agent-lib/internal/vendor/update.go

package vendor

import (
	"bytes"
	"crypto/sha256"
	"fmt"
	"io"
	"os"
	"path/filepath"
	"sort"
	"strings"

	"github.com/m3tam3re/agent-lib/internal/discovery"
	"github.com/m3tam3re/agent-lib/internal/gitsource"
	"github.com/m3tam3re/agent-lib/internal/lockfile"
)

// ChangeReport describes one update (or diff preview) for one source.
type ChangeReport struct {
	Source    string              `json:"source"`
	FromRev   string              `json:"from_rev,omitempty"`
	ToRev     string              `json:"to_rev,omitempty"`
	UpToDate  bool                `json:"up_to_date,omitempty"`
	Added     map[string][]string `json:"added,omitempty"`
	Removed   map[string][]string `json:"removed,omitempty"`
	Changed   map[string][]string `json:"changed,omitempty"`
	Available map[string][]string `json:"available,omitempty"`
}

// Update refreshes one source to the latest revision of its pinned ref,
// rewrites the external area deterministically and reports the changes.
func Update(workDir, name string, stdout io.Writer) (*ChangeReport, error) {
	lf, err := lockfile.Load(filepath.Join(workDir, lockfile.FileName))
	if err != nil {
		return nil, err
	}
	src, ok := lf.Sources[name]
	if !ok {
		return nil, fmt.Errorf("source %q is not in the lockfile", name)
	}

	plan, err := computeUpdate(workDir, lf, name, true)
	if err != nil {
		return nil, err
	}
	if plan.Report.UpToDate {
		fmt.Fprintf(stdout, "source %s already up to date (rev %s)\n", name, src.Rev)
		return plan.Report, nil
	}

	cleanupStaleStaging(workDir)
	staging := filepath.Join(workDir, externalDir, ".staging-"+name)
	if err := os.RemoveAll(staging); err != nil {
		return nil, err
	}
	if err := materializeTree(plan.Tree, plan.Selected, src.Renames, staging); err != nil {
		os.RemoveAll(staging)
		return nil, err
	}
	if err := fillChanged(plan.Report, name, lf, staging); err != nil {
		os.RemoveAll(staging)
		return nil, err
	}

	oldDir := filepath.Join(workDir, externalDir, name)
	if err := os.RemoveAll(oldDir); err != nil {
		return nil, fmt.Errorf("replacing %s: %w", oldDir, err)
	}
	if err := os.Rename(staging, oldDir); err != nil {
		return nil, fmt.Errorf("swapping in new external area: %w", err)
	}

	src.Rev = plan.Repo.Rev
	src.Inventory = inventoryOf(plan.Selected)
	src.Warnings = dedupe(warningsOf(plan.Items))
	if err := lf.Save(filepath.Join(workDir, lockfile.FileName)); err != nil {
		return nil, err
	}

	printChangeReport(stdout, plan.Report)
	return plan.Report, nil
}

// UpdateAll refreshes every source; a failing source does not block the
// others, but the run reports every failure and exits non-zero.
func UpdateAll(workDir string, stdout io.Writer) error {
	lf, err := lockfile.Load(filepath.Join(workDir, lockfile.FileName))
	if err != nil {
		return err
	}
	if len(lf.Sources) == 0 {
		fmt.Fprintln(stdout, "no sources to update")
		return nil
	}
	var failures []string
	for _, name := range sortedSourceNames(lf) {
		if _, err := Update(workDir, name, stdout); err != nil {
			fmt.Fprintf(stdout, "source %s: FAILED: %v\n", name, err)
			failures = append(failures, name)
		}
	}
	if len(failures) > 0 {
		return fmt.Errorf("update failed for: %s", strings.Join(failures, ", "))
	}
	return nil
}

