pat: add wildcard pattern

PATTERNS.md

@@ -149,9 +149,9 @@ is not equal to any of the strings in the array.
 If a Field in a Pattern contains an Anything-But Pattern,
 it **MUST NOT** contain any other values.
 
-### Shellstyle Pattern
+### Wildcard Pattern
 
-The Pattern Type of a Shellstyle Pattern is `shellstyle`
+The Pattern Type of a Wildcard Pattern is `wildcard`
 and its value **MUST** be a string which **MAY** contain
 `*` (“star”) characters. The star character
 functions exactly as the same character does in
@@ -164,13 +164,37 @@ Consider the following Event:
 ```json
 {"img": "https://example.com/9943.jpg"}
 ```
-The following Shellstyle Patterns would match it:
+The following Wildcard Patterns would match it:
+```json
+{"img": [ {"wildcard": "*.jpg"} ] }
+{"img": [ {"wildcard": "https://example.com/*"} ] }
+{"img": [ {"wildcard": "https://example.com/*.jpg"} ] }
+{"img": [ {"wildcard": "https://example.*/*.jpg"} ] }
+```
+
+If it is desired to match the actual character "*", it may be “escaped”
+with backslash, "\". For example, consider the following Event:
+
 ```json
-{"img": [ {"shellstyle": "*.jpg"} ] }
-{"img": [ {"shellstyle": "https://example.com/*"} ] }
-{"img": [ {"shellstyle": "https://example.com/*.jpg"} ] }
-{"img": [ {"shellstyle": "https://example.*/*.jpg"} ] }
+{"example-regex": "a**\\.b"}
 ```
+
+The following Wildcard pattern would match it.
+
+```json
+{"example-regex": [ {"wildcard":  "a\\*\\*\\\\.b"}]}
+```
+
+Note that the "\" backslashes must be doubled to deal with the
+fact that they are escape characters for JSON as well as for Quamina.
+
+After a "\", the appearance of any character other than "*" or "\" is an error.
+
+### Shellstyle Pattern
+
+This is an earlier version of the Wildcard pattern, differing only that 
+\-escaping the "*" and "\" characters is not supported.
+
 ### Equals-Ignore-Case Pattern
 
 The Pattern Type of an Equals-Ignore-Case pattern is `equals-ignore-case`
@@ -192,6 +216,6 @@ the AWS EventBridge service, as documented in
 As of release 1.0, Quamina supports Exists and
 Anything-But Patterns, but does not yet support any other
 EventBridge patterns. Note that a
-Shellstyle Pattern with a trailing `*` is equivalent
+Wildcard Pattern with a trailing `*` is equivalent
 to a `prefix` pattern.
 

README.md

@@ -89,23 +89,23 @@ The following Patterns would match it:
 ```json
 {
   "Image": {
-    "Thumbnail": { "Url": [ { "shellstyle": "*9943" } ] }
+    "Thumbnail": { "Url": [ { "wildcard": "*9943" } ] }
   }
 }
 ```
 ```json
 {
   "Image": {
     "Thumbnail": { "Url":
-      [ { "shellstyle": "http://www.example.com/*" } ] }
+      [ { "wildcard": "http://www.example.com/*" } ] }
   }
 }
 ```
 ```json
 {
   "Image": {
     "Thumbnail": { "Url":
-      [ { "shellstyle": "http://www.example.*/*9943" } ] }
+      [ { "wildcard": "http://www.example.*/*9943" } ] }
   }
 }
 ```
@@ -298,7 +298,7 @@ I used to say that the performance of
 `MatchesForEvent` was O(1) in the number of
 Patterns. That’s probably a reasonable way to think
 about it, because it’s *almost* right, except in the
-case where a very large number of `shellstyle` patterns
+case where a very large number of `wildcard` patterns
 have been added; this is discussed in the next section.
 
 To be correct, the performance is a little worse than
@@ -342,7 +342,7 @@ So, adding a new Pattern that only mentions fields which are
 already mentioned in previous Patterns is effectively free,
 i.e. O(1) in terms of run-time performance.
 
