eec-campaigns/internal/campaign/sync.go

package campaign

import (
	"bytes"
	"crypto/sha256"
	"encoding/hex"
	"fmt"
	"os"
	"os/exec"
	"path/filepath"
	"sort"
	"strings"

	"gitea.reground.org/will/eec-campaigns/internal/listmonk"
)

// SyncResult summarizes what a sync run did, split so the CI log (and the
// person reading it) can tell at a glance what actually happened — no bad
// campaign silently skipped or overwritten, no preview silently swallowed.
type SyncResult struct {
	Synced        []string // created or updated (content changed) in listmonk
	Unchanged     []string // existing draft, content identical — no API write, no preview
	Skipped       []string // already sent/running/paused/etc. in listmonk — a historical record, left untouched, not an error
	Rejected      []string // validation/segmentation/list-lookup failure — nothing written
	PreviewFailed []string // content synced fine, but the automatic preview send itself failed
	// Mode is purely for CI log visibility into how this run was scoped:
	// "full scan", or "since <sha>: N of M campaign(s) touched" — see
	// SyncChanged.
	Mode string
}

// SyncDir walks <root>/campaigns/*/campaign.md and upserts each as a
// listmonk draft campaign. Equivalent to SyncChanged with an empty since —
// see SyncChanged for the git-diff-scoped version this wraps.
func SyncDir(lm *listmonk.Client, root string, defaultPreviewEmails []string) (*SyncResult, error) {
	return SyncChanged(lm, root, defaultPreviewEmails, "")
}

// SyncChanged is SyncDir optionally scoped to just the campaigns whose
// files changed since a given commit (via ChangedSlugs' git diff), so a
// push that only touched one campaign doesn't need to fetch and diff every
// other campaign in the repo against listmonk to confirm nothing changed.
// An empty since, or ChangedSlugs failing to resolve it for any reason
// (e.g. a shallow clone missing that commit's history), falls back to a
// full scan exactly like SyncDir — this can never cause a campaign to be
// silently skipped, only ever process more than strictly necessary.
func SyncChanged(lm *listmonk.Client, root string, defaultPreviewEmails []string, since string) (*SyncResult, error) {
	pattern := filepath.Join(root, "campaigns", "*", "campaign.md")
	matches, err := filepath.Glob(pattern)
	if err != nil {
		return nil, err
	}

	mode := "full scan"
	if since != "" {
		if slugs, ok := ChangedSlugs(root, since); ok {
			total := len(matches)
			filtered := matches[:0]
			for _, path := range matches {
				if slugs[filepath.Base(filepath.Dir(path))] {
					filtered = append(filtered, path)
				}
			}
			matches = filtered
			mode = fmt.Sprintf("since %s: %d of %d campaign(s) touched", since, len(matches), total)
		} else {
			mode = fmt.Sprintf("full scan (could not resolve since %s)", since)
		}
	}

	result := &SyncResult{Mode: mode}
	for _, path := range matches {
		dir := filepath.Dir(path)
		slug := filepath.Base(dir)

		res, err := syncOne(lm, slug, dir, path, defaultPreviewEmails)
		if err != nil {
			result.Rejected = append(result.Rejected, fmt.Sprintf("%s: %v", slug, err))
			continue
		}
		if res.Skipped != "" {
			result.Skipped = append(result.Skipped, fmt.Sprintf("%s: %s", slug, res.Skipped))
			continue
		}
		if !res.Changed {
			result.Unchanged = append(result.Unchanged, slug)
			continue
		}
		result.Synced = append(result.Synced, slug)
		if res.PreviewErr != nil {
			result.PreviewFailed = append(result.PreviewFailed, fmt.Sprintf("%s: %v", slug, res.PreviewErr))
		}
	}
	return result, nil
}

