feat: refactor some shit

parser/blocks.go

@@ -1,51 +1,32 @@
 package parser
 
-type BlockType string
+import (
+	"fmt"
 
-const (
-	BlockData     BlockType = "data"     // content between lines "{" and "}"
-	BlockMatching BlockType = "matching" // everything outside data blocks
+	"git.max-richter.dev/max/marka/parser/blocks"
 )
 
-type Block struct {
-	Type       BlockType
-	Start, End int // byte offsets [Start, End)
-	src        *string
-}
-
-func (b Block) GetContent() string {
-	if b.src == nil || b.Start < 0 || b.End > len(*b.src) || b.Start > b.End {
-		return ""
-	}
-	return (*b.src)[b.Start:b.End]
-}
-
 // ExtractBlocks scans once, emitting:
 // - data blocks: inner content between a line that's exactly "{" and a line that's exactly "}"
 // - matching blocks: gaps between data blocks (excluding the brace lines themselves)
-func ExtractBlocks(src string) []Block {
-	var out []Block
+func ExtractBlocks(template string) ([]blocks.TemplateBlock, error) {
+	var out []blocks.TemplateBlock
 	var curlyIndex int
 
 	const CLOSING = '}'
 	const OPENING = '{'
 
-	if len(src) > 0 && src[0] == OPENING {
+	var start int
+	var blockType blocks.BlockType
+
+	if len(template) > 0 && template[0] == OPENING {
 		curlyIndex = 1
-		out = append(out, Block{
-			Start: 0,
-			Type:  BlockData,
-			src:   &src,
-		})
+		blockType = blocks.DataBlock
 	} else {
-		out = append(out, Block{
-			Start: 0,
-			Type:  BlockMatching,
-			src:   &src,
-		})
+		blockType = blocks.MatchingBlock
 	}
 
-	for i, r := range src {
+	for i, r := range template {
 
 		var nextCurlyIndex = curlyIndex
 
@@ -57,41 +38,42 @@ func ExtractBlocks(src string) []Block {
 		}
 
 		var nextChar rune = ' '
-		if i+1 < len(src) {
-			nextChar = rune(src[i+1])
+		if i+1 < len(template) {
+			nextChar = rune(template[i+1])
 		}
 
 		if curlyIndex == 0 && nextCurlyIndex == 1 {
-			out[len(out)-1].End = i
-			out = append(out, Block{
-				Start: i,
-				Type:  BlockData,
-				src:   &src,
-			})
+			block, err := blocks.ParseTemplateBlock(template[start:i], blockType)
+			if err != nil {
+				return nil, fmt.Errorf("Failed to parse block: %w", err)
+			}
+			out = append(out, block)
+			start = i
+			blockType = blocks.DataBlock
 		} else if curlyIndex == 1 && nextCurlyIndex == 0 {
-			out[len(out)-1].End = i + 1
+
+			block, err := blocks.ParseTemplateBlock(template[start:i+1], blockType)
+			if err != nil {
+				return nil, fmt.Errorf("Failed to parse block: %w", err)
+			}
+			out = append(out, block)
+
 			if nextChar == OPENING {
-				out = append(out, Block{
-					Start: i + 1,
-					Type:  BlockData,
-					src:   &src,
-				})
+				start = i + 1
+				blockType = blocks.DataBlock
 			} else {
-				out = append(out, Block{
-					Start: i + 1,
-					Type:  BlockMatching,
-					src:   &src,
-				})
+				start = i + 1
+				blockType = blocks.MatchingBlock
 			}
 		}
 
 		curlyIndex = nextCurlyIndex
 	}
 
-	var lastBlock = out[len(out)-1]
-	if lastBlock.End == 0 {
-		out = out[:len(out)-1]
-	}
+	// var lastBlock = out[len(out)-1]
+	// if lastBlock.End == 0 {
+	// 	out = out[:len(out)-1]
+	// }
 
-	return out
+	return out, nil
 }

parser/blocks/blocks.go

@@ -0,0 +1,98 @@
+package blocks
+
+import (
+	"fmt"
+	"strings"
+)
+
+// TemplateType represents whether a template is short, long, or invalid.
+type TemplateType int
+
+const (
+	InvalidTemplate TemplateType = iota
+	ShortTemplate
+	LongTemplate
+)
+
+// DetectTemplateType checks if the template is short or long.
+func DetectTemplateType(tmpl string) TemplateType {
+	trimmed := strings.TrimSpace(tmpl)
+
+	// Short type: starts with "{" and ends with "}" on a single line,
+	// and contains "|" or "," inside for inline definition
+	// Matchs for example { name | text,required }
+	if strings.HasPrefix(trimmed, "{") &&
+		strings.HasSuffix(trimmed, "}") &&
+		!strings.Contains(trimmed, "\n") {
+		return ShortTemplate
+	}
+
+	// Long type: multiline and contains keys like "path:" or "codec:" inside
+	// Matches for example:
+	// {
+	//   path: name
+	//   codec: text
+	//   required: true
+	// }
+	if strings.Contains(trimmed, "\n") &&
+		(strings.Contains(trimmed, "path:") || strings.Contains(trimmed, "codec:")) {
+		return LongTemplate
+	}
+
+	return InvalidTemplate
+}
+
+// CodecType represents the type of codec used to encode/render a value
+type CodecType string
+
+const (
+	CodecText   CodecType = "text"
+	CodecNumber CodecType = "number"
+	CodecYaml   CodecType = "yaml"
+	CodecList   CodecType = "list"
+)
+
+func parseCodecType(input string) (CodecType, error) {
+	switch input {
+	case "number":
+		return CodecNumber, nil
+	case "yaml":
+		return CodecYaml, nil
+	case "list":
+		return CodecList, nil
+	case "text":
+		return CodecText, nil
+	}
+	return CodecText, fmt.Errorf("unknown codec: '%s'", input)
+}
+
+type BlockType string
+
+const (
+	DataBlock     BlockType = "data"     // content between lines "{" and "}"
+	MatchingBlock BlockType = "matching" // everything outside data blocks
+)
+
+type TemplateBlock struct {
+	Type     BlockType
+	Path     string
+	Codec    CodecType
+	Required bool
+	content  string
+}
+
+func (b TemplateBlock) GetContent() string {
+	return b.content
+}
+
+func (p *TemplateBlock) Parse(input string) (key string, value any, err error) {
+	switch p.Codec {
+	case CodecText:
+		return p.Path, input, nil
+	case CodecYaml:
+		return p.ParseYamlBlock(input)
+	case CodecList:
+		return p.ParseListBlock(input)
+	}
+	return p.Path, "", nil
+}