-### Quamina instances with large numbers of `shellstyle` Patterns
+### Quamina instances with large numbers of `wildcard` Patterns
 
 A study of the theory of finite automata reveals that processing
 regular-expression constructs such as `*` increases the complexity of

pattern.go

@@ -21,6 +21,7 @@ const (
 	anythingButType
 	prefixType
 	monocaseType
+	wildcardType
 )
 
 // typedVal represents the value of a field in a pattern, giving the value and the type of pattern.
@@ -198,6 +199,8 @@ func readSpecialPattern(pb *patternBuild, valsIn []typedVal) (pathVals []typedVa
 		pathVals, err = readExistsSpecial(pb, pathVals)
 	case "shellstyle":
 		pathVals, err = readShellStyleSpecial(pb, pathVals)
+	case "wildcard":
+		pathVals, err = readWildcardSpecial(pb, pathVals)
 	case "prefix":
 		pathVals, err = readPrefixSpecial(pb, pathVals)
 	case "equals-ignore-case":

pattern_test.go

@@ -69,6 +69,10 @@ func TestPatternFromJSON(t *testing.T) {
 		`{"abc": [ {"prefix":  - "a" }, "foo" ] }`,
 		`{"abc": [ {"prefix":  "a" {, "foo" ] }`,
 		`{"abc": [ {"equals-ignore-case":23}, "foo" ] }`,
+		`{"abc": [ {"wildcard":"15", "x", 1} ] }`,
+		`{"abc": [ {"wildcard":"a**b"}, "foo" ] }`,
+		`{"abc": [ {"wildcard":"a\\b"}, "foo" ] }`,                                             // after JSON parsing, code sees `a/b`
+		`{"abc": [ {"wildcard":"a\\"}, "foo" ] }`,                                              // after JSON parsing, code sees `a\`
 		"{\"a\": [ { \"anything-but\": { \"equals-ignore-case\": [\"1\", \"2\" \"3\"] } } ] }", // missing ,
 		"{\"a\": [ { \"anything-but\": { \"equals-ignore-case\": [1, 2, 3] } } ] }",            // no numbers
 		"{\"a\": [ { \"anything-but\": { \"equals-ignore-case\": [\"1\", \"2\" } } ] }",        // missing ]
@@ -93,6 +97,10 @@ func TestPatternFromJSON(t *testing.T) {
 		`{"abc": [ {"shellstyle":"a*b"}, "foo" ] }`,
 		`{"abc": [ {"shellstyle":"a*b*c"} ] }`,
 		`{"x": [ {"equals-ignore-case":"a*b*c"} ] }`,
+		`{"abc": [ 3, {"wildcard":"a*b"} ] }`,
+		`{"abc": [ {"wildcard":"a*b"}, "foo" ] }`,
+		`{"abc": [ {"wildcard":"a*b*c"} ] }`,
+		`{"abc": [ {"wildcard":"a*b\\*c"} ] }`,
 	}
 	w1 := []*patternField{{path: "x", vals: []typedVal{{vType: numberType, val: "2"}}}}
 	w2 := []*patternField{{path: "x", vals: []typedVal{
@@ -156,7 +164,37 @@ func TestPatternFromJSON(t *testing.T) {
 			},
 		},
 	}
-	wanted := [][]*patternField{w1, w2, w3, w4, w5, w6, w7, w8, w9}
+	w10 := []*patternField{
+		{
+			path: "abc", vals: []typedVal{
+				{vType: stringType, val: "3"},
+				{vType: wildcardType, val: `"a*b"`},
+			},
+		},
+	}
+	w11 := []*patternField{
+		{
+			path: "abc", vals: []typedVal{
+				{vType: wildcardType, val: `"a*b"`},
+				{vType: stringType, val: `"foo"`},
+			},
+		},
+	}
+	w12 := []*patternField{
+		{
+			path: "abc", vals: []typedVal{
+				{vType: wildcardType, val: `"a*b*c"`},
+			},
+		},
+	}
+	w13 := []*patternField{
+		{
+			path: "abc", vals: []typedVal{
+				{vType: wildcardType, val: `"a*b\*c"`},
+			},
+		},
+	}
+	wanted := [][]*patternField{w1, w2, w3, w4, w5, w6, w7, w8, w9, w10, w11, w12, w13}
 
 	for i, good := range goods {
 		fields, err := patternFromJSON([]byte(good))

prettyprinter.go

@@ -58,7 +58,9 @@ func (pp *prettyPrinter) tableLabel(t *smallTable) string {
 
 func (pp *prettyPrinter) labelTable(table *smallTable, label string) {
 	pp.tableLabels[table] = label
-	pp.tableSerials[table] = uint(pp.randInts.Int63()%500 + 500)
+	newSerial := pp.randInts.Int63()%500 + 500
+	//nolint:gosec
+	pp.tableSerials[table] = uint(newSerial)
 }
 
