sow-tools/internal/pipeline/extract.go

package pipeline

import (
	"bytes"
	"encoding/json"
	"errors"
	"fmt"
	"os"
	"path/filepath"
	"slices"
	"strings"

	"gitea.westgate.pw/ShadowsOverWestgate/sow-tools/internal/erf"
	"gitea.westgate.pw/ShadowsOverWestgate/sow-tools/internal/gff"
	"gitea.westgate.pw/ShadowsOverWestgate/sow-tools/internal/project"
)

type ExtractResult struct {
	ModulePath  string
	HAKPaths    []string
	Written     int
	Overwritten int
	Removed     int
	Skipped     int
}

func Extract(p *project.Project, files ...string) (ExtractResult, error) {
	var result ExtractResult
	var failures []error
	desired := map[string]struct{}{}

	allowed := make(map[string]bool)
	for _, f := range files {
		allowed[f] = true
	}

	shouldExtractMod := len(allowed) == 0 || allowed[p.Config.Module.ResRef+".mod"]
	if shouldExtractMod {
		modulePath := filepath.Join(p.BuildDir(), p.Config.Module.ResRef+".mod")
		input, err := os.Open(modulePath)
		if err != nil {
			return ExtractResult{}, fmt.Errorf("open module archive: %w", err)
		}
		defer input.Close()

		archive, err := erf.Read(input)
		if err != nil {
			return ExtractResult{}, fmt.Errorf("read module archive: %w", err)
		}

		result.ModulePath = modulePath
		written, overwritten, skipped, errs := extractArchiveResources(p, archive, desired)
		result.Written += written
		result.Overwritten += overwritten
		result.Skipped += skipped
		failures = append(failures, errs...)
	}

	hakPaths, err := filepath.Glob(filepath.Join(p.BuildDir(), "*.hak"))
	if err != nil {
		return result, fmt.Errorf("scan hak archives: %w", err)
	}
	slices.Sort(hakPaths)
	for _, hakPath := range hakPaths {
		filename := filepath.Base(hakPath)
		if len(allowed) > 0 && !allowed[filename] {
			continue
		}

		input, err := os.Open(hakPath)
		if err != nil {
			failures = append(failures, fmt.Errorf("open hak archive %s: %w", hakPath, err))
			continue
		}
		hakArchive, err := erf.Read(input)
		input.Close()
		if err != nil {
			failures = append(failures, fmt.Errorf("read hak archive %s: %w", hakPath, err))
			continue
		}

		result.HAKPaths = append(result.HAKPaths, hakPath)
		written, overwritten, skipped, errs := extractArchiveResources(p, hakArchive, desired)
		result.Written += written
		result.Overwritten += overwritten
		result.Skipped += skipped
		failures = append(failures, errs...)
	}

	removed, errs := cleanupStaleFiles(p, desired)
	result.Removed = removed
	failures = append(failures, errs...)

	if len(failures) > 0 {
		return result, errors.Join(failures...)
	}
	return result, nil
}

func extractArchiveResources(p *project.Project, archive erf.Archive, desired map[string]struct{}) (int, int, int, []error) {
	var failures []error
	writtenCount := 0
	overwrittenCount := 0
	skippedCount := 0

	ignored := make(map[string]bool)
	for _, ext := range p.Config.Extract.IgnoreExtensions {
		ignored[ext] = true
	}

	for _, resource := range archive.Resources {
		ext, ok := erf.ExtensionForResourceType(resource.Type)
		if !ok {
			failures = append(failures, fmt.Errorf("unsupported resource type 0x%04X for %s", resource.Type, resource.Name))
			continue
		}
		if ignored[ext] {
			skippedCount++
			continue
		}

		target, data, err := extractedFile(p, resource, ext)
		if err != nil {
			failures = append(failures, err)
			continue
		}
		desired[target] = struct{}{}

		state, err := writeManagedFile(target, data)
		if err != nil {
			failures = append(failures, err)
			continue
		}
		switch state {
		case writeNew:
			writtenCount++
		case writeOverwritten:
			overwrittenCount++
		case writeSkipped:
			skippedCount++
		}
	}

	return writtenCount, overwrittenCount, skippedCount, failures
}

func extractedFile(p *project.Project, resource erf.Resource, extension string) (string, []byte, error) {
	resref := strings.ToLower(resource.Name)

