .range = U"-", .by = U"per", .score = U" a ", .minus = U"meno", .plus = U"più"
   };
 
-  // Recognized Italian ordinal suffixes attached to digits (e.g. "1º", "1ª",
-  // plural "1ºs", "1ªs").
-  inline const std::unordered_set<std::u32string> kOrdinalSuffixes = {
-    U"º", U"ª", U"ºs", U"ªs"
+  // 24/03/2022
+  inline const DateFormat kDateFormat = {.boundaries = U".,/-"};
+
+  // 333 1234567, 06 1234 5678, 06 12345678
+  inline const std::vector<PhoneFormat> kPhoneFormats = {
+    {.groups = {{3, 3}, {6, 7}}, .separators = U" "},
+    {.groups = {{2, 4}, {3, 4}, {4, 4}}, .separators = U" ", .leading_zero = true},
+    {.groups = {{2, 4}, {6, 8}}, .separators = U" ", .leading_zero = true}
   };
+
+  // 1º, 2 ª: a sign, possibly after a space
+  inline const Keys kOrdinalSigns = {U"º", U"ª"};
+  inline const SuffixFormat kOrdinalSignFormat = {.space = true};
+
+  // 21esimo, 33esima
+  inline const Keys kOrdinalSuffixes = {U"esimo", U"esima", U"esimi", U"esime"};
 } // namespace num2word
 
+// --- Symbols to Word Normalization ---
+namespace sym2word {
+  using ::phonemis::processor::Readings;
+
+  // "&" only joins English names: AT&T, R&B
+  inline const Readings kSymbols = {
+    {U"&", {U"and"}},         {U"+", {U"più"}},       {U"=", {U"uguale a"}}, {U"@", {U"chiocciola"}}, {U"%", {U"per cento"}},
+    {U"°", {U"gradi"}},       {U"#", {U"hashtag"}},   {U"²", {U"al quadrato"}}, {U"³", {U"al cubo"}}, {U"π", {U"pi greco"}},
+    {U"×", {U"per"}},         {U"€", {U"euro"}},      {U"$", {U"dollari"}},  {U"£", {U"sterline"}},   {U"§", {U"paragrafo"}}
+  };
+} // namespace sym2word
+
+// --- Terms to Word Normalization ---
+namespace term2word {
+  using ::phonemis::processor::before_number;
+  using ::phonemis::processor::term2word::Terms;
+
+  inline const Terms kTerms = {
+    .abbreviations = {
+      // Before a number only: n. 5
+      {U"n.", {{U"numero", before_number}}},
+      {U"N.", {{U"numero", before_number}}},
+      {U"nr.", {{U"numero", before_number}}},
+      {U"n°", {{U"numero", before_number}}},
+      {U"N°", {{U"numero", before_number}}},
+      {U"art.", {{U"articolo", before_number}}},
+      {U"Art.", {{U"articolo", before_number}}},
+      {U"pag.", {{U"pagina", before_number}}},
+      {U"cap.", {{U"capitolo", before_number}}},
+      {U"§", {{U"paragrafo", before_number}}},
+    },
+    .web = {.dot = U"punto", .at = U"chiocciola", .slash = U"slash"}
+  };
+} // namespace term2word
+
 // --- Tokenization Rules ---
 namespace tokenizer {
   using ::phonemis::tokenizer::split::Rule;