relspecgo/cmd/relspec/merge.go

package main

import (
	"fmt"
	"os"
	"path/filepath"
	"strings"

	"github.com/spf13/cobra"

	"git.warky.dev/wdevs/relspecgo/pkg/merge"
	"git.warky.dev/wdevs/relspecgo/pkg/models"
	"git.warky.dev/wdevs/relspecgo/pkg/readers"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/bun"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/dbml"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/dctx"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/drawdb"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/drizzle"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/gorm"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/graphql"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/json"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/pgsql"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/prisma"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/typeorm"
	"git.warky.dev/wdevs/relspecgo/pkg/readers/yaml"
	"git.warky.dev/wdevs/relspecgo/pkg/writers"
	wbun "git.warky.dev/wdevs/relspecgo/pkg/writers/bun"
	wdbml "git.warky.dev/wdevs/relspecgo/pkg/writers/dbml"
	wdctx "git.warky.dev/wdevs/relspecgo/pkg/writers/dctx"
	wdrawdb "git.warky.dev/wdevs/relspecgo/pkg/writers/drawdb"
	wdrizzle "git.warky.dev/wdevs/relspecgo/pkg/writers/drizzle"
	wgorm "git.warky.dev/wdevs/relspecgo/pkg/writers/gorm"
	wgraphql "git.warky.dev/wdevs/relspecgo/pkg/writers/graphql"
	wjson "git.warky.dev/wdevs/relspecgo/pkg/writers/json"
	wpgsql "git.warky.dev/wdevs/relspecgo/pkg/writers/pgsql"
	wprisma "git.warky.dev/wdevs/relspecgo/pkg/writers/prisma"
	wtypeorm "git.warky.dev/wdevs/relspecgo/pkg/writers/typeorm"
	wyaml "git.warky.dev/wdevs/relspecgo/pkg/writers/yaml"
)

var (
	mergeTargetType    string
	mergeTargetPath    string
	mergeTargetConn    string
	mergeSourceType    string
	mergeSourcePath    string
	mergeSourceConn    string
	mergeOutputType    string
	mergeOutputPath    string
	mergeOutputConn    string
	mergeSkipDomains   bool
	mergeSkipRelations bool
	mergeSkipEnums     bool
	mergeSkipViews     bool
	mergeSkipSequences bool
	mergeSkipTables    string // Comma-separated table names to skip
	mergeVerbose       bool
)

var mergeCmd = &cobra.Command{
	Use:   "merge",
	Short: "Merge database schemas (additive only - adds missing items)",
	Long: `Merge one database schema into another. Performs additive merging only:
adds missing schemas, tables, columns, and other objects without modifying
or deleting existing items.

The target database is loaded first, then the source database is merged into it.
The result can be saved to a new format or updated in place.

Examples:
  # Merge two JSON schemas
  relspec merge --target json --target-path base.json \
                --source json --source-path additional.json \
                --output json --output-path merged.json

  # Merge from PostgreSQL into JSON
  relspec merge --target json --target-path mydb.json \
                --source pgsql --source-conn "postgres://user:pass@localhost/source_db" \
                --output json --output-path combined.json

  # Merge DBML and YAML, skip relations
  relspec merge --target dbml --target-path schema.dbml \
                --source yaml --source-path tables.yaml \
                --output dbml --output-path merged.dbml \
                --skip-relations

  # Merge and save back to target format
  relspec merge --target json --target-path base.json \
                --source json --source-path patch.json \
                --output json --output-path base.json`,
	RunE: runMerge,
}

func init() {
	// Target database flags
	mergeCmd.Flags().StringVar(&mergeTargetType, "target", "", "Target format (required): dbml, dctx, drawdb, graphql, json, yaml, gorm, bun, drizzle, prisma, typeorm, pgsql")
	mergeCmd.Flags().StringVar(&mergeTargetPath, "target-path", "", "Target file path (required for file-based formats)")
	mergeCmd.Flags().StringVar(&mergeTargetConn, "target-conn", "", "Target connection string (required for pgsql)")

	// Source database flags
	mergeCmd.Flags().StringVar(&mergeSourceType, "source", "", "Source format (required): dbml, dctx, drawdb, graphql, json, yaml, gorm, bun, drizzle, prisma, typeorm, pgsql")
	mergeCmd.Flags().StringVar(&mergeSourcePath, "source-path", "", "Source file path (required for file-based formats)")
	mergeCmd.Flags().StringVar(&mergeSourceConn, "source-conn", "", "Source connection string (required for pgsql)")

	// Output flags
	mergeCmd.Flags().StringVar(&mergeOutputType, "output", "", "Output format (required): dbml, dctx, drawdb, graphql, json, yaml, gorm, bun, drizzle, prisma, typeorm, pgsql")
	mergeCmd.Flags().StringVar(&mergeOutputPath, "output-path", "", "Output file path (required for file-based formats)")
	mergeCmd.Flags().StringVar(&mergeOutputConn, "output-conn", "", "Output connection string (for pgsql)")

