Introduce "Segments Tree Tracker"

Currently the flattener and more specifically "NameTracker" is not hierarchy aware, to explain that let's examine this case:
Pattern: `{ "context": { "user_id": [1] } }`
Event: `{ "context": { "user_id": 1 }, "payload": { /* lots of fields, large payload */ }`

In this case the flattener will read context and then user_id, all the other values (strings, objects and arrays) will be skipped, but still will get traversed.
By replacing the "NameTracker" with "SegmentsTreeTracker" which is aware of the hierarchy, we can:
* Know that once we read "user_id" we can stop consuming the event and exit
* Cache the paths - so `pathForChild` is not needed and we reduce allocations.

```
name                                         old time/op    new time/op    delta
CityLots-10                                    4.42µs ± 0%    4.28µs ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattener_ContextFields-10                9.06µs ± 0%    0.21µs ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattener_MiddleNestedField-10            10.9µs ± 0%     1.8µs ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattener_LastField-10                    9.65µs ± 0%    9.68µs ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_ContextFields-10        9.42µs ± 0%    0.50µs ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_MiddleNestedField-10    11.0µs ± 0%     2.0µs ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_LastField-10            9.81µs ± 0%    9.87µs ± 0%   ~     (p=1.000 n=1+1)

name                                         old alloc/op   new alloc/op   delta
CityLots-10                                      985B ± 0%      832B ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattener_ContextFields-10                 48.0B ± 0%      0.0B        ~     (p=1.000 n=1+1)
_JsonFlattener_MiddleNestedField-10             64.0B ± 0%      0.0B        ~     (p=1.000 n=1+1)
_JsonFlattener_LastField-10                     32.0B ± 0%      0.0B        ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_ContextFields-10          232B ± 0%      184B ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_MiddleNestedField-10      240B ± 0%       48B ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_LastField-10             80.0B ± 0%     48.0B ± 0%   ~     (p=1.000 n=1+1)

name                                         old allocs/op  new allocs/op  delta
CityLots-10                                      38.0 ± 0%      31.0 ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattener_ContextFields-10                  3.00 ± 0%      0.00        ~     (p=1.000 n=1+1)
_JsonFlattener_MiddleNestedField-10              4.00 ± 0%      0.00        ~     (p=1.000 n=1+1)
_JsonFlattener_LastField-10                      2.00 ± 0%      0.00        ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_ContextFields-10          9.00 ± 0%      6.00 ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_MiddleNestedField-10      9.00 ± 0%      3.00 ± 0%   ~     (p=1.000 n=1+1)
_JsonFlattner_Evaluate_LastField-10              5.00 ± 0%      3.00 ± 0%   ~     (p=1.000 n=1+1)
```

Signed-off-by: Yosi Attias <[email protected]>

anything_but_test.go

@@ -129,7 +129,7 @@ func TestParseAnythingButPattern(t *testing.T) {
 	}
 
 	for i, good := range goods {
-		fields, _, err := patternFromJSON([]byte(good))
+		fields, err := patternFromJSON([]byte(good))
 		if err != nil {
 			t.Errorf("parse anything-but i=%d: "+err.Error(), i)
 		}
@@ -139,7 +139,7 @@ func TestParseAnythingButPattern(t *testing.T) {
 	}
 
 	for _, bad := range bads {
-		_, _, err := patternFromJSON([]byte(bad))
+		_, err := patternFromJSON([]byte(bad))
 		if err == nil {
 			t.Errorf(`accepted anything-but "%s"`, bad)
 		}