// Diff previews an update without touching the work repository, surfacing
// new upstream items that the current selection would not bring in.
func Diff(workDir, name string, stdout io.Writer) (*ChangeReport, error) {
	lf, err := lockfile.Load(filepath.Join(workDir, lockfile.FileName))
	if err != nil {
		return nil, err
	}
	src, ok := lf.Sources[name]
	if !ok {
		return nil, fmt.Errorf("source %q is not in the lockfile", name)
	}

	plan, err := computeUpdate(workDir, lf, name, false)
	if err != nil {
		return nil, err
	}
	if plan.Report.UpToDate {
		fmt.Fprintf(stdout, "source %s up to date (rev %s); no changes upstream\n", name, src.Rev)
		return plan.Report, nil
	}

	tmp, err := os.MkdirTemp("", "agent-lib-diff-")
	if err != nil {
		return nil, err
	}
	defer os.RemoveAll(tmp)
	staged := filepath.Join(tmp, "staged")
	if err := materializeTree(plan.Tree, plan.Selected, src.Renames, staged); err != nil {
		return nil, err
	}
	if err := fillChanged(plan.Report, name, lf, staged); err != nil {
		return nil, err
	}

	fmt.Fprintf(stdout, "source %s: pending changes vs upstream %s…\n", name, short(plan.Report.ToRev))
	printChangeReport(stdout, plan.Report)
	return plan.Report, nil
}

// Remove deletes a source's external area and lockfile entry.
func Remove(workDir, name string, stdout io.Writer) error {
	lockPath := filepath.Join(workDir, lockfile.FileName)
	lf, err := lockfile.Load(lockPath)
	if err != nil {
		return err
	}
	if _, ok := lf.Sources[name]; !ok {
		return fmt.Errorf("source %q is not in the lockfile", name)
	}
	if err := os.RemoveAll(filepath.Join(workDir, externalDir, name)); err != nil {
		return fmt.Errorf("removing external area: %w", err)
	}
	delete(lf.Sources, name)
	if err := lf.Save(lockPath); err != nil {
		return err
	}
	externalRoot := filepath.Join(workDir, externalDir)
	if entries, err := os.ReadDir(externalRoot); err == nil && len(entries) == 0 {
		os.Remove(externalRoot)
	}
	fmt.Fprintf(stdout, "source %s removed (external area and lockfile entry)\n", name)
	return nil
}

type updatePlan struct {
	Report   *ChangeReport
	Selected []discovery.Item
	Items    []discovery.Item
	Tree     discovery.Tree
	Repo     *gitsource.Repo
}

// computeUpdate clones the pinned ref, rediscovers and reselects. With
// withCollisions it also runs the namespace checks that must gate a real
// update; diff skips them to stay a pure preview.
func computeUpdate(workDir string, lf *lockfile.Lockfile, name string, withCollisions bool) (*updatePlan, error) {
	src := lf.Sources[name]
	repo, err := gitsource.Clone(src.URL, src.Ref)
	if err != nil {
		return nil, err
	}
	tree, err := repo.Tree()
	if err != nil {
		return nil, err
	}
	items, err := discovery.Scan(tree, discovery.FromLockfileDiscovery(derefDiscovery(src)))
	if err != nil {
		return nil, err
	}

	if src.Selection.Mode == lockfile.ModeInclude {
		if err := checkIncludeEntries(items, src.Selection.Include, name); err != nil {
			return nil, err
		}
	}
	selected := applySelection(items, src.Selection)
	if withCollisions {
		if err := checkCollisions(workDir, lf, name, selected, src.Renames); err != nil {
			return nil, err
		}
	}

	report := &ChangeReport{
		Source:   name,
		FromRev:  src.Rev,
		ToRev:    repo.Rev,
		UpToDate: src.Rev == repo.Rev,
	}
	report.Added, report.Removed = membershipDiff(inventorySet(src.Inventory), idSet(selected))
	report.Available = availableSince(inventorySet(src.Inventory), items, selected)
	return &updatePlan{Report: report, Selected: selected, Items: items, Tree: tree, Repo: repo}, nil
}

