Implemented an initial transcript merge stage

This commit is contained in:
2026-04-26 13:57:13 -05:00
parent c32349a017
commit 18f1873776
8 changed files with 535 additions and 28 deletions

View File

@@ -11,11 +11,25 @@ import (
"gitea.maximumdirect.net/eric/seriatim/internal/report"
)
func TestMergeWritesPlaceholderOutputAndReport(t *testing.T) {
func TestMergeWritesMergedOutputAndReport(t *testing.T) {
dir := t.TempDir()
inputA := writeFile(t, dir, "a.json")
inputB := writeFile(t, dir, "b.json")
speakers := writeFile(t, dir, "speakers.yml")
inputA := writeJSONFile(t, dir, "a.json", `{
"segments": [
{"start": 10, "end": 11, "text": " second a ", "words": [{"word": "ignored"}]},
{"start": 1, "end": 2, "text": "first a"}
]
}`)
inputB := writeJSONFile(t, dir, "b.json", `{
"segments": [
{"start": 5, "end": 6, "text": "first b"}
]
}`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
a.json:
speaker: Alice
b.json:
speaker: Bob
`)
output := filepath.Join(dir, "merged.json")
reportPath := filepath.Join(dir, "report.json")
@@ -37,9 +51,6 @@ func TestMergeWritesPlaceholderOutputAndReport(t *testing.T) {
t.Fatalf("read output bytes: %v", err)
}
outputJSON := string(outputBytes)
if !strings.Contains(outputJSON, `"segments": []`) {
t.Fatalf("expected segments to serialize as an empty array, got:\n%s", outputJSON)
}
if !strings.Contains(outputJSON, `"overlap_groups": []`) {
t.Fatalf("expected overlap_groups to serialize as an empty array, got:\n%s", outputJSON)
}
@@ -49,8 +60,17 @@ func TestMergeWritesPlaceholderOutputAndReport(t *testing.T) {
if got, want := transcript.Metadata.InputFiles, []string{inputA, inputB}; !equalStrings(got, want) {
t.Fatalf("input files not sorted deterministically: got %v want %v", got, want)
}
if len(transcript.Segments) != 0 {
t.Fatalf("expected placeholder output to contain no segments, got %d", len(transcript.Segments))
if got, want := len(transcript.Segments), 3; got != want {
t.Fatalf("expected merged output to contain %d segments, got %d", want, got)
}
assertSegment(t, transcript.Segments[0], 1, inputA, 1, "Alice", 1, 2, "first a")
assertSegment(t, transcript.Segments[1], 2, inputB, 0, "Bob", 5, 6, "first b")
assertSegment(t, transcript.Segments[2], 3, inputA, 0, "Alice", 10, 11, "second a")
if strings.Contains(outputJSON, "internal_ref") {
t.Fatalf("did not expect internal_ref in output:\n%s", outputJSON)
}
if strings.Contains(outputJSON, "words") {
t.Fatalf("did not expect words in output:\n%s", outputJSON)
}
if len(transcript.OverlapGroups) != 0 {
t.Fatalf("expected placeholder output to contain no overlap groups, got %d", len(transcript.OverlapGroups))
@@ -79,10 +99,62 @@ func TestMergeWritesPlaceholderOutputAndReport(t *testing.T) {
}
}
func TestMergeTieBreakOrder(t *testing.T) {
dir := t.TempDir()
inputA := writeJSONFile(t, dir, "a.json", `{
"segments": [
{"start": 1, "end": 4, "text": "a-late-end"},
{"start": 1, "end": 2, "text": "a-index-one"}
]
}`)
inputB := writeJSONFile(t, dir, "b.json", `{
"segments": [
{"start": 1, "end": 2, "text": "b-same-time"}
]
}`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
a.json:
speaker: Alice
b.json:
speaker: Bob
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
"--input-file", inputB,
"--input-file", inputA,
"--speakers", speakers,
"--output-file", output,
)
if err != nil {
t.Fatalf("merge failed: %v", err)
}
var transcript model.FinalTranscript
readJSON(t, output, &transcript)
got := []string{
transcript.Segments[0].Text,
transcript.Segments[1].Text,
transcript.Segments[2].Text,
}
want := []string{"a-index-one", "b-same-time", "a-late-end"}
if !equalStrings(got, want) {
t.Fatalf("tie-break order mismatch: got %v want %v", got, want)
}
for index, segment := range transcript.Segments {
if segment.ID != index+1 {
