code wiki / _hdl_build / nx_html_extract_data_test.nx
nx_html_extract_data_test.nx source
↩ module page · 65 lines · 3251 B
1// nx_html_extract_data_test.nx -- KAT: pull JSON-LD prose from HTML; then a REAL
2// run on /tmp/cadhub_raw.html to show what rung-1 recovers (honest). expect_exit: 0.
3
4import "nx_syscalls.nx"
5import "nx_research_extract.nx"
6import "nx_html_extract_data.nx"
7
8func w(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 }
9func wn(v: i64) -> i64 {
10 let b: *u8 = sys_mmap(28); var m: i64 = v
11 if m < 0 { m = 0 - m; sys_write(1, "-" as *u8, 1) }
12 let t: *u8 = sys_mmap(28); var k: i64 = 0
13 if m == 0 { t[0] = 48 as u8; k = 1 }
14 while m > 0 { t[k] = (48 + (m % 10)) as u8; m = m / 10; k = k + 1 }
15 var i: i64 = 0; while i < k { b[i] = t[k - 1 - i]; i = i + 1 }
16 sys_write(1, b, k); return 0
17}
18func contains(hay: *u8, hn: i64, ndl: *u8) -> i64 {
19 var m: i64 = 0; while ndl[m] != (0 as u8) { m = m + 1 }
20 var i: i64 = 0
21 while i + m <= hn { var j: i64 = 0; var ok: i64 = 1; while j < m { if hay[i + j] != ndl[j] { ok = 0; j = m } else { j = j + 1 } } if ok == 1 { return 1 } i = i + 1 }
22 return 0
23}
24func chk(cond: i64, pass: *i64, fail: *i64, label: *u8) -> i64 {
25 if cond == 1 { pass[0] = pass[0] + 1; w(" ok " as *u8); w(label); w("\n" as *u8) }
26 else { fail[0] = fail[0] + 1; w(" XX " as *u8); w(label); w("\n" as *u8) }
27 return 0
28}
29
30func main() -> i64 {
31 let pass: *i64 = (sys_mmap(8)) as *i64
32 let fail: *i64 = (sys_mmap(8)) as *i64
33 pass[0] = 0
34 fail[0] = 0
35 w("=== nx_html_extract_data KAT (JSON-LD prose from HTML) ===\n" as *u8)
36
37 let html: *u8 = "<html><head><script type=\"application/ld+json\">{\"headline\":\"Best 3D Printing Software\",\"description\":\"covers Cura and PrusaSlicer\",\"articleBody\":\"Blender is great for modeling\"}</script></head></html>" as *u8
38 let hn: i64 = re_strlen(html)
39 let out: *u8 = sys_mmap(16384)
40 let len: i64 = nx_html_extract_data(html, hn, out, 16384)
41 w("--- recovered ---\n" as *u8); w(out); w("--- end ---\n" as *u8)
42 let on: i64 = re_strlen(out)
43 chk(contains(out, on, "Best 3D Printing Software" as *u8) == 1, pass, fail, "headline recovered" as *u8)
44 chk(contains(out, on, "Cura and PrusaSlicer" as *u8) == 1, pass, fail, "description recovered" as *u8)
45 chk(contains(out, on, "Blender is great" as *u8) == 1, pass, fail, "articleBody recovered (JS-only content)" as *u8)
46
47 // real run on the CADHub page we fetched raw
48 w("=== REAL: nx_html_extract_data on /tmp/cadhub_raw.html ===\n" as *u8)
49 let rl: *i64 = (sys_mmap(8)) as *i64
50 let raw: *u8 = sys_read_file("/tmp/cadhub_raw.html" as *u8, rl)
51 if rl[0] > 0 {
52 let rout: *u8 = sys_mmap(131072)
53 let rn: i64 = nx_html_extract_data(raw, rl[0], rout, 131072)
54 w("recovered bytes=" as *u8); wn(rn); w(" (vs 71 from static strip)\n" as *u8)
55 var show: i64 = rn; if show > 700 { show = 700 }
56 sys_write(1, rout, show); w("\n" as *u8)
57 chk(rn > 71, pass, fail, "recovered MORE than the 71-byte static strip" as *u8)
58 } else {
59 w("(no /tmp/cadhub_raw.html -- run nx_page_fetch_raw first)\n" as *u8)
60 }
61
62 w("=== VERDICT pass=" as *u8); wn(pass[0]); w(" fail=" as *u8); wn(fail[0]); w(" ===\n" as *u8)
63 if fail[0] == 0 { return 0 }
64 return 1
65}