code wiki / _hdl_build / nx_research_relation_clean.nx
nx_research_relation_clean.nx
buildroot/runtime/_hdl_build/nx_research_relation_clean.nx
about
nx_research_relation_clean.nx -- Researcher EXTRACT, rung 3: CLAIM CLEANUP.
The measured noise from rung 2 (pronoun subjects, "[1]"/"^" citation markers,
"Wikipedia(R) trademark" boilerplate, junk) is fixed here -- by IMPROVING the
claims, not by stripping the feature. Inspectable + data-driven (stopword /
pronoun / boilerplate tables). nx_claim_clean(subj,obj -> clean) -> 1 good / 0 reject.
license_tier: ORIGINAL.
dependencies 2 imports · 3 importers
imports: nx_syscalls.nxnx_research_extract.nx
imported by: nx_research_landscape.nxnx_research_relation_clean_test.nxnx_research_relations.nx
structs
| none |
consts
| none |
functions
| 11 | func cl_is_alpha(c: u8) -> i64 called by 1: cl_is_alnum |
| 16 | func cl_is_alnum(c: u8) -> i64 |
| 21 | func cl_strlen(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } return n } |
| 23 | func cl_max_subj_words() -> i64 { return 6 } called by 1: nx_claim_clean |
| 24 | func cl_word_count(s: *u8) -> i64 called by 1: nx_claim_clean |
| 32 | func cl_streq(a: *u8, b: *u8) -> i64 called by 1: tc |
| 37 | func cl_lc(c: u8) -> u8 { if c >= (65 as u8) { if c <= (90 as u8) { return (c + 32) as u8 } } return c } called by 1: cl_streq_ci |
| 39 | func cl_streq_ci(a: *u8, b: *u8) -> i64 |
| 46 | func cl_first_token(s: *u8, tok: *u8) -> i64 |
| 54 | func cl_is_stopword(t: *u8) -> i64 |
| 69 | func cl_is_pronoun(t: *u8) -> i64 |
| 79 | func cl_boiler(s: *u8) -> i64 |
| 91 | func cl_drop_first_token(out: *u8, tl: i64) -> i64 called by 1: cl_clean_subject |
| 103 | func cl_copy_from_alpha(src: *u8, out: *u8) -> i64 |
| 116 | func cl_clean_subject(src: *u8, out: *u8) -> i64 |
| 129 | func cl_is_verb(t: *u8) -> i64 |
| 167 | func nx_claim_clean(subj_in: *u8, obj_in: *u8, subj_out: *u8, obj_out: *u8) -> i64 |