neuron-technologies/neuron
2
0
Fork 0
Code Issues 2 Pull Requests 3 Actions Packages Projects Releases Wiki Activity
Files
test/layer-composition
neuron/dist/morphology-sga.c
T
will.anderson 48ecd83421 fix: restore elb build — import paths, morphology deps, C master declarations header 
- Fix wrong ELP import paths in soul.el, elp-input.el, studio.el
  (../foundation/elp/src → ../foundation/el/elp/src)
- Add missing import "morphology.el" to all 29 language morphology modules
- Recompile all affected dist/*.c with correct cross-module declarations
- Add dist/elp-c-decls.h: C-level master forward declarations for ELP package
  (enables elb --force-include to resolve undeclared cross-module calls)
2026-05-08 19:43:57 -05:00

572 lines
14 KiB
C
Raw Permalink Blame History

#include <stdint.h>
#include <stdlib.h>
#include "el_runtime.h"
el_val_t str_ends(el_val_t s, el_val_t suf);
el_val_t str_last_char(el_val_t s);
el_val_t str_last2(el_val_t s);
el_val_t str_last3(el_val_t s);
el_val_t str_drop_last(el_val_t s, el_val_t n);
el_val_t is_vowel(el_val_t c);
el_val_t morph_apply_suffix(el_val_t base, el_val_t suffix);
el_val_t en_irregular_plural(el_val_t word);
el_val_t en_irregular_singular(el_val_t word);
el_val_t en_irregular_verb(el_val_t base);
el_val_t en_verb_3sg(el_val_t base);
el_val_t en_should_double_final(el_val_t base);
el_val_t en_verb_past(el_val_t base);
el_val_t en_verb_gerund(el_val_t base);
el_val_t en_pluralize_regular(el_val_t singular);
el_val_t en_verb_form(el_val_t base, el_val_t tense, el_val_t person, el_val_t number);
el_val_t agree_determiner(el_val_t det, el_val_t noun);
el_val_t morph_pluralize(el_val_t noun, el_val_t profile);
el_val_t morph_map_canonical(el_val_t verb, el_val_t code);
el_val_t morph_conjugate(el_val_t verb, el_val_t tense, el_val_t person, el_val_t number, el_val_t profile);
el_val_t morph_inflect(el_val_t word, el_val_t features, el_val_t profile);
el_val_t pluralize(el_val_t singular);
el_val_t singularize(el_val_t plural);
el_val_t verb_form(el_val_t base, el_val_t tense, el_val_t person, el_val_t number);
el_val_t irregular_plural(el_val_t word);
el_val_t irregular_singular(el_val_t word);
el_val_t sga_drop(el_val_t s, el_val_t n);
el_val_t sga_first(el_val_t s);
el_val_t sga_rest(el_val_t s);
el_val_t sga_slot(el_val_t person, el_val_t number);
el_val_t sga_lenite(el_val_t word);
el_val_t sga_copula_present(el_val_t slot);
el_val_t sga_bith_present(el_val_t slot);
el_val_t sga_bith_past(el_val_t slot);
el_val_t sga_teit_present(el_val_t slot);
el_val_t sga_teit_past(el_val_t slot);
el_val_t sga_gaibid_present(el_val_t slot);
el_val_t sga_adci_present(el_val_t slot);
el_val_t sga_asbeir_present(el_val_t slot);
el_val_t sga_map_canonical(el_val_t verb);
el_val_t sga_ai_present(el_val_t stem, el_val_t slot);
el_val_t sga_conjugate(el_val_t verb, el_val_t tense, el_val_t person, el_val_t number);
el_val_t sga_decline_ostem(el_val_t noun, el_val_t gram_case, el_val_t number);
el_val_t sga_decline_astem(el_val_t noun, el_val_t gram_case, el_val_t number);
el_val_t sga_detect_gender(el_val_t noun);
el_val_t sga_decline(el_val_t noun, el_val_t gram_case, el_val_t number);
el_val_t sga_noun_phrase(el_val_t noun, el_val_t gram_case, el_val_t number, el_val_t definite);
el_val_t sga_drop(el_val_t s, el_val_t n) {
el_val_t len = str_len(s);
if (n >= len) {
return EL_STR("");
}
return str_slice(s, 0, (len - n));
return 0;
}
el_val_t sga_first(el_val_t s) {
if (str_len(s) == 0) {
return EL_STR("");
}
return str_slice(s, 0, 1);
return 0;
}