 func (pp *prettyPrinter) printNFA(t *smallTable) string {

value_matcher.go

@@ -116,6 +116,9 @@ func (m *valueMatcher) addTransition(val typedVal, printer printer) *fieldMatche
 		case shellStyleType:
 			newFA, nextField = makeShellStyleFA(valBytes, printer)
 			fields.isNondeterministic = true
+		case wildcardType:
+			newFA, nextField = makeWildcardFA(valBytes, printer)
+			fields.isNondeterministic = true
 		case prefixType:
 			newFA, nextField = makePrefixFA(valBytes)
 		case monocaseType:
@@ -156,6 +159,12 @@ func (m *valueMatcher) addTransition(val typedVal, printer printer) *fieldMatche
 			fields.isNondeterministic = true
 			m.update(fields)
 			return nextField
+		case wildcardType:
+			newAutomaton, nextField := makeWildcardFA(valBytes, printer)
+			fields.startTable = newAutomaton
+			fields.isNondeterministic = true
+			m.update(fields)
+			return nextField
 		case prefixType:
 			newFA, nextField := makePrefixFA(valBytes)
 			fields.startTable = newFA
@@ -194,6 +203,9 @@ func (m *valueMatcher) addTransition(val typedVal, printer printer) *fieldMatche
 	case shellStyleType:
 		newFA, nextField = makeShellStyleFA(valBytes, printer)
 		fields.isNondeterministic = true
+	case wildcardType:
+		newFA, nextField = makeWildcardFA(valBytes, printer)
+		fields.isNondeterministic = true
 	case prefixType:
 		newFA, nextField = makePrefixFA(valBytes)
 	case monocaseType:

value_matcher_test.go

@@ -347,3 +347,75 @@ func TestMakeFAFragment(t *testing.T) {
 		}
 	}
 }
+func TestExerciseSingletonReplacement(t *testing.T) {
+	cm := newCoreMatcher()
+	err := cm.addPattern("x", `{"x": [ "a"]}`)
+	if err != nil {
+		t.Error("AP: " + err.Error())
+	}
+	err = cm.addPattern("x", `{"x": [1]}`)
+	if err != nil {
+		t.Error("AP: " + err.Error())
+	}
+	events := []string{`{"x": 1}`, `{"x": "a"}`}
+	for _, e := range events {
+		matches, err := cm.matchesForJSONEvent([]byte(e))
+		if err != nil {
+			t.Error("m4: " + err.Error())
+		}
+		if len(matches) != 1 || matches[0] != "x" {
+			t.Error("match failed on: " + e)
+		}
+	}
+	events = []string{`{"x": 1}`, `{"x": "a"}`}
+	for _, e := range events {
+		matches, err := cm.matchesForJSONEvent([]byte(e))
+		if err != nil {
+			t.Error("m4: " + err.Error())
+		}
+		if len(matches) != 1 || matches[0] != "x" {
+			t.Error("match failed on: " + e)
+		}
+	}
+	cm = newCoreMatcher()
+	err = cm.addPattern("x", `{"x": ["x"]}`)
+	if err != nil {
+		t.Error("AP: " + err.Error())
+	}
+	err = cm.addPattern("x", `{"x": [ {"wildcard": "x*y"}]}`)
+	if err != nil {
+		t.Error("AP: " + err.Error())
+	}
+	events = []string{`{"x": "x"}`, `{"x": "x..y"}`}
+	for _, e := range events {
+		matches, err := cm.matchesForJSONEvent([]byte(e))
+		if err != nil {
+			t.Error("m4: " + err.Error())
+		}
+		if len(matches) != 1 || matches[0] != "x" {
+			t.Error("match failed on: " + e)
+		}
+	}
+}
+
+func TestMergeNfaAndNumeric(t *testing.T) {
+	cm := newCoreMatcher()
+	err := cm.addPattern("x", `{"x": [{"wildcard":"x*y"}]}`)
+	if err != nil {
+		t.Error("AP: " + err.Error())
+	}
+	err = cm.addPattern("x", `{"x": [3]}`)
+	if err != nil {
+		t.Error("AP: " + err.Error())
+	}
+	events := []string{`{"x": 3}`, `{"x": "xasdfy"}`}
+	for _, e := range events {
+		matches, err := cm.matchesForJSONEvent([]byte(e))
+		if err != nil {
+			t.Error("M4: " + err.Error())
+		}
+		if len(matches) != 1 || matches[0] != "x" {
+			t.Error("Match failed on " + e)
+		}
+	}
+}