	switch extension {
	case "nss":
		return filepath.Join(p.SourceDir(), "scripts", resref+".nss"), resource.Data, nil
	case "utc", "utd", "ute", "uti", "utm", "utp", "uts", "utt", "utw",
		"are", "dlg", "fac", "gic", "git", "ifo", "itp", "jrl":
		document, err := gff.Read(bytes.NewReader(resource.Data))
		if err != nil {
			return "", nil, fmt.Errorf("decode gff %s.%s: %w", resource.Name, extension, err)
		}
		formatted, err := json.MarshalIndent(document, "", "  ")
		if err != nil {
			return "", nil, fmt.Errorf("marshal json %s.%s: %w", resource.Name, extension, err)
		}
		formatted = append(formatted, '\n')
		return filepath.Join(p.SourceDir(), sourceSubdir(extension), resref+"."+extension+".json"), formatted, nil
	default:
		return filepath.Join(p.AssetsDir(), extension, resref+"."+extension), resource.Data, nil
	}
}

type writeState int

const (
	writeSkipped writeState = iota
	writeNew
	writeOverwritten
)

func writeManagedFile(path string, data []byte) (writeState, error) {
	if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
		return writeSkipped, fmt.Errorf("create parent directory for %s: %w", path, err)
	}

	existing, err := os.ReadFile(path)
	if err == nil {
		if bytes.Equal(existing, data) {
			return writeSkipped, nil
		}
		if err := os.WriteFile(path, data, 0o644); err != nil {
			return writeSkipped, fmt.Errorf("overwrite %s: %w", path, err)
		}
		return writeOverwritten, nil
	}
	if !errors.Is(err, os.ErrNotExist) {
		return writeSkipped, fmt.Errorf("check existing file %s: %w", path, err)
	}

	if err := os.WriteFile(path, data, 0o644); err != nil {
		return writeSkipped, fmt.Errorf("write %s: %w", path, err)
	}
	return writeNew, nil
}

func cleanupStaleFiles(p *project.Project, desired map[string]struct{}) (int, []error) {
	candidates := make([]string, 0, len(p.Inventory.SourceFiles)+len(p.Inventory.ScriptFiles)+len(p.Inventory.AssetFiles))
	for _, rel := range p.Inventory.SourceFiles {
		candidates = append(candidates, filepath.Join(p.SourceDir(), filepath.FromSlash(rel)))
	}
	for _, rel := range p.Inventory.ScriptFiles {
		candidates = append(candidates, filepath.Join(p.SourceDir(), filepath.FromSlash(rel)))
	}
	for _, rel := range p.Inventory.AssetFiles {
		candidates = append(candidates, filepath.Join(p.AssetsDir(), filepath.FromSlash(rel)))
	}

	removed := 0
	var failures []error
	for _, path := range candidates {
		if _, keep := desired[path]; keep {
			continue
		}
		if err := os.Remove(path); err != nil {
			if errors.Is(err, os.ErrNotExist) {
				continue
			}
			failures = append(failures, fmt.Errorf("remove stale file %s: %w", path, err))
			continue
		}
		removed++
		cleanupEmptyParents(filepath.Dir(path), p.SourceDir(), p.AssetsDir())
	}
	return removed, failures
}

func cleanupEmptyParents(dir string, roots ...string) {
	for {
		if dir == "." || dir == string(filepath.Separator) {
			return
		}
		stop := false
		for _, root := range roots {
			if dir == root {
				stop = true
				break
			}
		}
		if stop {
			return
		}
		if err := os.Remove(dir); err != nil {
			return
		}
		dir = filepath.Dir(dir)
	}
}

func sourceSubdir(extension string) string {
	switch strings.ToLower(extension) {
	case "are":
		return "areas"
	case "dlg":
		return "dialogs"
	case "fac":
		return "factions"
	case "gic":
		return "instance"
	case "git":
		return "instance"
	case "ifo":
		return "module"
	case "itp":
		return "palettes"
	case "jrl":
		return "journal"
	case "utc":
		return filepath.Join("blueprints", "creatures")
	case "utd":
		return filepath.Join("blueprints", "doors")
	case "ute":
		return filepath.Join("blueprints", "encounters")
	case "uti":
		return filepath.Join("blueprints", "items")
	case "utm":
		return filepath.Join("blueprints", "merchants")
	case "utp":
		return filepath.Join("blueprints", "placeables")
	case "uts":
		return filepath.Join("blueprints", "sounds")
	case "utt":
		return filepath.Join("blueprints", "triggers")
	case "utw":
		return filepath.Join("blueprints", "waypoints")
	default:
		return extension
	}
}