el_val_t sga_rest(el_val_t s) {
el_val_t n = str_len(s);
if (n <= 1) {
return EL_STR("");
}
return str_slice(s, 1, n);
return 0;
}
el_val_t sga_slot(el_val_t person, el_val_t number) {
if (str_eq(person, EL_STR("first"))) {
if (str_eq(number, EL_STR("singular"))) {
return 0;
}
return 3;
}
if (str_eq(person, EL_STR("second"))) {
if (str_eq(number, EL_STR("singular"))) {
return 1;
}
return 4;
}
if (str_eq(number, EL_STR("singular"))) {
return 2;
}
return 5;
return 0;
}
el_val_t sga_lenite(el_val_t word) {
el_val_t init = sga_first(word);
el_val_t tail = sga_rest(word);
if (str_eq(init, EL_STR("b"))) {
return el_str_concat(EL_STR("bh"), tail);
}
if (str_eq(init, EL_STR("c"))) {
return el_str_concat(EL_STR("ch"), tail);
}
if (str_eq(init, EL_STR("d"))) {
return el_str_concat(EL_STR("dh"), tail);
}
if (str_eq(init, EL_STR("f"))) {
return el_str_concat(EL_STR("fh"), tail);
}
if (str_eq(init, EL_STR("g"))) {
return el_str_concat(EL_STR("gh"), tail);
}
if (str_eq(init, EL_STR("m"))) {
return el_str_concat(EL_STR("mh"), tail);
}
if (str_eq(init, EL_STR("p"))) {
return el_str_concat(EL_STR("ph"), tail);
}
if (str_eq(init, EL_STR("s"))) {
return el_str_concat(EL_STR("sh"), tail);
}
if (str_eq(init, EL_STR("t"))) {
return el_str_concat(EL_STR("th"), tail);
}
return word;
return 0;
}
el_val_t sga_copula_present(el_val_t slot) {
if (slot == 0) {
return EL_STR("am");
}
if (slot == 1) {
return EL_STR("at");
}
if (slot == 2) {
return EL_STR("is");
}
if (slot == 3) {
return EL_STR("am");
}
if (slot == 4) {
return EL_STR("adib");
}
return EL_STR("it");
return 0;
}
el_val_t sga_bith_present(el_val_t slot) {
if (slot == 0) {
return EL_STR("am");
}
if (slot == 1) {
return EL_STR("at");
}
if (slot == 2) {
return EL_STR("is");
}
if (slot == 3) {
return EL_STR("am");
}
if (slot == 4) {
return EL_STR("adib");
}
return EL_STR("at");
return 0;
}
el_val_t sga_bith_past(el_val_t slot) {
if (slot == 0) {
return EL_STR("ba");
}
if (slot == 1) {
return EL_STR("ba");
}
if (slot == 2) {
return EL_STR("ba");
}
if (slot == 3) {
return EL_STR("b\xc3\xa1mmar");
}
if (slot == 4) {
return EL_STR("b\xc3\xa1""daid");
}
return EL_STR("batar");
return 0;
}
el_val_t sga_teit_present(el_val_t slot) {
if (slot == 0) {
return EL_STR("t\xc3\xad""agu");
}
if (slot == 1) {
return EL_STR("t\xc3\xa9it");
}
if (slot == 2) {
return EL_STR("t\xc3\xa9it");
}
if (slot == 3) {
return EL_STR("t\xc3\xad""agmai");
}
if (slot == 4) {
return EL_STR("t\xc3\xad""agid");
}
return EL_STR("t\xc3\xad""agat");
return 0;
}
el_val_t sga_teit_past(el_val_t slot) {
if (slot == 0) {
return EL_STR("lod");
}
if (slot == 1) {
return EL_STR("lod");
}
if (slot == 2) {
return EL_STR("luid");
}
if (slot == 3) {
return EL_STR("lodmar");
}
if (slot == 4) {
return EL_STR("lodaid");
}
return EL_STR("lotar");
return 0;
}
el_val_t sga_gaibid_present(el_val_t slot) {
if (slot == 0) {
return EL_STR("gaibim");
}
if (slot == 1) {
return EL_STR("gaibi");
}
if (slot == 2) {
return EL_STR("gaibid");
}
if (slot == 3) {
return EL_STR("gaibmi");
}
if (slot == 4) {
return EL_STR("gaibthe");
}
return EL_STR("gaibid");
return 0;
}
el_val_t sga_adci_present(el_val_t slot) {
if (slot == 0) {
return EL_STR("ad\xc2\xb7""ciu");
}
if (slot == 1) {
return EL_STR("ad\xc2\xb7""c\xc3\xad");
}
if (slot == 2) {
return EL_STR("ad\xc2\xb7""c\xc3\xad");
}
if (slot == 3) {
return EL_STR("ad\xc2\xb7""c\xc3\xadmi");
}
if (slot == 4) {
return EL_STR("ad\xc2\xb7""c\xc3\xadthe");
}
return EL_STR("ad\xc2\xb7""ciat");
return 0;
}
el_val_t sga_asbeir_present(el_val_t slot) {
if (slot == 0) {
