From 573618e4cef8f25d0081fe93a93eb1f2d1f8dd41 Mon Sep 17 00:00:00 2001
From: Mike Farah <mikefarah@gmail.com>
Date: Wed, 7 Jul 2021 22:47:16 +1000
Subject: [PATCH] (wip) regex match op

---
 pkg/yqlib/doc/String Operators.md  | 39 ++-----------------------
 pkg/yqlib/operator_strings.go      | 23 +++++++++------
 pkg/yqlib/operator_strings_test.go | 46 ++++++++++++++++++------------
 3 files changed, 44 insertions(+), 64 deletions(-)

diff --git a/pkg/yqlib/doc/String Operators.md b/pkg/yqlib/doc/String Operators.md
index f6d30637..581b2149 100644
--- a/pkg/yqlib/doc/String Operators.md	
+++ b/pkg/yqlib/doc/String Operators.md	
@@ -1,47 +1,13 @@
 # String Operators
 
-## Match string
-Given a sample.yml file of:
-```yaml
-cat
-```
-then
-```bash
-yq eval 'match("at")' sample.yml
-```
-will output
-```yaml
-string: at
-offset: 1
-length: 2
-captures: []
-```
-
-## Match string, case insensitive
-Given a sample.yml file of:
-```yaml
-cAt
-```
-then
-```bash
-yq eval 'match("(?i)at")' sample.yml
-```
-will output
-```yaml
-string: At
-offset: 1
-length: 2
-captures: []
-```
-
-## Match with capture groups
+## Match with names capture groups
 Given a sample.yml file of:
 ```yaml
 a cat
 ```
 then
 ```bash
-yq eval 'match("c(.t)")' sample.yml
+yq eval 'match("c(?P<cool>.t)")' sample.yml
 ```
 will output
 ```yaml
@@ -52,5 +18,6 @@ captures:
   - string: at
     offset: 3
     length: 2
+    name: cool
 ```
 
diff --git a/pkg/yqlib/operator_strings.go b/pkg/yqlib/operator_strings.go
index e5b24cd1..911e207e 100644
--- a/pkg/yqlib/operator_strings.go
+++ b/pkg/yqlib/operator_strings.go
@@ -74,24 +74,29 @@ func substituteStringOperator(d *dataTreeNavigator, context Context, expressionN
 
 }
 
-func addMatch(original []*yaml.Node, match string, offset int) []*yaml.Node {
-	return append(original,
+func addMatch(original []*yaml.Node, match string, offset int, name string) []*yaml.Node {
+	newContent := append(original,
 		createScalarNode("string", "string"),
 		createScalarNode(match, match),
 		createScalarNode("offset", "offset"),
 		createScalarNode(offset, fmt.Sprintf("%v", offset)),
 		createScalarNode("length", "length"),
 		createScalarNode(len(match), fmt.Sprintf("%v", len(match))))
+
+	if name != "" {
+		newContent = append(newContent,
+			createScalarNode("name", "name"),
+			createScalarNode(name, name),
+		)
+	}
+	return newContent
 }
 
 func match(regEx *regexp.Regexp, candidate *CandidateNode, value string, results *list.List) {
-	// captures = FindAllStringSubmatch
-	// FindAllStringSubmatchIndex = offset?
 
-	//string array
-	// subNames := regEx.SubexpNames()
+	subNames := regEx.SubexpNames()
+	log.Debugf("subNames %v", subNames)
 
-	//array of arrays
 	allMatches := regEx.FindAllStringSubmatch(value, -1)
 	allIndices := regEx.FindAllStringSubmatchIndex(value, -1)
 
@@ -100,12 +105,12 @@ func match(regEx *regexp.Regexp, candidate *CandidateNode, value string, results
 		match, submatches := matches[0], matches[1:]
 		for j, submatch := range submatches {
 			captureNode := &yaml.Node{Kind: yaml.MappingNode}
-			captureNode.Content = addMatch(capturesNode.Content, submatch, allIndices[i][2+j*2])
+			captureNode.Content = addMatch(capturesNode.Content, submatch, allIndices[i][2+j*2], subNames[j+1])
 			capturesNode.Content = append(capturesNode.Content, captureNode)
 		}
 
 		node := &yaml.Node{Kind: yaml.MappingNode}
-		node.Content = addMatch(node.Content, match, allIndices[i][0])
+		node.Content = addMatch(node.Content, match, allIndices[i][0], "")
 		node.Content = append(node.Content,
 			createScalarNode("captures", "captures"),
 			capturesNode,
diff --git a/pkg/yqlib/operator_strings_test.go b/pkg/yqlib/operator_strings_test.go
index 3bd3b43b..655dc2ec 100644
--- a/pkg/yqlib/operator_strings_test.go
+++ b/pkg/yqlib/operator_strings_test.go
@@ -13,28 +13,36 @@ var stringsOperatorScenarios = []expressionScenario{
 	// 		"D0, P[], (!!str)::cat; meow; 1; ; true\n",
 	// 	},
 	// },
+	// {
+	// 	description: "Match string",
+	// 	document:    `cat`,
+	// 	expression:  `match("at")`,
+	// 	expected: []string{
+	// 		"D0, P[], ()::string: at\noffset: 1\nlength: 2\ncaptures: []\n",
+	// 	},
+	// },
+	// {
+	// 	description: "Match string, case insensitive",
+	// 	document:    `cAt`,
+	// 	expression:  `match("(?i)at")`,
+	// 	expected: []string{
+	// 		"D0, P[], ()::string: At\noffset: 1\nlength: 2\ncaptures: []\n",
+	// 	},
+	// },
+	// {
+	// 	description: "Match with capture groups",
+	// 	document:    `a cat`,
+	// 	expression:  `match("c(.t)")`,
+	// 	expected: []string{
+	// 		"D0, P[], ()::string: cat\noffset: 2\nlength: 3\ncaptures:\n    - string: at\n      offset: 3\n      length: 2\n",
+	// 	},
+	// },
 	{
-		description: "Match string",
-		document:    `cat`,
-		expression:  `match("at")`,
-		expected: []string{
-			"D0, P[], ()::string: at\noffset: 1\nlength: 2\ncaptures: []\n",
-		},
-	},
-	{
-		description: "Match string, case insensitive",
-		document:    `cAt`,
-		expression:  `match("(?i)at")`,
-		expected: []string{
-			"D0, P[], ()::string: At\noffset: 1\nlength: 2\ncaptures: []\n",
-		},
-	},
-	{
-		description: "Match with capture groups",
+		description: "Match with names capture groups",
 		document:    `a cat`,
-		expression:  `match("c(.t)")`,
+		expression:  `match("c(?P<cool>.t)")`,
 		expected: []string{
-			"D0, P[], ()::string: cat\noffset: 2\nlength: 3\ncaptures:\n    - string: at\n      offset: 3\n      length: 2\n",
+			"D0, P[], ()::string: cat\noffset: 2\nlength: 3\ncaptures:\n    - string: at\n      offset: 3\n      length: 2\n      name: cool\n",
 		},
 	},
 	// {