Bugfixes and documentation cleanup for v1.0 release.

internal/config/config.go

@@ -13,22 +13,23 @@ import (
 const (
 	DefaultInputReader            = "json-files"
 	DefaultOutputModules          = "json"
-	DefaultOutputSchema           = OutputSchemaDefault
+	DefaultOutputSchema           = OutputSchemaIntermediate
 	DefaultPreprocessingModules   = "validate-raw,normalize-speakers,trim-text"
 	DefaultPostprocessingModules  = "detect-overlaps,resolve-overlaps,backchannel,filler,resolve-danglers,coalesce,detect-overlaps,autocorrect,assign-ids,validate-output"
-	DefaultOverlapWordRunGap      = 0.75
+	DefaultOverlapWordRunGap      = 1.0
 	DefaultWordRunReorderWindow   = 1.0
 	DefaultCoalesceGap            = 3.0
 	DefaultCoalesceGapValue       = "3.0"
 	DefaultBackchannelMaxDuration = 2.0
 	DefaultFillerMaxDuration      = 1.25
+	OutputSchemaEnv               = "SERIATIM_OUTPUT_SCHEMA"
 	OverlapWordRunGapEnv          = "SERIATIM_OVERLAP_WORD_RUN_GAP"
 	WordRunReorderWindowEnv       = "SERIATIM_OVERLAP_WORD_RUN_REORDER_WINDOW"
 	BackchannelMaxDurationEnv     = "SERIATIM_BACKCHANNEL_MAX_DURATION"
 	FillerMaxDurationEnv          = "SERIATIM_FILLER_MAX_DURATION"
-	OutputSchemaDefault           = "default"
-	OutputSchemaSeriatim          = "seriatim"
-	OutputSchemaMinimal           = "minimal"
+	OutputSchemaMinimal           = "seriatim-minimal"
+	OutputSchemaIntermediate      = "seriatim-intermediate"
+	OutputSchemaFull              = "seriatim-full"
 )
 
 // MergeOptions captures raw CLI option values before validation.
@@ -70,7 +71,7 @@ func NewMergeConfig(opts MergeOptions) (Config, error) {
 	cfg := Config{
 		InputReader:            strings.TrimSpace(opts.InputReader),
 		OutputModules:          nil,
-		OutputSchema:           strings.TrimSpace(opts.OutputSchema),
+		OutputSchema:           "",
 		PreprocessingModules:   nil,
 		PostprocessingModules:  nil,
 		OverlapWordRunGap:      DefaultOverlapWordRunGap,
@@ -83,14 +84,12 @@ func NewMergeConfig(opts MergeOptions) (Config, error) {
 	if cfg.InputReader == "" {
 		return Config{}, errors.New("--input-reader is required")
 	}
-	if cfg.OutputSchema == "" {
-		cfg.OutputSchema = DefaultOutputSchema
-	}
-	if err := validateOutputSchema(cfg.OutputSchema); err != nil {
+	var err error
+	cfg.OutputSchema, err = resolveOutputSchema(opts.OutputSchema)
+	if err != nil {
 		return Config{}, err
 	}
 
-	var err error
 	cfg.OutputModules, err = parseModuleList(opts.OutputModules)
 	if err != nil {
 		return Config{}, fmt.Errorf("--output-modules: %w", err)
@@ -189,13 +188,27 @@ func parseModuleList(value string) ([]string, error) {
 
 func validateOutputSchema(value string) error {
 	switch value {
-	case OutputSchemaDefault, OutputSchemaSeriatim, OutputSchemaMinimal:
+	case OutputSchemaMinimal, OutputSchemaIntermediate, OutputSchemaFull:
 		return nil
 	default:
-		return fmt.Errorf("--output-schema must be one of %q, %q, or %q", OutputSchemaDefault, OutputSchemaMinimal, OutputSchemaSeriatim)
+		return fmt.Errorf("--output-schema must be one of %q, %q, or %q", OutputSchemaMinimal, OutputSchemaIntermediate, OutputSchemaFull)
 	}
 }
 
+func resolveOutputSchema(value string) (string, error) {
+	value = strings.TrimSpace(value)
+	if value == "" {
+		value = strings.TrimSpace(os.Getenv(OutputSchemaEnv))
+	}
+	if value == "" {
+		value = DefaultOutputSchema
+	}
+	if err := validateOutputSchema(value); err != nil {
+		return "", err
+	}
+	return value, nil
+}
+
 func normalizeInputFiles(paths []string) ([]string, error) {
 	if len(paths) == 0 {
 		return nil, errors.New("at least one --input-file is required")