return EL_STR("as\xc2\xb7""biur");
}
if (slot == 1) {
return EL_STR("as\xc2\xb7""beir");
}
if (slot == 2) {
return EL_STR("as\xc2\xb7""beir");
}
if (slot == 3) {
return EL_STR("as\xc2\xb7""beram");
}
if (slot == 4) {
return EL_STR("as\xc2\xb7""berid");
}
return EL_STR("as\xc2\xb7""berat");
return 0;
}
el_val_t sga_map_canonical(el_val_t verb) {
if (str_eq(verb, EL_STR("be"))) {
return EL_STR("is");
}
if (str_eq(verb, EL_STR("go"))) {
return EL_STR("t\xc3\xa9it");
}
if (str_eq(verb, EL_STR("take"))) {
return EL_STR("gaibid");
}
if (str_eq(verb, EL_STR("hold"))) {
return EL_STR("gaibid");
}
if (str_eq(verb, EL_STR("see"))) {
return EL_STR("ad\xc2\xb7""c\xc3\xad");
}
if (str_eq(verb, EL_STR("say"))) {
return EL_STR("as\xc2\xb7""beir");
}
return verb;
return 0;
}
el_val_t sga_ai_present(el_val_t stem, el_val_t slot) {
if (slot == 0) {
return el_str_concat(stem, EL_STR("aim"));
}
if (slot == 1) {
return el_str_concat(stem, EL_STR("ai"));
}
if (slot == 2) {
return el_str_concat(stem, EL_STR("aid"));
}
if (slot == 3) {
return el_str_concat(stem, EL_STR("am"));
}
if (slot == 4) {
return el_str_concat(stem, EL_STR("aid"));
}
return el_str_concat(stem, EL_STR("at"));
return 0;
}
el_val_t sga_conjugate(el_val_t verb, el_val_t tense, el_val_t person, el_val_t number) {
el_val_t v = sga_map_canonical(verb);
el_val_t slot = sga_slot(person, number);
if (str_eq(v, EL_STR("is"))) {
if (str_eq(tense, EL_STR("present"))) {
return sga_copula_present(slot);
}
return EL_STR("ba");
}
if (str_eq(v, EL_STR("bith"))) {
if (str_eq(tense, EL_STR("present"))) {
return sga_bith_present(slot);
}
if (str_eq(tense, EL_STR("past"))) {
return sga_bith_past(slot);
}
return v;
}
if (str_eq(v, EL_STR("t\xc3\xa9it"))) {
if (str_eq(tense, EL_STR("present"))) {
return sga_teit_present(slot);
}
if (str_eq(tense, EL_STR("past"))) {
return sga_teit_past(slot);
}
return v;
}
if (str_eq(v, EL_STR("gaibid"))) {
if (str_eq(tense, EL_STR("present"))) {
return sga_gaibid_present(slot);
}
return EL_STR("gab");
}
if (str_eq(v, EL_STR("ad\xc2\xb7""c\xc3\xad"))) {
if (str_eq(tense, EL_STR("present"))) {
return sga_adci_present(slot);
}
return v;
}
if (str_eq(v, EL_STR("as\xc2\xb7""beir"))) {
if (str_eq(tense, EL_STR("present"))) {
return sga_asbeir_present(slot);
}
return v;
}
if (str_ends_with(v, EL_STR("id"))) {
el_val_t stem = sga_drop(v, 2);
if (str_eq(tense, EL_STR("present"))) {
return sga_ai_present(stem, slot);
}
return v;
}
return v;
return 0;
}
el_val_t sga_decline_ostem(el_val_t noun, el_val_t gram_case, el_val_t number) {
if (str_eq(noun, EL_STR("fer"))) {
if (str_eq(number, EL_STR("singular"))) {
if (str_eq(gram_case, EL_STR("nominative"))) {
return EL_STR("fer");
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return EL_STR("fhir");
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return EL_STR("fer");
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return EL_STR("fir");
}
if (str_eq(gram_case, EL_STR("dative"))) {
return EL_STR("fiur");
}
return EL_STR("fer");
}
if (str_eq(gram_case, EL_STR("nominative"))) {
return EL_STR("fir");
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return EL_STR("firu");
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return EL_STR("firu");
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return EL_STR("fer");
}
if (str_eq(gram_case, EL_STR("dative"))) {
return EL_STR("feraib");
}
return EL_STR("fir");
}
if (str_eq(number, EL_STR("singular"))) {
if (str_eq(gram_case, EL_STR("nominative"))) {
return noun;
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return sga_lenite(noun);
