Janady13/missionmatch / internal/modules/opportunities/normalize_test.go

Source identity

Field	Recorded value
Source type	GitHub source
Original record	Janady13/missionmatch / internal/modules/opportunities/normalize_test.go
Date	Not specified in source
Repository	Janady13/missionmatch
Revision	720d47fae476b0d38477ae4401697d372b1ea2bc
Branch	feature/opportunity-to-outcome
Path or archive member	internal/modules/opportunities/normalize_test.go
Original SHA-256	0584a6ac1ca7742d72212512a1e85deee6d8c42acf24309cc03b7ca75635d5e8
Git blob SHA	d98611f1b6a1967ef080a9c532b840a2d1a9e071
Reading-copy SHA-256	0584a6ac1ca7742d72212512a1e85deee6d8c42acf24309cc03b7ca75635d5e8

Original source ↗

Recovered source transcript

Download complete source text ↓102 lines · 4.4 KB

package opportunities

import (
	"strings"
	"testing"
)

const srcText = `Statement of Work. The contractor shall provide senior data engineering support to the
Department of Veterans Affairs. Personnel must hold an active Secret clearance. The contractor shall
deliver a monthly status report. Two Senior Data Engineers are required at the Austin, Texas facility.`

// TestUncitedModelOutputIsRejected is the Gate 2 criterion: an item the model produced without a quote
// cannot become a scope item, no matter how plausible it reads.
func TestUncitedModelOutputIsRejected(t *testing.T) {
	raw := []byte(`{"items":[{"scope_type":"task","title":"Provide data engineering","quote":""}]}`)
	got, rejected, err := ValidateScopeCandidates(raw, srcText)
	if err != nil {
		t.Fatalf("validation should not error on a well formed payload: %v", err)
	}
	if len(got) != 0 {
		t.Fatalf("an uncited item must be rejected, got %d accepted", len(got))
	}
	if len(rejected) != 1 || !strings.Contains(rejected[0].Reason, "uncited") {
		t.Fatalf("expected one uncited rejection, got %+v", rejected)
	}
}

// TestFabricatedCitationIsRejected is the strongest grounding guarantee here: a quote that does not appear
// in the source text is a FABRICATED citation, which is worse than no citation because it reads as evidence.
// The check is exact containment against the source, so the model cannot invent support for a claim.
func TestFabricatedCitationIsRejected(t *testing.T) {
	raw := []byte(`{"items":[{"scope_type":"clearance","title":"TS SCI required",
	  "quote":"Personnel must hold an active TS SCI clearance with polygraph."}]}`)
	got, rejected, err := ValidateScopeCandidates(raw, srcText)
	if err != nil {
		t.Fatalf("unexpected error: %v", err)
	}
	if len(got) != 0 {
		t.Fatal("a quote absent from the source must be rejected as a fabricated citation")
	}
	if len(rejected) != 1 || !strings.Contains(rejected[0].Reason, "not found in the source") {
		t.Fatalf("expected a fabricated citation rejection, got %+v", rejected)
	}
}

// TestGroundedItemIsAcceptedAndHashed proves a properly cited item passes and carries the hash of the exact
// quote, which is what the scope item stores as its citation.
func TestGroundedItemIsAcceptedAndHashed(t *testing.T) {
	raw := []byte(`{"items":[{"scope_type":"clearance","title":"Secret clearance required",
	  "quote":"Personnel must hold an active Secret clearance.","confidence":0.9}]}`)
	got, _, err := ValidateScopeCandidates(raw, srcText)
	if err != nil {
		t.Fatalf("unexpected error: %v", err)
	}
	if len(got) != 1 {
		t.Fatalf("a grounded item must be accepted, got %d", len(got))
	}
	if got[0].QuoteHash == "" || len(got[0].QuoteHash) != 64 {
		t.Fatalf("expected a sha256 hex quote hash, got %q", got[0].QuoteHash)
	}
	if got[0].ScopeType != "clearance" {
		t.Errorf("scope type not preserved: %q", got[0].ScopeType)
	}
}

// TestInvalidScopeTypeIsRejected proves the closed vocabulary is enforced before the database sees it, so a
// model inventing a category is refused rather than becoming an 'other' catch all.
func TestInvalidScopeTypeIsRejected(t *testing.T) {
	raw := []byte(`{"items":[{"scope_type":"vibes","title":"x","quote":"The contractor shall deliver a monthly status report."}]}`)
	got, rejected, err := ValidateScopeCandidates(raw, srcText)
	if err != nil {
		t.Fatalf("unexpected error: %v", err)
	}
	if len(got) != 0 || len(rejected) != 1 {
		t.Fatalf("an out of vocabulary scope type must be rejected, got %d accepted", len(got))
	}
	if !strings.Contains(rejected[0].Reason, "scope_type") {
		t.Errorf("expected a scope_type rejection, got %q", rejected[0].Reason)
	}
}

// TestMalformedPayloadFailsClosed proves an unparseable model response yields an error and zero items,
// never a partial write.
func TestMalformedPayloadFailsClosed(t *testing.T) {
	if _, _, err := ValidateScopeCandidates([]byte("not json at all"), srcText); err == nil {
		t.Fatal("a malformed model payload must fail closed")
	}
}

// TestEmptySourceRejectsEverything proves that with no source text nothing can be grounded, so nothing is
// accepted. A normalization run against an empty notice writes no scope.
func TestEmptySourceRejectsEverything(t *testing.T) {
	raw := []byte(`{"items":[{"scope_type":"task","title":"x","quote":"anything at all"}]}`)
	got, _, err := ValidateScopeCandidates(raw, "")
	if err != nil {
		t.Fatalf("unexpected error: %v", err)
	}
	if len(got) != 0 {
		t.Fatal("nothing can be grounded against an empty source")
	}
}