feat: Replace VaryLayerStorage to plain storage system with algorithm (#93)

* feat: Replace VaryLayerStorage to plain storage system with algorithm

* Remove caddy update

* Fix coalescing + vary mannagement

* Use plain concatenation instead of Sprintf method for varied key

Author: darkweak

cache/coalescing/requestCoalescing.go

@@ -12,7 +12,9 @@ import (
 
 // Temporise will run one call to proxy then use the response for other requests that couldn't reach cached response
 func (r *RequestCoalescing) Temporise(req *http.Request, rw http.ResponseWriter, nextMiddleware func(http.ResponseWriter, *http.Request) error) {
-	ch := r.requestGroup.DoChan(rfc.GetCacheKey(req), func() (interface{}, error) {
+	key := rfc.GetCacheKey(req)
+	ch := r.requestGroup.DoChan(key, func() (interface{}, error) {
+		defer r.requestGroup.Forget(key)
 		e := nextMiddleware(rw, req)
 
 		return nil, e

cache/providers/abstractProvider.go

@@ -3,8 +3,12 @@ package providers
 import (
 	"github.com/darkweak/souin/cache/types"
 	"github.com/darkweak/souin/configurationtypes"
+	"net/http"
+	"strings"
 )
 
+const VarySeparator = "{-VARY-}"
+
 // InitializeProvider allow to generate the providers array according to the configuration
 func InitializeProvider(configuration configurationtypes.AbstractConfigurationInterface) types.AbstractProviderInterface {
 	var r types.AbstractProviderInterface
@@ -23,3 +27,27 @@ func InitializeProvider(configuration configurationtypes.AbstractConfigurationIn
 	}
 	return r
 }
+
+func varyVoter(baseKey string, req *http.Request, currentKey string) bool {
+	if currentKey == baseKey {
+		return true
+	}
+
+	if strings.Contains(currentKey, VarySeparator) {
+		list := currentKey[(strings.LastIndex(currentKey, VarySeparator) + len(VarySeparator)):]
+		if len(list) == 0 {
+			return false
+		}
+
+		for _, item := range strings.Split(list, ";") {
+			index := strings.LastIndex(item, ":")
+			if req.Header.Get(item[:index]) != item[index+1:] {
+				return false
+			}
+		}
+
+		return true
+	}
+
+	return false
+}

cache/providers/badgerProvider.go

@@ -4,6 +4,7 @@ import (
 	"fmt"
 	t "github.com/darkweak/souin/configurationtypes"
 	badger "github.com/dgraph-io/badger/v3"
+	"net/http"
 	"regexp"
 	"time"
 )
@@ -65,6 +66,28 @@ func (provider *Badger) Get(key string) []byte {
 	return result
 }
 
+// Prefix method returns the populated response if exists, empty response then
+func (provider *Badger) Prefix(key string, req *http.Request) []byte {
+	var result []byte
+
+	_ = provider.DB.View(func(txn *badger.Txn) error {
+		prefix := []byte(key)
+		it := txn.NewIterator(badger.DefaultIteratorOptions)
+		defer it.Close()
+		for it.Seek(prefix); it.ValidForPrefix(prefix); it.Next() {
+			if varyVoter(key, req, string(it.Item().Key())) {
+				_ = it.Item().Value(func(val []byte) error {
+					result = val
+					return nil
+				})
+			}
+		}
+		return nil
+	})
+
+	return result
+}
+
 // Set method will store the response in Badger provider
 func (provider *Badger) Set(key string, value []byte, url t.URL, duration time.Duration) {
 	if duration == 0 {

cache/providers/embeddedOlricProvider.go

@@ -11,6 +11,7 @@ import (
 	"go.uber.org/zap"
 	"gopkg.in/yaml.v3"
 	"io/ioutil"
+	"net/http"
 	"os"
 	"time"
 )
@@ -123,6 +124,33 @@ func (provider *EmbeddedOlric) ListKeys() []string {
 	return keys
 }
 