// availableSince lists upstream items that appeared since the pinned
// inventory and that the current selection would not bring in (relevant in
// include mode; empty in all mode by construction).
func availableSince(old map[string]map[string]bool, items []discovery.Item, selected []discovery.Item) map[string][]string {
	inSelection := idSet(selected)
	out := map[string][]string{}
	for _, it := range items {
		if old[it.Type][it.UpstreamID] || inSelection[it.Type][it.UpstreamID] {
			continue
		}
		out[it.Type] = append(out[it.Type], it.UpstreamID)
	}
	for _, ids := range out {
		sort.Strings(ids)
	}
	return out
}

func cleanupStaleStaging(workDir string) {
	entries, err := os.ReadDir(filepath.Join(workDir, externalDir))
	if err != nil {
		return
	}
	for _, e := range entries {
		if strings.HasPrefix(e.Name(), ".staging-") {
			os.RemoveAll(filepath.Join(workDir, externalDir, e.Name()))
		}
	}
}

func materializeTree(tree discovery.Tree, selected []discovery.Item, renames map[string]string, dst string) error {
	for _, it := range selected {
		dep := deployedName(it.UpstreamID, renames)
		base := filepath.Join(dst, it.Type)
		var target string
		var err error
		switch it.Type {
		case lockfile.TypeSkill:
			target = filepath.Join(base, dep)
			err = copyTreeDir(tree, it.RelPath, target)
		default:
			target = filepath.Join(base, dep+filepath.Ext(it.RelPath))
			err = copyTreeFile(tree, it.RelPath, target)
		}
		if err != nil {
			return fmt.Errorf("materializing %s/%s: %w", it.Type, it.UpstreamID, err)
		}
	}
	return nil
}

// fillChanged byte-compares the staged tree against the previous external
// area to populate the changed-section of the report.
func fillChanged(report *ChangeReport, name string, lf *lockfile.Lockfile, staging string) error {
	report.Changed = map[string][]string{}
	src := lf.Sources[name]
	for _, typ := range lockfile.Types {
		for _, id := range src.Inventory[typ] {
			if report.itemIn(typ, id, report.Added) || report.itemIn(typ, id, report.Removed) {
				continue
			}
			dep := deployedName(id, src.Renames)
			ext := extFor(typ, staging, dep)
			oldDir := filepath.Join(externalDir, name, typ)
			newDir := filepath.Join(staging, typ)
			same, err := treesEqual(filepath.Join(oldDir, dep+ext), filepath.Join(newDir, dep+ext))
			if typ == lockfile.TypeSkill {
				same, err = treesEqual(filepath.Join(oldDir, dep), filepath.Join(newDir, dep))
			}
			if err != nil {
				return fmt.Errorf("comparing %s/%s: %w", typ, id, err)
			}
			if !same {
				report.Changed[typ] = append(report.Changed[typ], id)
			}
		}
	}
	return nil
}

func extFor(typ, staging, dep string) string {
	switch typ {
	case lockfile.TypeSkill:
		return ""
	case lockfile.TypeMcp:
		if _, err := os.Stat(filepath.Join(staging, typ, dep+".yaml")); err == nil {
			return ".yaml"
		}
		return ".yml"
	default:
		return ".md"
	}
}

func (r *ChangeReport) itemIn(typ, id string, m map[string][]string) bool {
	for _, x := range m[typ] {
		if x == id {
			return true
		}
	}
	return false
}

func membershipDiff(oldSets, newSets map[string]map[string]bool) (added, removed map[string][]string) {
	added = map[string][]string{}
	removed = map[string][]string{}
	for _, typ := range lockfile.Types {
		for id := range newSets[typ] {
			if !oldSets[typ][id] {
				added[typ] = append(added[typ], id)
			}
		}
		for id := range oldSets[typ] {
			if !newSets[typ][id] {
				removed[typ] = append(removed[typ], id)
			}
		}
		sort.Strings(added[typ])
		sort.Strings(removed[typ])
	}
	return added, removed
}

func inventorySet(inv map[string][]string) map[string]map[string]bool {
	out := map[string]map[string]bool{}
	for typ, ids := range inv {
		if out[typ] == nil {
			out[typ] = map[string]bool{}
		}
		for _, id := range ids {
			out[typ][id] = true
		}
	}
	return out
}