// ChangedSlugs returns the set of campaign slugs (directory names under
// campaigns/) whose files differ between since and HEAD, by shelling out to
// git diff. The second return is false when since is empty, the all-zero
// placeholder SHA (what gitea's push event sends for a brand-new branch's
// first push, since there's no real "before" commit), or the diff itself
// fails for any reason — most commonly a shallow clone that doesn't have
// since's history available locally. All of those mean "can't determine
// what changed," never "nothing changed," so callers must treat false as
// "fall back to processing everything," not as an empty result.
func ChangedSlugs(root, since string) (map[string]bool, bool) {
	if since == "" || since == strings.Repeat("0", 40) {
		return nil, false
	}
	cmd := exec.Command("git", "diff", "--name-only", since, "HEAD", "--", "campaigns")
	cmd.Dir = root
	out, err := cmd.Output()
	if err != nil {
		return nil, false
	}
	slugs := map[string]bool{}
	for _, line := range strings.Split(strings.TrimSpace(string(out)), "\n") {
		if line == "" {
			continue
		}
		// Lines look like "campaigns/<slug>/campaign.md" or
		// "campaigns/<slug>/assets/whatever.pdf" — the pathspec above
		// already guarantees the "campaigns/" prefix.
		parts := strings.SplitN(line, "/", 3)
		if len(parts) >= 2 {
			slugs[parts[1]] = true
		}
	}
	return slugs, true
}

type syncOneResult struct {
	Changed    bool
	Skipped    string // non-empty (the listmonk status) when left untouched because it's no longer a draft
	PreviewErr error  // set only when Changed and the automatic preview send failed
}

func syncOne(lm *listmonk.Client, slug, dir, path string, defaultPreviewEmails []string) (syncOneResult, error) {
	fm, body, err := ParseFile(path)
	if err != nil {
		return syncOneResult{}, err
	}

	existing, err := lm.FindCampaignByName(slug)
	if err != nil {
		return syncOneResult{}, err
	}
	if existing != nil && existing.Status != "draft" {
		// Once a campaign has actually gone out (or is running/paused/etc.),
		// its campaign.md is a historical record of what was sent and when —
		// not something sync should touch or fail the build over. Forcing it
		// to be deleted just to keep CI green would throw away exactly the
		// record someone would want to look back at later. Check this before
		// resolving anything else below, so a since-renamed list or a
		// since-changed attachment on an old campaign can never break the
		// build either — an archived campaign has nothing left to resolve.
		return syncOneResult{Skipped: existing.Status}, nil
	}

	listIDs, err := resolveLists(lm, fm.Lists)
	if err != nil {
		return syncOneResult{}, err
	}
	if fm.SegmentQuery != "" {
		segID, err := resolveSegment(lm, slug, fm.SegmentQuery)
		if err != nil {
			return syncOneResult{}, err
		}
		listIDs = append(listIDs, segID)
	}

	mediaIDs, err := resolveAttachments(lm, slug, dir, fm.Attachments)
	if err != nil {
		return syncOneResult{}, err
	}

	input := listmonk.CampaignInput{
		Name:       slug, // the directory name IS the identity — see internal/campaign/parse.go doc
		Subject:    fm.Subject,
		FromEmail:  fm.FromEmail,
		TemplateID: fm.TemplateID,
		Type:       fm.Type,
		Tags:       fm.Tags,
		ListIDs:    listIDs,
		Body:       body,
		MediaIDs:   mediaIDs,
	}

	var campaignID int
	var changed bool
	switch {
	case existing == nil:
		created, err := lm.CreateCampaign(input)
		if err != nil {
			return syncOneResult{}, err
		}
		campaignID, changed = created.ID, true

	case campaignUnchanged(existing, input):
		campaignID, changed = existing.ID, false

	default:
		if _, err := lm.UpdateCampaign(existing.ID, input); err != nil {
			return syncOneResult{}, err
		}
		campaignID, changed = existing.ID, true
	}

	if !changed {
		return syncOneResult{Changed: false}, nil
	}

	recipients := fm.PreviewEmails
	if len(recipients) == 0 {
		recipients = defaultPreviewEmails
	}
	var previewErr error
	if len(recipients) > 0 {
		previewErr = lm.TestCampaign(campaignID, input, recipients)
	}
	return syncOneResult{Changed: true, PreviewErr: previewErr}, nil
}

func resolveLists(lm *listmonk.Client, names []string) ([]int, error) {
	ids := make([]int, 0, len(names))
	for _, name := range names {
		id, err := lm.FindListByName(name)
		if err != nil {
			return nil, fmt.Errorf("list %q: %w", name, err)
		}
		ids = append(ids, id)
	}
	return ids, nil
}

