Add a pure-Go zim reader and writer

ZIM is the open single-file archive format Kiwix uses for offline content:
a fixed header, a MIME list, URL/title/cluster pointer lists, directory
entries, zstd-compressed or stored clusters, and a trailing MD5. The writer
lays out a mirror in two passes (assign positions, then emit bytes) and
derives the UUID from the content so packing is deterministic. The reader
random-accesses entries by namespace and url, follows redirects, and reads
xz clusters too so archives from other tooling open.

Cross-checked against an independent reader (gozim): header, MIME list,
namespaces, urls, dirents, and a non-last cluster's blob all read back
byte-for-byte.

zim/codec.go

@@ -0,0 +1,48 @@
+package zim
+
+import (
+	"sync"
+
+	"github.com/klauspost/compress/zstd"
+)
+
+// A single shared zstd codec. Both EncodeAll and DecodeAll are safe for
+// concurrent use, so one encoder and one decoder serve the whole process.
+var (
+	zstdOnce sync.Once
+	zstdEnc  *zstd.Encoder
+	zstdDec  *zstd.Decoder
+)
+
+func initZstd() {
+	zstdOnce.Do(func() {
+		zstdEnc, _ = zstd.NewWriter(nil, zstd.WithEncoderLevel(zstd.SpeedBetterCompression))
+		zstdDec, _ = zstd.NewReader(nil)
+	})
+}
+
+func zstdEncode(p []byte) []byte {
+	initZstd()
+	return zstdEnc.EncodeAll(p, nil)
+}
+
+func zstdDecode(p []byte) ([]byte, error) {
+	initZstd()
+	return zstdDec.DecodeAll(p, nil)
+}
+
+// isTextMime reports whether content of this MIME type is worth compressing.
+// Already-compressed media (images, fonts, audio, video, archives) is stored
+// uncompressed so we do not burn CPU inflating it by a few bytes.
+func isTextMime(mime string) bool {
+	switch mime {
+	case "application/json", "application/xml", "application/javascript",
+		"application/x-javascript":
+		return true
+	}
+	if len(mime) >= 5 && mime[:5] == "text/" {
+		return true
+	}
+	// Any structured-XML type: application/rss+xml, image/svg+xml, ...
+	return len(mime) >= 4 && mime[len(mime)-4:] == "+xml"
+}