code wiki / _hdl_build / nx_site_import.nx
`nx_site_import.nx` source

1// nx_site_import.nx -- the MIGRATION IMPORTER: the no-lock-in CAPABILITY (not architecture). Ingest a foreign site's
2// HTML (e.g., a Wix/Webflow/any export) and produce our PORTABLE, SOVEREIGN blueprint: title + structure counts +
3// stripped visible content, with all third-party JS/tracking REMOVED. Lock-in platforms can't import a competitor's
4// site; we can -- so "no-lock-in" becomes a thing you can DO (migrate TO us from anywhere), measurably better-at-job.
5// 100% sovereign string processing.  license_tier: ORIGINAL
6import "nx_syscalls.nx"
7const K_MAGIC_65536: i64 = 65536
8
9func mi_strlen(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} return n }
10func mi_match_at(h: *u8, i: i64, n: i64, lit: *u8) -> i64 {
11    let ll: i64 = mi_strlen(lit)
12    if i + ll > n { return 0 }
13    var j: i64 = 0
14    while j < ll { if h[i+j] != lit[j] { return 0 } j = j + 1 }
15    return 1
16}
17func mi_find_from(h: *u8, start: i64, n: i64, lit: *u8) -> i64 {
18    let ll: i64 = mi_strlen(lit)
19    if ll == 0 { return 0 - 1 }
20    var i: i64 = start
21    while i + ll <= n { if mi_match_at(h, i, n, lit) == 1 { return i } i = i + 1 }
22    return 0 - 1
23}
24func mi_count(h: *u8, n: i64, lit: *u8) -> i64 {
25    let ll: i64 = mi_strlen(lit)
26    if ll == 0 { return 0 }
27    var c: i64 = 0; var i: i64 = 0
28    while i + ll <= n { if mi_match_at(h, i, n, lit) == 1 { c = c + 1; i = i + ll } else { i = i + 1 } }
29    return c
30}
31// extract visible text: skip <script>/<style> bodies entirely, drop all tags, collapse whitespace.
32func mi_strip_tags(h: *u8, n: i64, out: *u8, cap: i64) -> i64 {
33    var i: i64 = 0; var o: i64 = 0; var in_tag: i64 = 0; var last_space: i64 = 1
34    while i < n {
35        let c: u8 = h[i]
36        if c == (60 as u8) {                       // '<'
37            if mi_match_at(h, i, n, "<script" as *u8) == 1 { let e: i64 = mi_find_from(h, i, n, "</script>" as *u8); if e < 0 { return o } i = e + 9; in_tag = 0; continue }
38            if mi_match_at(h, i, n, "<style" as *u8) == 1 { let e: i64 = mi_find_from(h, i, n, "</style>" as *u8); if e < 0 { return o } i = e + 8; in_tag = 0; continue }
39            in_tag = 1; i = i + 1; continue
40        }
41        if c == (62 as u8) {                       // '>'
42            in_tag = 0; i = i + 1
43            if last_space == 0 { if o < cap { out[o] = 32 as u8; o = o + 1 } last_space = 1 }
44            continue
45        }
46        if in_tag == 0 {
47            var sp: i64 = 0
48            if c == (32 as u8) { sp = 1 }
49            if c == (10 as u8) { sp = 1 }
50            if c == (9 as u8) { sp = 1 }
51            if c == (13 as u8) { sp = 1 }
52            if sp == 1 { if last_space == 0 { if o < cap { out[o] = 32 as u8; o = o + 1 } last_space = 1 } }
53            else { if o < cap { out[o] = c; o = o + 1 } last_space = 0 }
54        }
55        i = i + 1
56    }
57    return o
58}
59func mi_extract_between(h: *u8, n: i64, open: *u8, close: *u8, out: *u8, cap: i64) -> i64 {
60    let s0: i64 = mi_find_from(h, 0, n, open)
61    if s0 < 0 { return 0 }
62    let s: i64 = s0 + mi_strlen(open)
63    let e: i64 = mi_find_from(h, s, n, close)
64    if e < 0 { return 0 }
65    var o: i64 = 0; var i: i64 = s
66    while i < e { if o < cap { out[o] = h[i]; o = o + 1 } i = i + 1 }
67    return o
68}
69
70func mi_w(fd: i64, s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(fd,s,n); return 0 }
71func mi_wn(fd: i64, v: i64) -> i64 { let bb: *u8=sys_mmap(28); var m: i64=v; if m<0{m=0-m;sys_write(fd,"-" as *u8,1)} let t: *u8=sys_mmap(28); var k: i64=0; if m==0{t[0]=48 as u8;k=1} while m>0{t[k]=(48+(m%10)) as u8;m=m/10;k=k+1} var i: i64=0; while i<k{bb[i]=t[k-1-i];i=i+1} sys_write(fd,bb,k); return 0 }
72
73// emit our portable, sovereign blueprint from foreign HTML.
74func mi_to_blueprint(h: *u8, n: i64, fd: i64) -> i64 {
75    let buf: *u8 = sys_mmap(K_MAGIC_65536)
76    mi_w(fd, "site|title|" as *u8)
77    let tl: i64 = mi_extract_between(h, n, "<title>" as *u8, "</title>" as *u8, buf, K_MAGIC_65536)
78    sys_write(fd, buf, tl); mi_w(fd, "\n" as *u8)
79    mi_w(fd, "site|headings|" as *u8); mi_wn(fd, mi_count(h, n, "<h1" as *u8) + mi_count(h, n, "<h2" as *u8) + mi_count(h, n, "<h3" as *u8)); mi_w(fd, "\n" as *u8)
80    mi_w(fd, "site|paragraphs|" as *u8); mi_wn(fd, mi_count(h, n, "<p" as *u8)); mi_w(fd, "\n" as *u8)
81    mi_w(fd, "site|links|" as *u8); mi_wn(fd, mi_count(h, n, "<a " as *u8)); mi_w(fd, "\n" as *u8)
82    mi_w(fd, "block|content|" as *u8)
83    let cl: i64 = mi_strip_tags(h, n, buf, K_MAGIC_65536)
84    sys_write(fd, buf, cl); mi_w(fd, "\n" as *u8)
85    return 0
86}
nx_site_import.nx source

`nx_site_import.nx` source