	// Merge options
	mergeCmd.Flags().BoolVar(&mergeSkipDomains, "skip-domains", false, "Skip domains during merge")
	mergeCmd.Flags().BoolVar(&mergeSkipRelations, "skip-relations", false, "Skip relations during merge")
	mergeCmd.Flags().BoolVar(&mergeSkipEnums, "skip-enums", false, "Skip enums during merge")
	mergeCmd.Flags().BoolVar(&mergeSkipViews, "skip-views", false, "Skip views during merge")
	mergeCmd.Flags().BoolVar(&mergeSkipSequences, "skip-sequences", false, "Skip sequences during merge")
	mergeCmd.Flags().StringVar(&mergeSkipTables, "skip-tables", "", "Comma-separated list of table names to skip during merge")
	mergeCmd.Flags().BoolVar(&mergeVerbose, "verbose", false, "Show verbose output")
}

func runMerge(cmd *cobra.Command, args []string) error {
	fmt.Fprintf(os.Stderr, "\n=== RelSpec Merge ===\n")
	fmt.Fprintf(os.Stderr, "Started at: %s\n\n", getCurrentTimestamp())

	// Validate required flags
	if mergeTargetType == "" {
		return fmt.Errorf("--target format is required")
	}
	if mergeSourceType == "" {
		return fmt.Errorf("--source format is required")
	}
	if mergeOutputType == "" {
		return fmt.Errorf("--output format is required")
	}

	// Validate and expand file paths
	if mergeTargetType != "pgsql" {
		if mergeTargetPath == "" {
			return fmt.Errorf("--target-path is required for %s format", mergeTargetType)
		}
		mergeTargetPath = expandPath(mergeTargetPath)
	} else if mergeTargetConn == "" {

		return fmt.Errorf("--target-conn is required for pgsql format")

	}

	if mergeSourceType != "pgsql" {
		if mergeSourcePath == "" {
			return fmt.Errorf("--source-path is required for %s format", mergeSourceType)
		}
		mergeSourcePath = expandPath(mergeSourcePath)
	} else if mergeSourceConn == "" {
		return fmt.Errorf("--source-conn is required for pgsql format")
	}

	if mergeOutputType != "pgsql" {
		if mergeOutputPath == "" {
			return fmt.Errorf("--output-path is required for %s format", mergeOutputType)
		}
		mergeOutputPath = expandPath(mergeOutputPath)
	}

	// Step 1: Read target database
	fmt.Fprintf(os.Stderr, "[1/3] Reading target database...\n")
	fmt.Fprintf(os.Stderr, "      Format: %s\n", mergeTargetType)
	if mergeTargetPath != "" {
		fmt.Fprintf(os.Stderr, "      Path:   %s\n", mergeTargetPath)
	}
	if mergeTargetConn != "" {
		fmt.Fprintf(os.Stderr, "      Conn:   %s\n", maskPassword(mergeTargetConn))
	}

	targetDB, err := readDatabaseForMerge(mergeTargetType, mergeTargetPath, mergeTargetConn, "Target")
	if err != nil {
		return fmt.Errorf("failed to read target database: %w", err)
	}
	fmt.Fprintf(os.Stderr, "      ✓ Successfully read target database '%s'\n", targetDB.Name)
	printDatabaseStats(targetDB)

	// Step 2: Read source database
	fmt.Fprintf(os.Stderr, "\n[2/3] Reading source database...\n")
	fmt.Fprintf(os.Stderr, "      Format: %s\n", mergeSourceType)
	if mergeSourcePath != "" {
		fmt.Fprintf(os.Stderr, "      Path:   %s\n", mergeSourcePath)
	}
	if mergeSourceConn != "" {
		fmt.Fprintf(os.Stderr, "      Conn:   %s\n", maskPassword(mergeSourceConn))
	}

	sourceDB, err := readDatabaseForMerge(mergeSourceType, mergeSourcePath, mergeSourceConn, "Source")
	if err != nil {
		return fmt.Errorf("failed to read source database: %w", err)
	}
	fmt.Fprintf(os.Stderr, "      ✓ Successfully read source database '%s'\n", sourceDB.Name)
	printDatabaseStats(sourceDB)

	// Step 3: Merge databases
	fmt.Fprintf(os.Stderr, "\n[3/3] Merging databases...\n")

	opts := &merge.MergeOptions{
		SkipDomains:   mergeSkipDomains,
		SkipRelations: mergeSkipRelations,
		SkipEnums:     mergeSkipEnums,
		SkipViews:     mergeSkipViews,
		SkipSequences: mergeSkipSequences,
	}

	// Parse skip-tables flag
	if mergeSkipTables != "" {
		opts.SkipTableNames = parseSkipTables(mergeSkipTables)
		if len(opts.SkipTableNames) > 0 {
			fmt.Fprintf(os.Stderr, "      Skipping tables: %s\n", mergeSkipTables)
		}
	}

	result := merge.MergeDatabases(targetDB, sourceDB, opts)

	// Update timestamp
	targetDB.UpdateDate()

	// Print merge summary
	fmt.Fprintf(os.Stderr, "      ✓ Merge complete\n\n")
	fmt.Fprintf(os.Stderr, "%s\n", merge.GetMergeSummary(result))

