nx_doc_catalog_gate.nx source
↩ module page · 83 lines · 4734 B
1// nx_doc_catalog_gate.nx -- R2 GATE for the doc-intelligence LIBRARIAN: proves documents are cataloged as
2// content-addressed structured records, retrievable + decodable, deduped by content (same doc -> same CID),
3// and additive (distinct docs coexist; re-catalog is idempotent). Hermetic /tmp prefix. Exits 0 iff ALL
4// pass. license_tier: ORIGINAL
5import "nx_syscalls.nx"
6import "nx_asset_record.nx"
7import "nx_doc_catalog.nx"
8
9func g_puts(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 }
10func g_putn(v: i64) -> i64 {
11 if v == 0 { sys_write(1, "0" as *u8, 1); return 0 }
12 var m: i64 = v; if m < 0 { sys_write(1, "-" as *u8, 1); m = 0 - m }
13 let d: *u8 = sys_mmap(24); var k: i64 = 0
14 while m > 0 { d[k] = (48 + (m % 10)) as u8; m = m / 10; k = k + 1 }
15 var j: i64 = k - 1
16 while j >= 0 { sys_write(1, ((d as i64)+j) as *u8, 1); j = j - 1 }
17 return 0
18}
19func chk(name: *u8, got: i64, want: i64, st: *i64) -> i64 {
20 if got == want { st[0] = st[0] + 1; g_puts(" PASS "); g_puts(name); g_puts("\n") }
21 else { st[1] = st[1] + 1; g_puts(" FAIL "); g_puts(name); g_puts(" got="); g_putn(got); g_puts(" want="); g_putn(want); g_puts("\n") }
22 return 0
23}
24
25func main() -> i64 {
26 let st: *i64 = sys_mmap(16) as *i64
27 st[0] = 0; st[1] = 0
28
29 // hermetic prefix "/tmp/doccat_<us>-"
30 let pfx: *u8 = sys_mmap(64)
31 var pp: i64 = 0
32 let pre: *u8 = "/tmp/doccat_\x00" as *u8
33 while pre[pp] != (0 as u8) { pfx[pp] = pre[pp]; pp = pp + 1 }
34 let us: i64 = sys_now_us()
35 let tmp: *u8 = sys_mmap(32); var m: i64 = us; var kk: i64 = 0
36 if m == 0 { tmp[0] = 0x30 as u8; kk = 1 }
37 while m > 0 { tmp[kk] = (0x30 + (m % 10)) as u8; m = m / 10; kk = kk + 1 }
38 var z: i64 = kk - 1
39 while z >= 0 { pfx[pp] = tmp[z]; pp = pp + 1; z = z - 1 }
40 pfx[pp] = 0x2d as u8; pp = pp + 1; pfx[pp] = 0 as u8
41
42 let cid1: *u8 = sys_mmap(96)
43 let cid1b: *u8 = sys_mmap(96)
44 let cid2: *u8 = sys_mmap(96)
45
46 chk("catalog doc1 (Memorial bill)", dc_catalog_doc(pfx, "Memorial Hospital Itemized Statement\x00" as *u8, "Memorial Hospital\x00" as *u8, "2025-03-10\x00" as *u8, "ACCT-12345\x00" as *u8, "5 line-items; 4 audit flags; $2356 fair-price overcharge\x00" as *u8, cid1), 0, st)
47
48 // retrieve + decode doc1
49 let pq: *i64 = sys_mmap(16) as *i64
50 let lq: *i64 = sys_mmap(16) as *i64
51 chk("doc1 retrievable by CID", dc_get(pfx, cid1, pq, lq), 1, st)
52 let keys: *i64 = sys_mmap(8 * 64) as *i64
53 let vals: *i64 = sys_mmap(8 * 64) as *i64
54 let nf: i64 = ar_decode(pq[0] as *u8, lq[0], keys, vals, 64)
55 var dok: i64 = 0
56 if nf > 0 { dok = 1 }
57 chk("doc1 decodes", dok, 1, st)
58 chk("doc1 type = doc", ar_streq(ar_get(keys, vals, nf, "type\x00" as *u8), "doc\x00" as *u8), 1, st)
59 chk("doc1 title intact", ar_streq(ar_get(keys, vals, nf, "title\x00" as *u8), "Memorial Hospital Itemized Statement\x00" as *u8), 1, st)
60 chk("doc1 is_current = 1", ar_streq(ar_get(keys, vals, nf, "is_current\x00" as *u8), "1\x00" as *u8), 1, st)
61 chk("doc1 carries structured extract summary", ar_streq(ar_get(keys, vals, nf, "extract_summary\x00" as *u8), "5 line-items; 4 audit flags; $2356 fair-price overcharge\x00" as *u8), 1, st)
62 chk("doc1 classification = private", ar_streq(ar_get(keys, vals, nf, "classification\x00" as *u8), "private\x00" as *u8), 1, st)
63
64 // a DIFFERENT document -> different CID
65 chk("catalog doc2 (different statement)", dc_catalog_doc(pfx, "City Clinic Statement\x00" as *u8, "City Clinic\x00" as *u8, "2025-04-02\x00" as *u8, "ACCT-99\x00" as *u8, "2 line-items; 0 flags\x00" as *u8, cid2), 0, st)
66 chk("doc2 CID differs from doc1 (content-addressed)", ar_streq(cid1, cid2), 0, st)
67
68 // re-catalog doc1 (same content) -> same CID (idempotent dedup)
69 chk("re-catalog doc1 (dedup)", dc_catalog_doc(pfx, "Memorial Hospital Itemized Statement\x00" as *u8, "Memorial Hospital\x00" as *u8, "2025-03-10\x00" as *u8, "ACCT-12345\x00" as *u8, "5 line-items; 4 audit flags; $2356 fair-price overcharge\x00" as *u8, cid1b), 0, st)
70 chk("dedup: same content -> same CID", ar_streq(cid1, cid1b), 1, st)
71
72 // additive: both documents still coexist + retrievable
73 chk("doc1 STILL retrievable (additive)", dc_get(pfx, cid1, pq, lq), 1, st)
74 chk("doc2 retrievable (additive)", dc_get(pfx, cid2, pq, lq), 1, st)
75
76 // negative control
77 chk("bogus CID not found (-1)", dc_get(pfx, "nxc1-deadbeef\x00" as *u8, pq, lq), 0 - 1, st)
78
79 g_puts("nx_doc_catalog_gate: PASS="); g_putn(st[0]); g_puts(" FAIL="); g_putn(st[1]); g_puts("\n")
80 if st[1] == 0 { g_puts("DOC-INTEL R2 nx_doc_catalog: GREEN\n"); return 0 }
81 g_puts("DOC-INTEL R2 nx_doc_catalog: RED\n")
82 return 1
83}