func idSet(items []discovery.Item) map[string]map[string]bool {
	out := map[string]map[string]bool{}
	for _, it := range items {
		if out[it.Type] == nil {
			out[it.Type] = map[string]bool{}
		}
		out[it.Type][it.UpstreamID] = true
	}
	return out
}

func treesEqual(a, b string) (bool, error) {
	filesA, errA := listFilesRecursive(a)
	filesB, errB := listFilesRecursive(b)
	if errA != nil || errB != nil {
		if os.IsNotExist(errA) || os.IsNotExist(errB) {
			return false, nil
		}
		return false, fmt.Errorf("walking %s or %s: %w/%w", a, b, errA, errB)
	}
	if len(filesA) != len(filesB) {
		return false, nil
	}
	for i := range filesA {
		if filesA[i] != filesB[i] {
			return false, nil
		}
		ha, err := hashFile(filepath.Join(a, filesA[i]))
		if err != nil {
			return false, err
		}
		hb, err := hashFile(filepath.Join(b, filesB[i]))
		if err != nil {
			return false, err
		}
		if !bytes.Equal(ha, hb) {
			return false, nil
		}
	}
	return true, nil
}

func listFilesRecursive(root string) ([]string, error) {
	var out []string
	err := filepath.WalkDir(root, func(p string, d os.DirEntry, err error) error {
		if err != nil {
			return err
		}
		if d.IsDir() {
			return nil
		}
		rel, err := filepath.Rel(root, p)
		if err != nil {
			return err
		}
		out = append(out, filepath.ToSlash(rel))
		return nil
	})
	if err != nil {
		return nil, err
	}
	sort.Strings(out)
	return out, nil
}

func hashFile(path string) ([]byte, error) {
	f, err := os.Open(path)
	if err != nil {
		return nil, err
	}
	defer f.Close()
	h := sha256.New()
	if _, err := io.Copy(h, f); err != nil {
		return nil, err
	}
	return h.Sum(nil), nil
}

func derefDiscovery(src *lockfile.Source) lockfile.Discovery {
	if src.Discovery == nil {
		return lockfile.Discovery{}
	}
	return *src.Discovery
}

func inventoryOf(selected []discovery.Item) map[string][]string {
	out := map[string][]string{}
	for _, it := range selected {
		out[it.Type] = append(out[it.Type], it.UpstreamID)
	}
	for _, ids := range out {
		sort.Strings(ids)
	}
	return out
}

func warningsOf(items []discovery.Item) []string {
	var out []string
	for _, it := range items {
		out = append(out, it.Warnings...)
	}
	return out
}

func printChangeReport(w io.Writer, r *ChangeReport) {
	empty := true
	for _, typ := range lockfile.Types {
		added := r.Added[typ]
		removed := r.Removed[typ]
		changed := r.Changed[typ]
		available := r.Available[typ]
		if len(added)+len(removed)+len(changed)+len(available) == 0 {
			continue
		}
		if empty {
			if r.FromRev != "" && r.ToRev != "" && r.FromRev != r.ToRev {
				fmt.Fprintf(w, "source %s: %s… → %s…\n", r.Source, short(r.FromRev), short(r.ToRev))
			} else {
				fmt.Fprintf(w, "source %s:\n", r.Source)
			}
			empty = false
		}
		fmt.Fprintf(w, "  %s:\n", typ)
		for _, id := range added {
			fmt.Fprintf(w, "    + %s (added)\n", id)
		}
		for _, id := range changed {
			fmt.Fprintf(w, "    ~ %s (changed)\n", id)
		}
		for _, id := range removed {
			fmt.Fprintf(w, "    - %s (removed)\n", id)
		}
		for _, id := range available {
			fmt.Fprintf(w, "    ? %s (available upstream, not selected)\n", id)
		}
	}
	if empty {
		fmt.Fprintf(w, "source %s: no item changes\n", r.Source)
	}
}

func short(rev string) string {
	if len(rev) > 8 {
		return rev[:8]
	}
	return rev
}