+// Prefix method returns the populated response if exists, empty response then
+func (provider *EmbeddedOlric) Prefix(key string, req *http.Request) []byte {
+	c, err := provider.dm.Query(query.M{
+		"$onKey": query.M{
+			"$regexMatch": "^" + key,
+		},
+	})
+
+	if err != nil {
+		fmt.Println(fmt.Sprintf("An error occurred while trying to retrieve data in Olric: %s", err))
+		return []byte{}
+	}
+	defer c.Close()
+
+	res := []byte{}
+	err = c.Range(func(k string, v interface{}) bool {
+		if varyVoter(key, req, k) {
+			res = v.([]byte)
+			return false
+		}
+
+		return true
+	})
+
+	return res
+}
+
 // Get method returns the populated response if exists, empty response then
 func (provider *EmbeddedOlric) Get(key string) []byte {
 	val2, err := provider.dm.Get(key)

cache/providers/olricProvider.go

@@ -6,6 +6,7 @@ import (
 	"github.com/buraksezer/olric/config"
 	"github.com/buraksezer/olric/query"
 	t "github.com/darkweak/souin/configurationtypes"
+	"net/http"
 	"time"
 )
 
@@ -62,6 +63,33 @@ func (provider *Olric) ListKeys() []string {
 	return keys
 }
 
+// Prefix method returns the populated response if exists, empty response then
+func (provider *Olric) Prefix(key string, req *http.Request) []byte {
+	c, err := provider.dm.Query(query.M{
+		"$onKey": query.M{
+			"$regexMatch": "^" + key,
+		},
+	})
+
+	if err != nil {
+		fmt.Println(fmt.Sprintf("An error occurred while trying to retrieve data in Olric: %s", err))
+		return []byte{}
+	}
+	defer c.Close()
+
+	res := []byte{}
+	err = c.Range(func(k string, v interface{}) bool {
+		if varyVoter(key, req, k) {
+			res = v.([]byte)
+			return false
+		}
+
+		return true
+	})
+
+	return res
+}
+
 // Get method returns the populated response if exists, empty response then
 func (provider *Olric) Get(key string) []byte {
 	val2, err := provider.dm.Get(key)

cache/types/layerStorage.go

@@ -4,39 +4,6 @@ import (
 	"github.com/dgraph-io/ristretto"
 )
 
-// VaryLayerStorage is the layer for Vary support storage
-type VaryLayerStorage struct {
-	*ristretto.Cache
-}
-
-// InitializeVaryLayerStorage initialize the storage
-func InitializeVaryLayerStorage() *VaryLayerStorage {
-	storage, _ := ristretto.NewCache(&ristretto.Config{
-		NumCounters: 1e7,     // number of keys to track frequency of (10M).
-		MaxCost:     1 << 30, // maximum cost of cache (1GB).
-		BufferItems: 64,      // number of keys per Get buffer.
-	})
-
-	return &VaryLayerStorage{Cache: storage}
-}
-
-// Get method returns the varied headers list if exists, empty array then
-func (provider *VaryLayerStorage) Get(key string) []string {
-	val, found := provider.Cache.Get(key)
-	if !found {
-		return []string{}
-	}
-	return val.([]string)
-}
-
-// Set method will store the response in Ristretto provider
-func (provider *VaryLayerStorage) Set(key string, headers []string) {
-	isSet := provider.Cache.Set(key, headers, 1)
-	if !isSet {
-		panic("Impossible to set value into Ristretto")
-	}
-}
-
 // CoalescingLayerStorage is the layer to be able to not coalesce uncoalesceable request
 type CoalescingLayerStorage struct {
 	*ristretto.Cache

cache/types/providers.go

@@ -2,12 +2,14 @@ package types
 
 import (
 	"github.com/darkweak/souin/configurationtypes"
+	"net/http"
 	"time"
 )
 
 // AbstractProviderInterface should be implemented in any providers
 type AbstractProviderInterface interface {
 	ListKeys() []string
+	Prefix(key string, req *http.Request) []byte
 	Get(key string) []byte
 	Set(key string, value []byte, url configurationtypes.URL, duration time.Duration)
 	Delete(key string)

