ZDDC/zddc/internal/archive/index.go

package archive

import (
	"log/slog"
	"os"
	"path/filepath"
	"regexp"
	"sort"
	"strings"
	"sync"
)

// RevisionEntry holds the resolved file paths for one base revision.
type RevisionEntry struct {
	BasePath  string            // server-relative path for trackingNumber_rev.html
	Modifiers map[string]string // modifier key (e.g. "C1") → server-relative path
	Date      string            // transmittal date (YYYY-MM-DD) for first-seen logic
}

// TrackingEntry holds all revision data for one tracking number.
type TrackingEntry struct {
	HighestBaseRev string                    // highest base revision (for trackingNumber.html)
	ByRevision     map[string]*RevisionEntry // base revision → entry
}

// ProjectEntry buckets all tracking numbers under one top-level segment of
// fsRoot (the "project"). Each project is its own namespace — the same
// tracking number issued under two different projects does NOT collide; each
// project's .archive/ surfaces only its own.
type ProjectEntry struct {
	ByTracking map[string]*TrackingEntry
}

// Index is the in-memory archive index, bucketed by project. The project key
// is the first slash-separated segment of an indexed file's server-relative
// path. .archive virtual requests under /<project>/.../.archive/ resolve
// against the named project's bucket; /.archive/ at the very root has no
// project and returns 404.
type Index struct {
	mu        sync.RWMutex
	ByProject map[string]*ProjectEntry
}

// NewIndex returns an empty Index.
func NewIndex() *Index {
	return &Index{
		ByProject: make(map[string]*ProjectEntry),
	}
}

// transmittalFolderRE matches: YYYY-MM-DD_anything (anything) - anything
var transmittalFolderRE = regexp.MustCompile(`^(\d{4}-\d{2}-\d{2})_[^_\s]+\s*\([^)]+\)\s*-\s*.+$`)

// zddc filename: trackingNumber_revision (status) - title.ext
// trackingNumber: no spaces or underscores
// revision: ~?[A-Z0-9]+(+[CBNQ][0-9]+)?
var zddcFilenameRE = regexp.MustCompile(
	`^([^_\s]+(?:-[^_\s]+)*)_(~?[A-Z0-9]+)(\+[CBNQ][0-9]+)?\s+\([^)]+\)\s*-\s*.+\.([^.]+)$`,
)

type parsedFile struct {
	trackingNumber string
	baseRev        string
	modifier       string // empty or e.g. "C1"
	date           string // transmittal folder date
	serverPath     string // server-relative path (slash-separated, no leading slash)
}

// BuildIndex walks fsRoot, finds all transmittal folders, and builds the index.
func BuildIndex(fsRoot string) (*Index, error) {
	idx := NewIndex()
	if err := walkAndIndex(idx, fsRoot, fsRoot, ""); err != nil {
		return nil, err
	}
	return idx, nil
}

// walkAndIndex recursively walks dirAbs looking for transmittal folders.
// serverDir is the server-relative path of dirAbs (slash-separated, no leading slash).
func walkAndIndex(idx *Index, fsRoot, dirAbs, serverDir string) error {
	entries, err := os.ReadDir(dirAbs)
	if err != nil {
		return err
	}

	for _, entry := range entries {
		name := entry.Name()
		if strings.HasPrefix(name, ".") {
			continue
		}
		if !entry.IsDir() {
			continue
		}

		var childServerDir string
		if serverDir == "" {
			childServerDir = name
		} else {
			childServerDir = serverDir + "/" + name
		}
		childAbs := filepath.Join(dirAbs, name)

		if m := transmittalFolderRE.FindStringSubmatch(name); m != nil {
			// This is a transmittal folder — index its files
			date := m[1]
			if err := indexTransmittalFolder(idx, fsRoot, childAbs, childServerDir, date); err != nil {
				// Non-fatal: log and continue
				continue
			}
		} else {
			// Recurse into grouping/portfolio/project folders
			if err := walkAndIndex(idx, fsRoot, childAbs, childServerDir); err != nil {
				continue
			}
		}
	}
	return nil
}