arrays_test.go

@@ -70,30 +70,29 @@ const bands = `{
 }`
 
 func TestArrayCorrectness(t *testing.T) {
-	// only pattern3 should match
-	pattern1 := `{"bands": { "members": { "given": [ "Mick" ], "surname": [ "Strummer" ] } } }`
-	pattern2 := `{"bands": { "members": { "given": [ "Wata" ], "role": [ "drums" ] } } }`
-	pattern3 := `{"bands": { "members": { "given": [ "Wata" ], "role": [ "guitar" ] } } }`
+	// only wataGuiterPattern should match
+	mickStrummerPattern := `{"bands": { "members": { "given": [ "Mick" ], "surname": [ "Strummer" ] } } }`
+	wataDrumsPattern := `{"bands": { "members": { "given": [ "Wata" ], "role": [ "drums" ] } } }`
+	wataGuiterPattern := `{"bands": { "members": { "given": [ "Wata" ], "role": [ "guitar" ] } } }`
+
 	m := newCoreMatcher()
-	err := m.addPattern("Mick strummer", pattern1)
-	if err != nil {
-		t.Error(err.Error())
+	if err := m.addPattern("Mick strummer", mickStrummerPattern); err != nil {
+		t.Errorf("Failed adding pattern: %s: %s", mickStrummerPattern, err)
 	}
-	err = m.addPattern("Wata drums", pattern2)
-	if err != nil {
-		t.Error(err.Error())
+
+	if err := m.addPattern("Wata drums", wataDrumsPattern); err != nil {
+		t.Errorf("Failed adding pattern: %s: %s", wataDrumsPattern, err)
 	}
-	err = m.addPattern("Wata guitar", pattern3)
-	if err != nil {
-		t.Error(err.Error())
+	if err := m.addPattern("Wata guitar", wataGuiterPattern); err != nil {
+		t.Errorf("Failed adding pattern: %s: %s", wataGuiterPattern, err)
 	}
 
 	matches, err := m.matchesForJSONEvent([]byte(bands))
 	if err != nil {
-		t.Error(err.Error())
+		t.Errorf("Failed 'matchesForJSONEvent': %s", err)
 	}
 
 	if len(matches) != 1 || matches[0].(string) != "Wata guitar" {
-		t.Error("Matches across array boundaries")
+		t.Errorf("Expected to get a single of 'Wata guiter', but got %d matches: %+v", len(matches), matches)
 	}
 }

benchmarks_test.go

@@ -160,7 +160,7 @@ func TestBigShellStyle(t *testing.T) {
 	before := time.Now()
 	fj := newJSONFlattener()
 	for _, line := range lines {
-		fields, err := fj.Flatten(line, m)
+		fields, err := fj.Flatten(line, m.getSegmentsTreeTracker())
 		if err != nil {
 			t.Error("Flatten: " + err.Error())
 		}

core_matcher.go

@@ -31,15 +31,12 @@ type coreMatcher struct {
 
 // coreFields groups the updateable fields in coreMatcher.
 // state is the start of the automaton.
-// namesUsed is a map of field names that are used in any of the patterns that this automaton encodes. Typically,
+// segmentsTree is a tree of segments that are used in any of the patterns that this automaton encodes.Typically,
 // patterns only consider a subset of the fields in an incoming data object, and there is no reason to consider
 // fields that do not appear in patterns when using the automaton for matching.
-// fakeField is used when the flattener for an event returns no fields, because it could still match if
-// there were patterns with "exists":false. So in this case we run one fake field through the matcher
-// which will cause it to notice that any "exists":false patterns should match.
 type coreFields struct {
-	state     *fieldMatcher
-	namesUsed map[string]bool
+	state        *fieldMatcher
+	segmentsTree *segmentsTree
 }
 
 func newCoreMatcher() *coreMatcher {
@@ -49,8 +46,8 @@ func newCoreMatcher() *coreMatcher {
 	// user-supplied path-name because it's not valid in UTF-8
 	m := coreMatcher{}
 	m.updateable.Store(&coreFields{
-		state:     newFieldMatcher(),
-		namesUsed: make(map[string]bool),
+		state:        newFieldMatcher(),
+		segmentsTree: newSegmentsIndex(),
 	})
 	return &m
 }
@@ -62,7 +59,7 @@ func (m *coreMatcher) start() *coreFields {
 // addPattern - the patternBytes is a JSON object. The X is what the matcher returns to indicate that the
 // provided pattern has been matched. In many applications it might be a string which is the pattern's name.
 func (m *coreMatcher) addPattern(x X, patternJSON string) error {
-	patternFields, patternNamesUsed, err := patternFromJSON([]byte(patternJSON))
+	patternFields, err := patternFromJSON([]byte(patternJSON))
 	if err != nil {
 		return err
 	}
@@ -75,15 +72,13 @@ func (m *coreMatcher) addPattern(x X, patternJSON string) error {
 