	// Step 4: Write output
	fmt.Fprintf(os.Stderr, "\n[4/4] Writing output...\n")
	fmt.Fprintf(os.Stderr, "      Format: %s\n", mergeOutputType)
	if mergeOutputPath != "" {
		fmt.Fprintf(os.Stderr, "      Path:   %s\n", mergeOutputPath)
	}

	err = writeDatabaseForMerge(mergeOutputType, mergeOutputPath, "", targetDB, "Output")
	if err != nil {
		return fmt.Errorf("failed to write output: %w", err)
	}

	fmt.Fprintf(os.Stderr, "      ✓ Successfully written merged database\n")
	fmt.Fprintf(os.Stderr, "\n=== Merge complete ===\n")

	return nil
}

func readDatabaseForMerge(dbType, filePath, connString, label string) (*models.Database, error) {
	var reader readers.Reader

	switch strings.ToLower(dbType) {
	case "dbml":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for DBML format", label)
		}
		reader = dbml.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "dctx":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for DCTX format", label)
		}
		reader = dctx.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "drawdb":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for DrawDB format", label)
		}
		reader = drawdb.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "graphql":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for GraphQL format", label)
		}
		reader = graphql.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "json":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for JSON format", label)
		}
		reader = json.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "yaml":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for YAML format", label)
		}
		reader = yaml.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "gorm":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for GORM format", label)
		}
		reader = gorm.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "bun":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for Bun format", label)
		}
		reader = bun.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "drizzle":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for Drizzle format", label)
		}
		reader = drizzle.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "prisma":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for Prisma format", label)
		}
		reader = prisma.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "typeorm":
		if filePath == "" {
			return nil, fmt.Errorf("%s: file path is required for TypeORM format", label)
		}
		reader = typeorm.NewReader(&readers.ReaderOptions{FilePath: filePath})
	case "pgsql":
		if connString == "" {
			return nil, fmt.Errorf("%s: connection string is required for PostgreSQL format", label)
		}
		reader = pgsql.NewReader(&readers.ReaderOptions{ConnectionString: connString})
	default:
		return nil, fmt.Errorf("%s: unsupported format '%s'", label, dbType)
	}

	db, err := reader.ReadDatabase()
	if err != nil {
		return nil, err
	}

	return db, nil
}

func writeDatabaseForMerge(dbType, filePath, connString string, db *models.Database, label string) error {
	var writer writers.Writer

	switch strings.ToLower(dbType) {
	case "dbml":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for DBML format", label)
		}
		writer = wdbml.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "dctx":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for DCTX format", label)
		}
		writer = wdctx.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "drawdb":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for DrawDB format", label)
		}
		writer = wdrawdb.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "graphql":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for GraphQL format", label)
		}
		writer = wgraphql.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "json":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for JSON format", label)
		}
		writer = wjson.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "yaml":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for YAML format", label)
		}
		writer = wyaml.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "gorm":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for GORM format", label)
		}
		writer = wgorm.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "bun":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for Bun format", label)
		}
		writer = wbun.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "drizzle":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for Drizzle format", label)
		}
		writer = wdrizzle.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "prisma":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for Prisma format", label)
		}
		writer = wprisma.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "typeorm":
		if filePath == "" {
			return fmt.Errorf("%s: file path is required for TypeORM format", label)
		}
		writer = wtypeorm.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	case "pgsql":
		writer = wpgsql.NewWriter(&writers.WriterOptions{OutputPath: filePath})
	default:
		return fmt.Errorf("%s: unsupported format '%s'", label, dbType)
	}

	return writer.WriteDatabase(db)
}

func expandPath(path string) string {
	if len(path) > 0 && path[0] == '~' {
		home, err := os.UserHomeDir()
		if err == nil {
			return filepath.Join(home, path[1:])
		}
	}
	return path
}

func printDatabaseStats(db *models.Database) {
	totalTables := 0
	totalColumns := 0
	totalConstraints := 0
	totalIndexes := 0

	for _, schema := range db.Schemas {
		totalTables += len(schema.Tables)
		for _, table := range schema.Tables {
			totalColumns += len(table.Columns)
			totalConstraints += len(table.Constraints)
			totalIndexes += len(table.Indexes)
		}
	}

	fmt.Fprintf(os.Stderr, "      Schemas: %d, Tables: %d, Columns: %d, Constraints: %d, Indexes: %d\n",
		len(db.Schemas), totalTables, totalColumns, totalConstraints, totalIndexes)
}

func parseSkipTables(skipTablesStr string) map[string]bool {
	skipTables := make(map[string]bool)
	if skipTablesStr == "" {
		return skipTables
	}

	// Split by comma and trim whitespace
	parts := strings.Split(skipTablesStr, ",")
	for _, part := range parts {
		trimmed := strings.TrimSpace(part)
		if trimmed != "" {
			// Store in lowercase for case-insensitive matching
			skipTables[strings.ToLower(trimmed)] = true
		}
	}

	return skipTables
}