cache/types/souin.go

@@ -13,7 +13,6 @@ type TransportInterface interface {
 	RoundTrip(req *http.Request) (resp *http.Response, err error)
 	SetURL(url configurationtypes.URL)
 	UpdateCacheEventually(req *http.Request) (resp *http.Response, err error)
-	GetVaryLayerStorage() *VaryLayerStorage
 	GetCoalescingLayerStorage() *CoalescingLayerStorage
 	GetYkeyStorage() *ykeys.YKeyStorage
 }
@@ -28,7 +27,6 @@ type Transport struct {
 	Provider               AbstractProviderInterface
 	ConfigurationURL       configurationtypes.URL
 	MarkCachedResponses    bool
-	VaryLayerStorage       *VaryLayerStorage
 	CoalescingLayerStorage *CoalescingLayerStorage
 	YkeyStorage            *ykeys.YKeyStorage
 }

plugins/base.go

@@ -27,10 +27,6 @@ func DefaultSouinPluginCallback(
 
 	go func() {
 		cacheKey := rfc.GetCacheKey(req)
-		varied := retriever.GetTransport().GetVaryLayerStorage().Get(cacheKey)
-		if len(varied) != 0 {
-			cacheKey = rfc.GetVariedCacheKey(req, varied)
-		}
 		go func() {
 			coalesceable <- retriever.GetTransport().GetCoalescingLayerStorage().Exists(cacheKey)
 		}()
@@ -69,7 +65,6 @@ func DefaultSouinPluginCallback(
 	} else {
 		_ = nextMiddleware(res, req)
 	}
-	close(coalesceable)
 }
 
 // DefaultSouinPluginInitializerFromConfiguration is the default initialization for plugins

plugins/caddy/app.go

@@ -4,6 +4,7 @@ import (
 	"github.com/caddyserver/caddy/v2"
 )
 
+// SouinApp contains the whole Souin necessary items
 type SouinApp struct {
 	*DefaultCache
 	LogLevel string `json:"log_level,omitempty"`
@@ -32,7 +33,7 @@ func (s SouinApp) Stop() error {
 }
 
 // CaddyModule implements caddy.ModuleInfo
-func (a SouinApp) CaddyModule() caddy.ModuleInfo {
+func (s SouinApp) CaddyModule() caddy.ModuleInfo {
 	return caddy.ModuleInfo{
 		ID:  moduleName,
 		New: func() caddy.Module { return new(SouinApp) },

plugins/caddy/go.sum

@@ -150,6 +150,8 @@ github.com/creack/pty v1.1.7/go.mod h1:lj5s0c3V2DBrqTV7llrYr5NG6My20zk30Fl46Y7Do
 github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
 github.com/daaku/go.zipexe v1.0.0/go.mod h1:z8IiR6TsVLEYKwXAoE/I+8ys/sDkgTzSL0CLnGVd57E=
 github.com/danwakefield/fnmatch v0.0.0-20160403171240-cbb64ac3d964/go.mod h1:Xd9hchkHSWYkEqJwUGisez3G1QY8Ryz0sdWrLPMGjLk=
+github.com/darkweak/souin v1.5.2 h1:WS/Q+qq7c+dIIw/abYYtWMpfj1ucjhfvaabW5Q+K6tc=
+github.com/darkweak/souin v1.5.2/go.mod h1:7TNt0JUJaB3Aq7wP0L45XQhwJJbKw5zMmfp6OTZTFMA=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=

plugins/caddy/main.go

@@ -37,11 +37,11 @@ type SouinCaddyPlugin struct {
 	Configuration *Configuration
 	logger        *zap.Logger
 	LogLevel      string `json:"log_level,omitempty"`
-	bufPool       sync.Pool
+	bufPool       *sync.Pool
 	Headers       []string                           `json:"headers,omitempty"`
 	Olric         configurationtypes.CacheProvider   `json:"olric,omitempty"`
 	TTL           string                             `json:"ttl,omitempty"`
-	ykeys         map[string]configurationtypes.YKey `json:"ykeys,omitempty"`
+	YKeys         map[string]configurationtypes.YKey `json:"ykeys,omitempty"`
 }
 
 // CaddyModule returns the Caddy module information.
@@ -105,6 +105,7 @@ func (s *SouinCaddyPlugin) configurationPropertyMapper() error {
 			LogLevel:     s.LogLevel,
 		}
 	}
+	s.Configuration.Ykeys = s.YKeys
 	s.Configuration.DefaultCache = defaultCache
 	return nil
 }
@@ -163,7 +164,7 @@ func (s *SouinCaddyPlugin) Provision(ctx caddy.Context) error {
 		return err
 	}
 