 	// we build up the new coreMatcher state in freshStart so we can atomically switch it in once complete
 	freshStart := &coreFields{}
-	freshStart.namesUsed = make(map[string]bool)
 	current := m.start()
+	freshStart.segmentsTree = current.segmentsTree.copy()
 	freshStart.state = current.state
 
-	for k := range current.namesUsed {
-		freshStart.namesUsed[k] = true
-	}
-	for used := range patternNamesUsed {
-		freshStart.namesUsed[used] = true
+	// Add paths to the segments tree index.
+	for _, field := range patternFields {
+		freshStart.segmentsTree.add(field.path)
 	}
 
 	// now we add each of the name/value pairs in fields slice to the automaton, starting with the start state -
@@ -132,7 +127,7 @@ func (m *coreMatcher) deletePatterns(_ X) error {
 // This is a leftover from previous times, is only used by tests, but it's used by a *lot*
 // so removing it would require a lot of tedious work
 func (m *coreMatcher) matchesForJSONEvent(event []byte) ([]X, error) {
-	fields, err := newJSONFlattener().Flatten(event, m)
+	fields, err := newJSONFlattener().Flatten(event, m.getSegmentsTreeTracker())
 	if err != nil {
 		return nil, err
 	}
@@ -251,7 +246,6 @@ func noArrayTrailConflict(from []ArrayPos, to []ArrayPos) bool {
 	return true
 }
 
-func (m *coreMatcher) IsNameUsed(label []byte) bool {
-	_, ok := m.start().namesUsed[string(label)]
-	return ok
+func (m *coreMatcher) getSegmentsTreeTracker() SegmentsTreeTracker {
+	return m.start().segmentsTree
 }

core_matcher_test.go

@@ -248,15 +248,6 @@ func TestSimpleaddPattern(t *testing.T) {
 	if err != nil {
 		t.Error(err.Error())
 	}
-	if len(m.start().namesUsed) != 2 {
-		t.Errorf("nameUsed = %d", len(m.start().namesUsed))
-	}
-	if !m.IsNameUsed([]byte("a")) {
-		t.Error("'a' not showing as used")
-	}
-	if !m.IsNameUsed([]byte("b")) {
-		t.Error("'b' not showing as used")
-	}
 	s0 := m.start().state
 	if len(s0.fields().transitions) != 1 {
 		t.Errorf("s0 trans len %d", len(s0.fields().transitions))

escaping_test.go

@@ -8,7 +8,7 @@ func TestReadMemberName(t *testing.T) {
 	j := `{"😀💋😺": 1, "x\u0078\ud83d\udc8by": "2"}`
 	m := fakeMatcher("😀💋😺", `xx💋y`)
 	f := newJSONFlattener()
-	fields, err := f.Flatten([]byte(j), m)
+	fields, err := f.Flatten([]byte(j), m.getSegmentsTreeTracker())
 	if err != nil {
 		t.Error("TRMN: " + err.Error())
 	}
@@ -26,7 +26,7 @@ func TestReadMemberName(t *testing.T) {
 func TestStringValuesWithEscapes(t *testing.T) {
 	j := `{"a": "x\u0078\ud83d\udc8by", "b": "\ud83d\ude00\ud83d\udc8b\ud83d\ude3a"}`
 	f := newJSONFlattener()
-	fields, err := f.Flatten([]byte(j), fakeMatcher("a", "b"))
+	fields, err := f.Flatten([]byte(j), fakeMatcher("a", "b").getSegmentsTreeTracker())
 	if err != nil {
 		t.Error("TSVWE: " + err.Error())
 	}

external_test.go

@@ -10,7 +10,7 @@ type fakeFlattener struct {
 	r []quamina.Field
 }
 
-func (f *fakeFlattener) Flatten(_ []byte, _ quamina.NameTracker) ([]quamina.Field, error) {
+func (f *fakeFlattener) Flatten(_ []byte, _ quamina.SegmentsTreeTracker) ([]quamina.Field, error) {
 	return f.r, nil
 }