cli,liturgy: lectio --clean prunes readings cache older than one year

1 files changed, 48 insertions, 0 deletions

diff --git a/internal/liturgy/fetch.go b/internal/liturgy/fetch.go
index 32eaaad..f4f6992 100644
--- a/internal/liturgy/fetch.go
+++ b/internal/liturgy/fetch.go
@@ -41,6 +41,12 @@ var publishedRe = regexp.MustCompile(`id="\w*0all"`)
 // validate its own input first.
 var dateRe = regexp.MustCompile(`^\d{4}-\d{2}-\d{2}$`)
 
+// cacheFileRe matches the cache file names Load/Harvest write into
+// cacheDir(): "<YYYY-MM-DD>.html" or "<YYYY-MM-DD>.json". CleanCache uses it
+// to tell cache entries apart from anything else that might be sitting in
+// the directory.
+var cacheFileRe = regexp.MustCompile(`^(\d{4}-\d{2}-\d{2})\.(html|json)$`)
+
 // Options controls how Load resolves a day's readings.
 type Options struct {
 	// Date is the day to load, formatted YYYY-MM-DD.
@@ -135,6 +141,48 @@ func Load(opts Options) ([]Section, error) {
 	return secs, nil
 }
 
+// CleanCache removes cached readings whose date is before `before` from
+// cacheDir(). It matches only files named "<YYYY-MM-DD>.html"/".json";
+// anything else in the directory (e.g. a stray notes.txt, or the sigla
+// store, which lives elsewhere entirely) is left alone. A missing cache dir
+// is not an error -- it just means there is nothing to clean yet.
+func CleanCache(before time.Time) (removed int, freed int64, err error) {
+	dir := cacheDir()
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		if os.IsNotExist(err) {
+			return 0, 0, nil
+		}
+		return 0, 0, err
+	}
+
+	for _, entry := range entries {
+		if entry.IsDir() {
+			continue
+		}
+		m := cacheFileRe.FindStringSubmatch(entry.Name())
+		if m == nil {
+			continue
+		}
+		date, perr := time.Parse("2006-01-02", m[1])
+		if perr != nil || !date.Before(before) {
+			continue
+		}
+
+		path := filepath.Join(dir, entry.Name())
+		info, serr := os.Stat(path)
+		if serr != nil {
+			return removed, freed, serr
+		}
+		if rerr := os.Remove(path); rerr != nil {
+			return removed, freed, rerr
+		}
+		removed++
+		freed += info.Size()
+	}
+	return removed, freed, nil
+}
+
 // loadJSONCache reads and unmarshals the parsed-sections cache file.
 func loadJSONCache(path string) ([]Section, error) {
 	data, err := os.ReadFile(path)