// resolveSegment materializes a segment_query into list membership, since a
// listmonk campaign can only target whole list(s), not an arbitrary query
// directly (still an open, unmerged upstream feature request — see the
// plan's Open items). listmonk applies the query server-side in one call
// (QueryAddToList) — no subscriber IDs are ever fetched client-side. The
// list is a point-in-time snapshot: re-running sync re-evaluates the query
// and re-syncs membership, so a later push before send picks up
// newly-matching subscribers.
func resolveSegment(lm *listmonk.Client, slug, query string) (int, error) {
	listName := "segment:" + slug
	listID, err := lm.FindOrCreateListByName(listName)
	if err != nil {
		return 0, fmt.Errorf("segment list %q: %w", listName, err)
	}
	if err := lm.QueryAddToList(query, listID); err != nil {
		return 0, fmt.Errorf("segment_query: %w", err)
	}
	return listID, nil
}

// resolveAttachments uploads each attachment under a content-hash-synthesized
// filename, so editing a file naturally produces a fresh upload while an
// unchanged file is recognized and reused — no local manifest of what's
// already been uploaded needed, listmonk's own media library is the only
// state store.
func resolveAttachments(lm *listmonk.Client, slug, dir string, relPaths []string) ([]int, error) {
	ids := make([]int, 0, len(relPaths))
	for _, rel := range relPaths {
		full := filepath.Join(dir, rel)
		content, err := os.ReadFile(full)
		if err != nil {
			return nil, fmt.Errorf("attachment %s: %w", rel, err)
		}
		name := attachmentFilename(slug, rel, content)

		id, found, err := lm.FindMediaByFilename(name)
		if err != nil {
			return nil, fmt.Errorf("attachment %s: %w", rel, err)
		}
		if !found {
			id, err = lm.UploadMedia(name, bytes.NewReader(content))
			if err != nil {
				return nil, fmt.Errorf("attachment %s: %w", rel, err)
			}
		}
		ids = append(ids, id)
	}
	return ids, nil
}

func attachmentFilename(slug, relPath string, content []byte) string {
	sum := sha256.Sum256(content)
	prefix := hex.EncodeToString(sum[:])[:8]
	return fmt.Sprintf("%s-%s-%s", slug, prefix, filepath.Base(relPath))
}

// campaignUnchanged reports whether input's content already matches what's
// stored in listmonk, so an unchanged campaign costs zero API writes and
// never re-fires a preview. List/tag/media comparisons are order-independent
// since neither side's ordering is meaningful.
func campaignUnchanged(existing *listmonk.Campaign, input listmonk.CampaignInput) bool {
	return existing.Subject == input.Subject &&
		existing.Body == input.Body &&
		// altbody always mirrors body (see CampaignInput.payload) -- comparing
		// it here is what catches a campaign synced before that started, whose
		// stored altbody is still null even though its body hasn't changed.
		existing.AltBody == input.Body &&
		existing.FromEmail == input.FromEmail &&
		// input.TemplateID == 0 means "unspecified, use whatever's assigned"
		// (see CampaignInput.TemplateID) -- listmonk always assigns some real,
		// nonzero template_id server-side even when it's omitted from the
		// create/update payload, so comparing 0 against that would flag every
		// campaign.md without an explicit template_id as changed on every
		// sync, forever.
		(input.TemplateID == 0 || existing.TemplateID == input.TemplateID) &&
		existing.Type == input.Type &&
		equalStringSets(existing.Tags, input.Tags) &&
		equalIntSets(existing.ListIDs, input.ListIDs) &&
		equalIntSets(existing.MediaIDs, input.MediaIDs)
}

func equalStringSets(a, b []string) bool {
	if len(a) != len(b) {
		return false
	}
	as, bs := append([]string(nil), a...), append([]string(nil), b...)
	sort.Strings(as)
	sort.Strings(bs)
	for i := range as {
		if as[i] != bs[i] {
			return false
		}
	}
	return true
}

func equalIntSets(a, b []int) bool {
	if len(a) != len(b) {
		return false
	}
	as, bs := append([]int(nil), a...), append([]int(nil), b...)
	sort.Ints(as)
	sort.Ints(bs)
	for i := range as {
		if as[i] != bs[i] {
			return false
		}
	}
	return true
}