t.Fatalf("segment %d has id %d; want %d", index, segment.ID, index+1)
}
}
}
func TestUnknownModulesFailDuringValidation(t *testing.T) {
dir := t.TempDir()
input := writeFile(t, dir, "input.json")
speakers := writeFile(t, dir, "speakers.yml")
input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
input.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
tests := []struct {
@@ -134,7 +206,7 @@ func TestUnknownModulesFailDuringValidation(t *testing.T) {
func TestInvalidPreprocessingOrderFails(t *testing.T) {
dir := t.TempDir()
input := writeFile(t, dir, "input.json")
input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
@@ -152,7 +224,10 @@ func TestInvalidPreprocessingOrderFails(t *testing.T) {
func TestMissingInputFileFailsBeforePipelineExecution(t *testing.T) {
dir := t.TempDir()
speakers := writeFile(t, dir, "speakers.yml")
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
missing.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
@@ -170,7 +245,7 @@ func TestMissingInputFileFailsBeforePipelineExecution(t *testing.T) {
func TestNormalizeSpeakersRequiresSpeakersFile(t *testing.T) {
dir := t.TempDir()
input := writeFile(t, dir, "input.json")
input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
@@ -187,8 +262,11 @@ func TestNormalizeSpeakersRequiresSpeakersFile(t *testing.T) {
func TestAutocorrectRequiresAutocorrectFile(t *testing.T) {
dir := t.TempDir()
input := writeFile(t, dir, "input.json")
speakers := writeFile(t, dir, "speakers.yml")
input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
input.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
@@ -207,9 +285,14 @@ func TestAutocorrectRequiresAutocorrectFile(t *testing.T) {
func TestOutputJSONIsByteStable(t *testing.T) {
dir := t.TempDir()
inputA := writeFile(t, dir, "a.json")
inputB := writeFile(t, dir, "b.json")
speakers := writeFile(t, dir, "speakers.yml")
inputA := writeJSONFile(t, dir, "a.json", `{"segments":[{"start":2,"end":3,"text":"a"}]}`)
inputB := writeJSONFile(t, dir, "b.json", `{"segments":[{"start":1,"end":2,"text":"b"}]}`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
a.json:
speaker: Alice
b.json:
speaker: Bob
`)
outputA := filepath.Join(dir, "merged-a.json")
outputB := filepath.Join(dir, "merged-b.json")
@@ -241,17 +324,192 @@ func TestOutputJSONIsByteStable(t *testing.T) {
}
}
func TestMissingSpeakerMappingFails(t *testing.T) {
dir := t.TempDir()
input := writeJSONFile(t, dir, "input.json", `{"segments":[]}`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
other.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
"--input-file", input,
"--speakers", speakers,
"--output-file", output,
)
if err == nil {
t.Fatal("expected error")
}
if !strings.Contains(err.Error(), `speaker map has no entry for input basename "input.json"`) {
t.Fatalf("unexpected error: %v", err)
}
}
func TestMalformedJSONFails(t *testing.T) {
dir := t.TempDir()
input := writeJSONFile(t, dir, "input.json", `{"segments":[`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
input.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
"--input-file", input,
"--speakers", speakers,
"--output-file", output,
)
if err == nil {
t.Fatal("expected error")
}
if !strings.Contains(err.Error(), "parse input file") {
t.Fatalf("unexpected error: %v", err)
}
}
func TestMissingTopLevelSegmentsFails(t *testing.T) {
dir := t.TempDir()
input := writeJSONFile(t, dir, "input.json", `{}`)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
input.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
"--input-file", input,
"--speakers", speakers,
"--output-file", output,
)
if err == nil {
t.Fatal("expected error")
}
if !strings.Contains(err.Error(), "must contain top-level segments array") {
t.Fatalf("unexpected error: %v", err)
}
}
