Implemented the initial Go framework

internal/builtin/input.go

@@ -0,0 +1,31 @@
+package builtin
+
+import (
+	"context"
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/model"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+)
+
+type jsonFilesReader struct{}
+
+func (jsonFilesReader) Name() string {
+	return "json-files"
+}
+
+func (jsonFilesReader) Read(ctx context.Context, cfg config.Config) ([]model.RawTranscript, []report.Event, error) {
+	if err := ctx.Err(); err != nil {
+		return nil, nil, err
+	}
+
+	raw := make([]model.RawTranscript, 0, len(cfg.InputFiles))
+	for _, inputFile := range cfg.InputFiles {
+		raw = append(raw, model.RawTranscript{Source: inputFile})
+	}
+
+	return raw, []report.Event{
+		report.Info("input", "json-files", fmt.Sprintf("accepted %d input file(s)", len(raw))),
+	}, nil
+}

internal/builtin/merge.go

@@ -0,0 +1,52 @@
+package builtin
+
+import (
+	"context"
+	"sort"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/model"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+)
+
+type placeholderMerger struct{}
+
+func (placeholderMerger) Name() string {
+	return "placeholder-merger"
+}
+
+func (placeholderMerger) Merge(ctx context.Context, in []model.CanonicalTranscript, cfg config.Config) (model.MergedTranscript, []report.Event, error) {
+	if err := ctx.Err(); err != nil {
+		return model.MergedTranscript{}, nil, err
+	}
+
+	segments := make([]model.Segment, 0)
+	for _, transcript := range in {
+		segments = append(segments, transcript.Segments...)
+	}
+
+	sort.SliceStable(segments, func(i, j int) bool {
+		left := segments[i]
+		right := segments[j]
+		if left.Start != right.Start {
+			return left.Start < right.Start
+		}
+		if left.End != right.End {
+			return left.End < right.End
+		}
+		if left.Source != right.Source {
+			return left.Source < right.Source
+		}
+		if left.SourceSegmentIndex != right.SourceSegmentIndex {
+			return left.SourceSegmentIndex < right.SourceSegmentIndex
+		}
+		return left.Speaker < right.Speaker
+	})
+
+	return model.MergedTranscript{
+			Segments:      segments,
+			OverlapGroups: nil,
+		}, []report.Event{
+			report.Info("merge", "placeholder-merger", "merged placeholder canonical transcript(s)"),
+		}, nil
+}

internal/builtin/output.go

@@ -0,0 +1,39 @@
+package builtin
+
+import (
+	"context"
+	"encoding/json"
+	"os"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/model"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+)
+
+type jsonOutputWriter struct{}
+
+func (jsonOutputWriter) Name() string {
+	return "json"
+}
+
+func (jsonOutputWriter) Write(ctx context.Context, out model.FinalTranscript, rpt report.Report, cfg config.Config) ([]report.Event, error) {
+	if err := ctx.Err(); err != nil {
+		return nil, err
+	}
+
+	file, err := os.Create(cfg.OutputFile)
+	if err != nil {
+		return nil, err
+	}
+	defer file.Close()
+
+	enc := json.NewEncoder(file)
+	enc.SetIndent("", "  ")
+	if err := enc.Encode(out); err != nil {
+		return nil, err
+	}
+
+	return []report.Event{
+		report.Info("output", "json", "wrote placeholder transcript JSON"),
+	}, nil
+}

internal/builtin/postprocess.go

@@ -0,0 +1,47 @@
+package builtin
+
+import (
+	"context"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/model"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+)
+
+type noopPostprocessor struct {
+	name string
+}
+
+func (p noopPostprocessor) Name() string {
+	return p.name
+}
+
+func (p noopPostprocessor) Process(ctx context.Context, in model.MergedTranscript, cfg config.Config) (model.MergedTranscript, []report.Event, error) {
+	if err := ctx.Err(); err != nil {
+		return model.MergedTranscript{}, nil, err
+	}
+
+	return in, []report.Event{
+		report.Info("postprocessing", p.name, "completed no-op postprocessing module"),
+	}, nil
+}
+
+type assignIDs struct{}
+
+func (assignIDs) Name() string {
+	return "assign-ids"
+}
+
+func (assignIDs) Process(ctx context.Context, in model.MergedTranscript, cfg config.Config) (model.MergedTranscript, []report.Event, error) {
+	if err := ctx.Err(); err != nil {
+		return model.MergedTranscript{}, nil, err
+	}
+
+	for index := range in.Segments {
+		in.Segments[index].ID = index + 1
+	}
+
+	return in, []report.Event{
+		report.Info("postprocessing", "assign-ids", "assigned final segment IDs"),
+	}, nil
+}

