Simplify the CLI interface and update documentation accordingly

internal/builtin/postprocess.go

@@ -58,6 +58,11 @@ func (autocorrectPostprocessor) Process(ctx context.Context, in model.MergedTran
 	if err := ctx.Err(); err != nil {
 		return model.MergedTranscript{}, nil, err
 	}
+	if cfg.AutocorrectFile == "" {
+		return in, []report.Event{
+			report.Info("postprocessing", "autocorrect", "skipped autocorrect because no autocorrect file was supplied"),
+		}, nil
+	}
 
 	rules, err := autocorrect.Load(cfg.AutocorrectFile)
 	if err != nil {

internal/builtin/preprocess.go

@@ -3,6 +3,7 @@ package builtin
 import (
 	"context"
 	"fmt"
+	"path/filepath"
 	"strings"
 
 	"gitea.maximumdirect.net/eric/seriatim/internal/config"
@@ -99,16 +100,25 @@ func (normalizeSpeakers) Process(ctx context.Context, in pipeline.PreprocessStat
 		return pipeline.PreprocessState{}, nil, fmt.Errorf("preprocessing module %q requires state %q but received %q", "normalize-speakers", pipeline.StateRaw, in.State)
 	}
 
-	speakers, err := speaker.LoadMap(cfg.SpeakersFile)
-	if err != nil {
-		return pipeline.PreprocessState{}, nil, err
+	var speakers speaker.Map
+	useSpeakerMap := cfg.SpeakersFile != ""
+	if useSpeakerMap {
+		var err error
+		speakers, err = speaker.LoadMap(cfg.SpeakersFile)
+		if err != nil {
+			return pipeline.PreprocessState{}, nil, err
+		}
 	}
 
 	canonical := make([]model.CanonicalTranscript, 0, len(in.Raw))
 	for _, raw := range in.Raw {
-		canonicalSpeaker, err := speakers.SpeakerForSource(raw.Source)
-		if err != nil {
-			return pipeline.PreprocessState{}, nil, err
+		canonicalSpeaker := filepath.Base(raw.Source)
+		if useSpeakerMap {
+			var err error
+			canonicalSpeaker, err = speakers.SpeakerForSource(raw.Source)
+			if err != nil {
+				return pipeline.PreprocessState{}, nil, err
+			}
 		}
 
 		segments := make([]model.Segment, 0, len(raw.Segments))
@@ -129,11 +139,16 @@ func (normalizeSpeakers) Process(ctx context.Context, in pipeline.PreprocessStat
 		})
 	}
 
+	message := "created canonical transcript(s) from raw input"
+	if !useSpeakerMap {
+		message = "created canonical transcript(s) using input basenames as speaker labels"
+	}
+
 	return pipeline.PreprocessState{
 			State:     pipeline.StateCanonical,
 			Raw:       append([]model.RawTranscript(nil), in.Raw...),
 			Canonical: canonical,
 		}, []report.Event{
-			report.Info("preprocessing", "normalize-speakers", "created canonical transcript(s) from raw input"),
+			report.Info("preprocessing", "normalize-speakers", message),
 		}, nil
 }