func TestInvalidSegmentFieldsFailWithSourceAndIndex(t *testing.T) {
tests := []struct {
name string
json string
want string
}{
{
name: "missing start",
json: `{"segments":[{"end":1,"text":"x"}]}`,
want: "segment 0 missing numeric start",
},
{
name: "wrong typed end",
json: `{"segments":[{"start":0,"end":"1","text":"x"}]}`,
want: "segment 0 end must be numeric",
},
{
name: "wrong typed text",
json: `{"segments":[{"start":0,"end":1,"text":7}]}`,
want: "segment 0 text must be a string",
},
{
name: "null text",
json: `{"segments":[{"start":0,"end":1,"text":null}]}`,
want: "segment 0 missing string text",
},
}
for _, test := range tests {
t.Run(test.name, func(t *testing.T) {
dir := t.TempDir()
input := writeJSONFile(t, dir, "input.json", test.json)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
input.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
"--input-file", input,
"--speakers", speakers,
"--output-file", output,
)
if err == nil {
t.Fatal("expected error")
}
if !strings.Contains(err.Error(), input) {
t.Fatalf("expected error to contain source path %q, got %v", input, err)
}
if !strings.Contains(err.Error(), test.want) {
t.Fatalf("expected error to contain %q, got %v", test.want, err)
}
})
}
}
func TestInvalidTimingFails(t *testing.T) {
tests := []struct {
name string
json string
want string
}{
{
name: "negative start",
json: `{"segments":[{"start":-1,"end":1,"text":"x"}]}`,
want: "segment 0 has negative start",
},
{
name: "end before start",
json: `{"segments":[{"start":2,"end":1,"text":"x"}]}`,
want: "segment 0 has end before start",
},
}
for _, test := range tests {
t.Run(test.name, func(t *testing.T) {
dir := t.TempDir()
input := writeJSONFile(t, dir, "input.json", test.json)
speakers := writeYAMLFile(t, dir, "speakers.yml", `inputs:
input.json:
speaker: Alice
`)
output := filepath.Join(dir, "merged.json")
err := executeMerge(
"--input-file", input,
"--speakers", speakers,
"--output-file", output,
)
if err == nil {
t.Fatal("expected error")
}
if !strings.Contains(err.Error(), test.want) {
t.Fatalf("expected error to contain %q, got %v", test.want, err)
}
})
}
}
func executeMerge(args ...string) error {
cmd := NewRootCommand()
cmd.SetArgs(append([]string{"merge"}, args...))
return cmd.Execute()
}
func writeFile(t *testing.T, dir string, name string) string {
func writeJSONFile(t *testing.T, dir string, name string, content string) string {
t.Helper()
path := filepath.Join(dir, name)
if err := os.WriteFile(path, []byte("{}\n"), 0o600); err != nil {
if err := os.WriteFile(path, []byte(content+"\n"), 0o600); err != nil {
t.Fatalf("write file: %v", err)
}
return path
}
func writeYAMLFile(t *testing.T, dir string, name string, content string) string {
t.Helper()
path := filepath.Join(dir, name)
if err := os.WriteFile(path, []byte(content), 0o600); err != nil {
t.Fatalf("write file: %v", err)
}
return path
@@ -280,3 +538,29 @@ func equalStrings(left []string, right []string) bool {
}
return true
}
func assertSegment(t *testing.T, segment model.Segment, id int, source string, sourceIndex int, speaker string, start float64, end float64, text string) {
t.Helper()
if segment.ID != id {
t.Fatalf("segment ID = %d, want %d", segment.ID, id)
}
if segment.Source != source {
t.Fatalf("segment source = %q, want %q", segment.Source, source)
}
if segment.SourceSegmentIndex != sourceIndex {
t.Fatalf("segment source index = %d, want %d", segment.SourceSegmentIndex, sourceIndex)
}
if segment.Speaker != speaker {
t.Fatalf("segment speaker = %q, want %q", segment.Speaker, speaker)
}
if segment.Start != start {
t.Fatalf("segment start = %f, want %f", segment.Start, start)
}
if segment.End != end {
t.Fatalf("segment end = %f, want %f", segment.End, end)
}
if segment.Text != text {
t.Fatalf("segment text = %q, want %q", segment.Text, text)
}
}