internal/builtin/preprocess.go

@@ -0,0 +1,82 @@
+package builtin
+
+import (
+	"context"
+	"fmt"
+
+	"gitea.maximumdirect.net/eric/seriatim/internal/config"
+	"gitea.maximumdirect.net/eric/seriatim/internal/model"
+	"gitea.maximumdirect.net/eric/seriatim/internal/pipeline"
+	"gitea.maximumdirect.net/eric/seriatim/internal/report"
+)
+
+type noopPreprocessor struct {
+	name     string
+	requires pipeline.ModelState
+	produces pipeline.ModelState
+}
+
+func (p noopPreprocessor) Name() string {
+	return p.name
+}
+
+func (p noopPreprocessor) Requires() pipeline.ModelState {
+	return p.requires
+}
+
+func (p noopPreprocessor) Produces() pipeline.ModelState {
+	return p.produces
+}
+
+func (p noopPreprocessor) Process(ctx context.Context, in pipeline.PreprocessState, cfg config.Config) (pipeline.PreprocessState, []report.Event, error) {
+	if err := ctx.Err(); err != nil {
+		return pipeline.PreprocessState{}, nil, err
+	}
+	if in.State != p.requires {
+		return pipeline.PreprocessState{}, nil, fmt.Errorf("preprocessing module %q requires state %q but received %q", p.name, p.requires, in.State)
+	}
+
+	in.State = p.produces
+	return in, []report.Event{
+		report.Info("preprocessing", p.name, "completed no-op preprocessing module"),
+	}, nil
+}
+
+type normalizeSpeakers struct{}
+
+func (normalizeSpeakers) Name() string {
+	return "normalize-speakers"
+}
+
+func (normalizeSpeakers) Requires() pipeline.ModelState {
+	return pipeline.StateRaw
+}
+
+func (normalizeSpeakers) Produces() pipeline.ModelState {
+	return pipeline.StateCanonical
+}
+
+func (normalizeSpeakers) Process(ctx context.Context, in pipeline.PreprocessState, cfg config.Config) (pipeline.PreprocessState, []report.Event, error) {
+	if err := ctx.Err(); err != nil {
+		return pipeline.PreprocessState{}, nil, err
+	}
+	if in.State != pipeline.StateRaw {
+		return pipeline.PreprocessState{}, nil, fmt.Errorf("preprocessing module %q requires state %q but received %q", "normalize-speakers", pipeline.StateRaw, in.State)
+	}
+
+	canonical := make([]model.CanonicalTranscript, 0, len(in.Raw))
+	for _, raw := range in.Raw {
+		canonical = append(canonical, model.CanonicalTranscript{
+			Source:   raw.Source,
+			Segments: nil,
+		})
+	}
+
+	return pipeline.PreprocessState{
+			State:     pipeline.StateCanonical,
+			Raw:       append([]model.RawTranscript(nil), in.Raw...),
+			Canonical: canonical,
+		}, []report.Event{
+			report.Info("preprocessing", "normalize-speakers", "created placeholder canonical transcript(s)"),
+		}, nil
+}

internal/builtin/registry.go

@@ -0,0 +1,23 @@
+package builtin
+
+import "gitea.maximumdirect.net/eric/seriatim/internal/pipeline"
+
+// NewRegistry registers the MVP built-in modules.
+func NewRegistry() *pipeline.Registry {
+	registry := pipeline.NewRegistry()
+
+	registry.RegisterInputReader(jsonFilesReader{})
+	registry.RegisterPreprocessor(noopPreprocessor{name: "validate-raw", requires: pipeline.StateRaw, produces: pipeline.StateRaw})
+	registry.RegisterPreprocessor(normalizeSpeakers{})
+	registry.RegisterPreprocessor(noopPreprocessor{name: "trim-text", requires: pipeline.StateCanonical, produces: pipeline.StateCanonical})
+	registry.RegisterPreprocessor(noopPreprocessor{name: "autocorrect", requires: pipeline.StateCanonical, produces: pipeline.StateCanonical})
+	registry.RegisterMerger(placeholderMerger{})
+	registry.RegisterPostprocessor(noopPostprocessor{name: "detect-overlaps"})
+	registry.RegisterPostprocessor(noopPostprocessor{name: "resolve-overlaps"})
+	registry.RegisterPostprocessor(assignIDs{})
+	registry.RegisterPostprocessor(noopPostprocessor{name: "validate-output"})
+	registry.RegisterPostprocessor(noopPostprocessor{name: "autocorrect"})
+	registry.RegisterOutputWriter(jsonOutputWriter{})
+
+	return registry
+}