// indexTransmittalFolder indexes all ZDDC files in a transmittal folder.
func indexTransmittalFolder(idx *Index, fsRoot, folderAbs, folderServerPath, date string) error {
	return filepath.WalkDir(folderAbs, func(path string, d os.DirEntry, err error) error {
		if err != nil {
			// Log the error but continue indexing other files
			_ = err // would log here: slog.Warn("walkdir error", "path", path, "err", err)
			return nil
		}
		if d.IsDir() {
			return nil
		}
		name := d.Name()
		if strings.HasPrefix(name, ".") {
			return nil
		}

		m := zddcFilenameRE.FindStringSubmatch(name)
		if m == nil {
			return nil
		}

		tracking := m[1]
		baseRev := m[2]
		modifierFull := m[3] // e.g. "+C1" or ""
		modifier := ""
		if modifierFull != "" {
			modifier = modifierFull[1:] // strip leading "+"
		}

		// Build server-relative path
		relPath, err := filepath.Rel(fsRoot, path)
		if err != nil {
			return nil
		}
		serverPath := filepath.ToSlash(relPath)

		pf := parsedFile{
			trackingNumber: tracking,
			baseRev:        baseRev,
			modifier:       modifier,
			date:           date,
			serverPath:     serverPath,
		}

		idx.recordFile(pf)
		return nil
	})
}

// projectOf returns the top-level slash-separated segment of a server-relative
// path. Files at the root (no slash) have no project and are not indexable.
func projectOf(serverPath string) string {
	i := strings.IndexByte(serverPath, '/')
	if i <= 0 {
		return ""
	}
	return serverPath[:i]
}

// recordFile adds a parsed file to the index using first-seen (oldest date)
// logic, bucketed under the project (top-level segment) the file lives in.
func (idx *Index) recordFile(pf parsedFile) {
	project := projectOf(pf.serverPath)
	if project == "" {
		// File sits directly at the served root with no project wrapper.
		// Skipping it means /.archive/ at the root surfaces nothing — which
		// is exactly the contract: stable references must include a project
		// directory. Such files are still reachable as ordinary static URLs.
		return
	}

	idx.mu.Lock()
	defer idx.mu.Unlock()

	pe, ok := idx.ByProject[project]
	if !ok {
		pe = &ProjectEntry{ByTracking: make(map[string]*TrackingEntry)}
		idx.ByProject[project] = pe
	}

	te, ok := pe.ByTracking[pf.trackingNumber]
	if !ok {
		te = &TrackingEntry{
			ByRevision: make(map[string]*RevisionEntry),
		}
		pe.ByTracking[pf.trackingNumber] = te
	}

	re, ok := te.ByRevision[pf.baseRev]
	if !ok {
		re = &RevisionEntry{
			Modifiers: make(map[string]string),
		}
		te.ByRevision[pf.baseRev] = re
	}

	if pf.modifier == "" {
		switch {
		case re.BasePath == "":
			re.BasePath = pf.serverPath
			re.Date = pf.date
		case re.BasePath == pf.serverPath:
			// same file, no-op (e.g. re-index from the watcher)
		default:
			// Two different files claim to be (project, tracking, rev) —
			// that's a within-project authoring mistake. Log once with both
			// paths so it's diagnosable; chronological winner still wins.
			slog.Warn("archive: within-project revision collision",
				"project", project,
				"tracking", pf.trackingNumber,
				"revision", pf.baseRev,
				"existing", re.BasePath,
				"existingDate", re.Date,
				"new", pf.serverPath,
				"newDate", pf.date,
			)
			if pf.date < re.Date {
				re.BasePath = pf.serverPath
				re.Date = pf.date
			}
		}
	} else {
		// Modifier file — record if no entry yet or this transmittal is older
		if existing, exists := re.Modifiers[pf.modifier]; !exists || pf.date < re.Date {
			_ = existing
			re.Modifiers[pf.modifier] = pf.serverPath
		}
	}

	// Update highest base revision
	te.HighestBaseRev = highestRevision(te)
}

// highestRevision returns the highest base revision among all revisions in te.
// Revision ordering: numeric revisions (0,1,2…) are lower than alphabetic (A,B,C…).
// Draft prefix ~ means lower than base.
func highestRevision(te *TrackingEntry) string {
	if len(te.ByRevision) == 0 {
		return ""
	}
	revs := make([]string, 0, len(te.ByRevision))
	for r := range te.ByRevision {
		revs = append(revs, r)
	}
	sort.Slice(revs, func(i, j int) bool {
		return compareRevisions(revs[i], revs[j]) < 0
	})
	return revs[len(revs)-1]
}