-	s.bufPool = sync.Pool{
+	s.bufPool = &sync.Pool{
 		New: func() interface{} {
 			return new(bytes.Buffer)
 		},

rfc/bridge.go

@@ -17,7 +17,7 @@ const (
 // otherwise.
 func CachedResponse(c types.AbstractProviderInterface, req *http.Request, cachedKey string, transport types.TransportInterface, update bool) (types.ReverseResponse, error) {
 	clonedReq := cloneRequest(req)
-	cachedVal := c.Get(cachedKey)
+	cachedVal := c.Prefix(cachedKey, req)
 	b := bytes.NewBuffer(cachedVal)
 	response, _ := http.ReadResponse(bufio.NewReader(b), clonedReq)
 	if update && nil != response && ValidateCacheControl(response) {
@@ -61,10 +61,6 @@ func (t *VaryTransport) BaseRoundTrip(req *http.Request, shouldReUpdate bool) (s
 	cacheable := IsVaryCacheable(req)
 	cachedResp := req.Response
 	if cacheable {
-		varied := t.GetVaryLayerStorage().Get(cacheKey)
-		if len(varied) != 0 {
-			cacheKey = GetVariedCacheKey(req, varied)
-		}
 		cr, _ := CachedResponse(t.GetProvider(), req, cacheKey, t, shouldReUpdate)
 		if cr.Response != nil {
 			cachedResp = cr.Response

rfc/standalone.go

@@ -5,6 +5,7 @@ import (
 	"bytes"
 	"errors"
 	"fmt"
+	"github.com/darkweak/souin/cache/providers"
 	"io"
 	"net/http"
 	"strings"
@@ -18,11 +19,10 @@ func GetCacheKey(req *http.Request) string {
 
 // GetVariedCacheKey returns the varied cache key for req and resp.
 func GetVariedCacheKey(req *http.Request, headers []string) string {
-	str := ""
-	for _, v := range headers {
-		str += fmt.Sprintf("%s:%s", v, req.Header.Get(v))
+	for i, v := range headers {
+		headers[i] = fmt.Sprintf("%s:%s", v, req.Header.Get(v))
 	}
-	return fmt.Sprintf("%s-[%s]", GetCacheKey(req), strings.Join(headers[:], ";"))
+	return GetCacheKey(req) + providers.VarySeparator + strings.Join(headers, ";")
 }
 
 // getFreshness will return one of fresh/stale/transparent based on the cache-control

rfc/transport.go

@@ -24,7 +24,6 @@ func IsVaryCacheable(req *http.Request) bool {
 func NewTransport(p types.AbstractProviderInterface, ykeyStorage *ykeys.YKeyStorage) *VaryTransport {
 	return &VaryTransport{
 		Provider:               p,
-		VaryLayerStorage:       types.InitializeVaryLayerStorage(),
 		CoalescingLayerStorage: types.InitializeCoalescingLayerStorage(),
 		MarkCachedResponses:    true,
 		YkeyStorage:            ykeyStorage,
@@ -41,11 +40,6 @@ func (t *VaryTransport) SetURL(url configurationtypes.URL) {
 	t.ConfigurationURL = url
 }
 
-// GetVaryLayerStorage get the vary layer storagecache/coalescing/requestCoalescing_test.go
-func (t *VaryTransport) GetVaryLayerStorage() *types.VaryLayerStorage {
-	return t.VaryLayerStorage
-}
-
 // GetCoalescingLayerStorage get the coalescing layer storage
 func (t *VaryTransport) GetCoalescingLayerStorage() *types.CoalescingLayerStorage {
 	return t.CoalescingLayerStorage

rfc/vary.go

@@ -23,9 +23,6 @@ func validateVary(req *http.Request, resp *http.Response, key string, t *VaryTra
 		variedHeaders := headerAllCommaSepValues(resp.Header, "vary")
 		cacheKey := key
 		if len(variedHeaders) > 0 {
-			go func() {
-				t.VaryLayerStorage.Set(key, variedHeaders)
-			}()
 			cacheKey = GetVariedCacheKey(req, variedHeaders)
 		}
 		switch req.Method {