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return noun;
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return el_str_concat(noun, EL_STR("a"));
}
if (str_eq(gram_case, EL_STR("dative"))) {
return el_str_concat(noun, EL_STR("u"));
}
return noun;
}
if (str_eq(gram_case, EL_STR("nominative"))) {
return el_str_concat(noun, EL_STR("i"));
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return el_str_concat(noun, EL_STR("u"));
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return el_str_concat(noun, EL_STR("u"));
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return noun;
}
if (str_eq(gram_case, EL_STR("dative"))) {
return el_str_concat(noun, EL_STR("aib"));
}
return el_str_concat(noun, EL_STR("i"));
return 0;
}
el_val_t sga_decline_astem(el_val_t noun, el_val_t gram_case, el_val_t number) {
if (str_eq(noun, EL_STR("ben"))) {
if (str_eq(number, EL_STR("singular"))) {
if (str_eq(gram_case, EL_STR("nominative"))) {
return EL_STR("ben");
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return EL_STR("ben");
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return EL_STR("bein");
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return EL_STR("mn\xc3\xa1");
}
if (str_eq(gram_case, EL_STR("dative"))) {
return EL_STR("mn\xc3\xa1ib");
}
return EL_STR("ben");
}
if (str_eq(gram_case, EL_STR("nominative"))) {
return EL_STR("mn\xc3\xa1");
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return EL_STR("mn\xc3\xa1");
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return EL_STR("mn\xc3\xa1");
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return EL_STR("ban");
}
if (str_eq(gram_case, EL_STR("dative"))) {
return EL_STR("mn\xc3\xa1ib");
}
return EL_STR("mn\xc3\xa1");
}
if (str_eq(number, EL_STR("singular"))) {
if (str_eq(gram_case, EL_STR("nominative"))) {
return noun;
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return noun;
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return el_str_concat(noun, EL_STR("i"));
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return el_str_concat(noun, EL_STR("e"));
}
if (str_eq(gram_case, EL_STR("dative"))) {
return el_str_concat(noun, EL_STR("aib"));
}
return noun;
}
if (str_eq(gram_case, EL_STR("nominative"))) {
return el_str_concat(noun, EL_STR("a"));
}
if (str_eq(gram_case, EL_STR("vocative"))) {
return el_str_concat(noun, EL_STR("a"));
}
if (str_eq(gram_case, EL_STR("accusative"))) {
return el_str_concat(noun, EL_STR("a"));
}
if (str_eq(gram_case, EL_STR("genitive"))) {
return noun;
}
if (str_eq(gram_case, EL_STR("dative"))) {
return el_str_concat(noun, EL_STR("aib"));
}
return el_str_concat(noun, EL_STR("a"));
return 0;
}
el_val_t sga_detect_gender(el_val_t noun) {
if (str_eq(noun, EL_STR("ben"))) {
return EL_STR("feminine");
}
if (str_eq(noun, EL_STR("mn\xc3\xa1"))) {
return EL_STR("feminine");
}
return EL_STR("masculine");
return 0;
}
el_val_t sga_decline(el_val_t noun, el_val_t gram_case, el_val_t number) {
el_val_t gender = sga_detect_gender(noun);
if (str_eq(gender, EL_STR("masculine"))) {
return sga_decline_ostem(noun, gram_case, number);
}
if (str_eq(gender, EL_STR("feminine"))) {
return sga_decline_astem(noun, gram_case, number);
}
return noun;
return 0;
}
el_val_t sga_noun_phrase(el_val_t noun, el_val_t gram_case, el_val_t number, el_val_t definite) {
el_val_t base = sga_decline(noun, gram_case, number);
if (!str_eq(definite, EL_STR("true"))) {
return base;
}
return el_str_concat(EL_STR("in "), base);
return 0;
}
Reference in New Issue View Git Blame Copy Permalink