// compareRevisions returns negative if a < b, 0 if equal, positive if a > b.
// Order: ~rev < numeric < alpha (A < B < C ...)
func compareRevisions(a, b string) int {
	isDraftA := strings.HasPrefix(a, "~")
	isDraftB := strings.HasPrefix(b, "~")
	baseA := strings.TrimPrefix(a, "~")
	baseB := strings.TrimPrefix(b, "~")

	// Draft < non-draft of same base
	if baseA == baseB {
		if isDraftA && !isDraftB {
			return -1
		}
		if !isDraftA && isDraftB {
			return 1
		}
		return 0
	}

	// Numeric vs alpha: numeric comes first
	aIsNum := len(baseA) > 0 && baseA[0] >= '0' && baseA[0] <= '9'
	bIsNum := len(baseB) > 0 && baseB[0] >= '0' && baseB[0] <= '9'
	if aIsNum && !bIsNum {
		return -1
	}
	if !aIsNum && bIsNum {
		return 1
	}

	// Both numeric or both alpha: string comparison (works for single-char alpha)
	if baseA < baseB {
		return -1
	}
	if baseA > baseB {
		return 1
	}
	return 0
}

// UpdateFromDir re-indexes a single transmittal folder (called by the watcher).
func (idx *Index) UpdateFromDir(fsRoot, transmittalDirPath string) error {
	// Determine the date from the folder name
	folderName := filepath.Base(transmittalDirPath)
	m := transmittalFolderRE.FindStringSubmatch(folderName)
	if m == nil {
		return nil // not a transmittal folder
	}
	date := m[1]

	// Compute server-relative path for this folder
	rel, err := filepath.Rel(fsRoot, transmittalDirPath)
	if err != nil {
		return err
	}
	serverDir := filepath.ToSlash(rel)

	return indexTransmittalFolder(idx, fsRoot, transmittalDirPath, serverDir, date)
}

// Entry is one virtual redirect file in the archive listing.
//
// URLName is the filename surfaced under .archive/ (e.g. "123.html",
// "123_~A.html"). TargetPath is the server-relative path the redirect
// resolves to — used both as the redirect target and as the input to the
// per-entry ACL check.
type Entry struct {
	URLName    string
	TargetPath string
}

// AllEntries returns a sorted snapshot of every redirect entry for the named
// project. Two kinds per tracking number:
//
//   - <tracking>.html        → first-chronological copy of the highest base rev
//   - <tracking>_<rev>.html  → first-chronological copy of that specific base rev
//
// Modifier files (e.g. <tracking>_<rev>+C1.html) remain reachable via the
// resolver but are not surfaced in the listing — they're return traffic
// (comments / markups), not items the user browses to as primary documents.
//
// Sort order is by URLName; the "." in <tracking>.html sorts before the "_"
// in <tracking>_<rev>.html, so each tracking number's highest-rev shortcut
// comes first, followed by its individual revisions in revision order.
//
// An empty project (or one with no indexed tracking numbers) returns nil,
// keeping the caller branch-free.
func (idx *Index) AllEntries(project string) []Entry {
	idx.mu.RLock()
	defer idx.mu.RUnlock()

	pe, ok := idx.ByProject[project]
	if !ok {
		return nil
	}

	var result []Entry
	for tn, te := range pe.ByTracking {
		if te.HighestBaseRev != "" {
			if re, ok := te.ByRevision[te.HighestBaseRev]; ok && re.BasePath != "" {
				result = append(result, Entry{
					URLName:    tn + ".html",
					TargetPath: re.BasePath,
				})
			}
		}
		for rev, re := range te.ByRevision {
			if re.BasePath == "" {
				continue
			}
			result = append(result, Entry{
				URLName:    tn + "_" + rev + ".html",
				TargetPath: re.BasePath,
			})
		}
	}
	sort.Slice(result, func(i, j int) bool {
		return result[i].URLName < result[j].URLName
	})
	return result
}