A dispute carries both readings with their own spelling; differences separated by a single shared word are one decision, two shared words keep them apart; compound splits/joins and a short run of filler words are resolved without a model, while numbers, content words, and 'nej' (an answer) are not. Red: FindDisputes, AutoResolve and ContextAround are undefined. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
132 lines
4.7 KiB
Go
132 lines
4.7 KiB
Go
package reconcile_test
|
|
|
|
import (
|
|
"testing"
|
|
|
|
"github.com/stretchr/testify/assert"
|
|
"github.com/stretchr/testify/require"
|
|
|
|
"git.d-ma.be/mathias/tapir/internal/reconcile"
|
|
)
|
|
|
|
func disputes(base, alt string) []reconcile.Dispute {
|
|
return reconcile.FindDisputes(reconcile.Tokenize(base), reconcile.Tokenize(alt))
|
|
}
|
|
|
|
func TestFindDisputes_NoneWhenTheTextsAgreeWordForWord(t *testing.T) {
|
|
assert.Empty(t, disputes("Vi går igenom läget.", "vi går igenom läget"),
|
|
"casing and punctuation are not disagreements")
|
|
}
|
|
|
|
func TestFindDisputes_OneReplacedWordIsOneDisputeCarryingBothReadings(t *testing.T) {
|
|
got := disputes("Vi ska onboardade kunderna.", "Vi ska omvårdade kunderna.")
|
|
|
|
require.Len(t, got, 1)
|
|
d := got[0]
|
|
assert.Equal(t, "onboardade", d.Base)
|
|
assert.Equal(t, "omvårdade", d.Alt)
|
|
assert.Equal(t, 2, d.BaseStart)
|
|
assert.Equal(t, 3, d.BaseEnd)
|
|
assert.Equal(t, 2, d.AltStart)
|
|
assert.Equal(t, 3, d.AltEnd)
|
|
}
|
|
|
|
func TestFindDisputes_ReadingsKeepTheirOwnSpellingAndPunctuation(t *testing.T) {
|
|
got := disputes("Det blir T+1, alltså.", "Det blir T 1, alltså.")
|
|
|
|
require.Len(t, got, 1)
|
|
assert.Equal(t, "T+1,", got[0].Base)
|
|
assert.Equal(t, "T 1,", got[0].Alt, "the alternative's words replace the span as the engine wrote them")
|
|
}
|
|
|
|
func TestFindDisputes_AWordOnlyOneEngineHeardHasAnEmptyOtherSide(t *testing.T) {
|
|
got := disputes("Man får tillbaka ett förslag.", "Man får tillbaka ett coolt förslag.")
|
|
|
|
require.Len(t, got, 1)
|
|
assert.Equal(t, "", got[0].Base)
|
|
assert.Equal(t, "coolt", got[0].Alt)
|
|
assert.Equal(t, got[0].BaseStart, got[0].BaseEnd, "an insertion has an empty base range at its position")
|
|
assert.Equal(t, 4, got[0].BaseStart)
|
|
}
|
|
|
|
func TestFindDisputes_TwoDifferencesSeparatedByOneSharedWordAreOneDispute(t *testing.T) {
|
|
// "x ... y" around a single agreed word is one decision for the reader, not two.
|
|
got := disputes("a b c d e", "a x c y e")
|
|
|
|
require.Len(t, got, 1)
|
|
assert.Equal(t, "b c d", got[0].Base)
|
|
assert.Equal(t, "x c y", got[0].Alt)
|
|
}
|
|
|
|
func TestFindDisputes_TwoSharedWordsKeepTheDisputesApart(t *testing.T) {
|
|
got := disputes("a b c d e f", "a x c d y f")
|
|
|
|
require.Len(t, got, 2)
|
|
assert.Equal(t, "b", got[0].Base)
|
|
assert.Equal(t, "e", got[1].Base)
|
|
}
|
|
|
|
func TestFindDisputes_ChainsOfCloseDifferencesMergeIntoOne(t *testing.T) {
|
|
got := disputes("a b c d e f g", "a x c y e z g")
|
|
|
|
require.Len(t, got, 1)
|
|
assert.Equal(t, "b c d e f", got[0].Base)
|
|
}
|
|
|
|
func TestFindDisputes_LeavesIndicesThatAddressBothTokenSlices(t *testing.T) {
|
|
base := reconcile.Tokenize("Hej Anders och Maria")
|
|
alt := reconcile.Tokenize("Hej Sanders och Maria")
|
|
|
|
got := reconcile.FindDisputes(base, alt)
|
|
|
|
require.Len(t, got, 1)
|
|
assert.Equal(t, "Anders", base[got[0].BaseStart].Text)
|
|
assert.Equal(t, "Sanders", alt[got[0].AltStart].Text)
|
|
}
|
|
|
|
func TestAutoResolve(t *testing.T) {
|
|
tests := []struct {
|
|
name string
|
|
base, alt string
|
|
resolved bool
|
|
}{
|
|
{"a compound written as two words or one", "i crossborder delen", "i crossborderdelen", true},
|
|
{"a compound with a hyphen", "crossborder-delen", "crossborder delen", true},
|
|
{"a filler only the alternative heard", "vi ska ju betala", "vi ska betala", true},
|
|
{"a filler only the base heard", "vi ska betala", "vi ska ju betala", true},
|
|
{"two fillers in a row", "vi ska ja ju betala", "vi ska betala", true},
|
|
{"a different word is for the model", "vi ska onboardade", "vi ska omvårdade", false},
|
|
{"a different number is for the model", "totalt 1050 kronor", "totalt 2050 kronor", false},
|
|
{"a content word only one side heard is for the model", "ett förslag", "ett coolt förslag", false},
|
|
{"three filler words are too many to wave through", "vi ska ju då liksom betala", "vi ska betala", false},
|
|
{"a filler swapped for a content word is for the model", "vi ska ju betala", "vi ska nu betala", false},
|
|
{"nej is an answer, not a filler", "men nej det går inte", "men det går inte", false},
|
|
}
|
|
for _, tt := range tests {
|
|
t.Run(tt.name, func(t *testing.T) {
|
|
got := disputes(tt.base, tt.alt)
|
|
require.Len(t, got, 1, "the fixture must be exactly one dispute")
|
|
|
|
assert.Equal(t, tt.resolved, reconcile.AutoResolve(got[0]))
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestContextAround_TakesWordsFromTheBaseOnBothSidesAndStopsAtTheEdges(t *testing.T) {
|
|
base := reconcile.Tokenize("ett två tre fyra fem sex sju åtta nio tio")
|
|
d := reconcile.Dispute{BaseStart: 4, BaseEnd: 6} // "fem sex"
|
|
|
|
before, after := reconcile.ContextAround(base, d, 3)
|
|
|
|
assert.Equal(t, "två tre fyra", before)
|
|
assert.Equal(t, "sju åtta nio", after)
|
|
|
|
b2, a2 := reconcile.ContextAround(base, reconcile.Dispute{BaseStart: 0, BaseEnd: 1}, 3)
|
|
assert.Equal(t, "", b2)
|
|
assert.Equal(t, "två tre fyra", a2)
|
|
|
|
b3, a3 := reconcile.ContextAround(base, reconcile.Dispute{BaseStart: 9, BaseEnd: 10}, 3)
|
|
assert.Equal(t, "sju åtta nio", b3)
|
|
assert.Equal(t, "", a3)
|
|
}
|