parser/blocks/list_block.go

@@ -0,0 +1,10 @@
+package blocks
+
+import "fmt"
+
+func (b TemplateBlock) ParseListBlock(input string) (key string, value any, error error) {
+
+	fmt.Printf("Parsing List: '%q'", input)
+
+	return "", nil, nil
+}

parser/blocks/template.go

@@ -0,0 +1,118 @@
+package blocks
+
+import (
+	"fmt"
+	"strings"
+
+	"go.yaml.in/yaml/v4"
+)
+
+func cleanTemplate(input string) string {
+	s := strings.TrimSpace(input)
+	s = strings.TrimPrefix(s, "{")
+	s = strings.TrimSuffix(s, "}")
+	return s
+}
+
+func parseShortTemplate(input string) (TemplateBlock, error) {
+
+	var split = strings.Split(cleanTemplate(input), "|")
+	if len(split) < 1 {
+		return TemplateBlock{}, fmt.Errorf("Invalid Short Template")
+	}
+
+	block := TemplateBlock{
+		Type:    DataBlock,
+		Path:    strings.TrimSpace(split[0]),
+		Codec:   CodecText,
+		content: input,
+	}
+
+	if len(split) > 1 {
+		var optionSplit = strings.Split(split[1], ",")
+		for _, option := range optionSplit {
+			switch strings.TrimSpace(option) {
+			case "required":
+				block.Required = true
+			case "number":
+				block.Codec = CodecNumber
+			}
+		}
+	}
+
+	return block, nil
+}
+
+type yamlBlock struct {
+	Path     string      `yaml:"path"`
+	Codec    string      `yaml:"codec"`
+	Required bool        `yaml:"required,omitempty"`
+	Fields   []yamlField `yaml:"fields"`
+	Item     *struct {
+		Template string `yaml:"template,omitempty"`
+	} `yaml:"item,omitempty"`
+	Template string `yaml:"template,omitempty"`
+}
+
+type yamlField struct {
+	Path     string `yaml:"path"`
+	Value    any    `yaml:"value,omitempty"`
+	Codec    string `yaml:"codec"`
+	Required bool   `yaml:"required"`
+}
+
+func parseYamlTemplate(input string) (block TemplateBlock, err error) {
+
+	var blk yamlBlock
+
+	cleaned := cleanTemplate(input)
+
+	dec := yaml.NewDecoder(strings.NewReader(cleaned))
+	dec.KnownFields(true)
+
+	if err := dec.Decode(&blk); err != nil {
+		fmt.Printf("Failed to parse:\n---\n%s\n---\n", cleaned)
+		return block, err
+	}
+
+	if blk.Path == "" {
+		return block, fmt.Errorf("missing top-level 'path'")
+	}
+
+	codec, err := parseCodecType(blk.Codec)
+	if err != nil {
+		return block, fmt.Errorf("failed to parse codec: %w", err)
+	}
+
+	return TemplateBlock{
+		Type:    DataBlock,
+		Path:    blk.Path,
+		Codec:   codec,
+		content: input,
+	}, nil
+
+}
+
+func ParseTemplateBlock(template string, blockType BlockType) (block TemplateBlock, err error) {
+
+	if blockType == MatchingBlock {
+		return TemplateBlock{
+			Type:    MatchingBlock,
+			content: template,
+		}, nil
+	}
+
+	block.Type = DataBlock
+	block.content = template
+
+	templateType := DetectTemplateType(template)
+	if templateType == InvalidTemplate {
+		return block, fmt.Errorf("Invalid Template")
+	}
+
+	if templateType == ShortTemplate {
+		return parseShortTemplate(template)
+	}
+
+	return parseYamlTemplate(template)
+}

parser/blocks/yaml_block.go

@@ -0,0 +1,18 @@
+package blocks
+
+import (
+	"fmt"
+
+	"go.yaml.in/yaml/v4"
+)
+
+func (b TemplateBlock) ParseYamlBlock(input string) (key string, value any, error error) {
+
+	res := make(map[string]any)
+	err := yaml.Unmarshal([]byte(input), &res)
+	if err != nil {
+		return "", nil, fmt.Errorf("failed to parse yaml: %w", err)
+	}
+
+	return "", nil, nil
+}

parser/blocks_test.go

@@ -1,50 +1,54 @@
-package parser
+package parser_test
 
 import (
-	"os"
-	"path/filepath"
-	"strings"
 	"testing"
+
+	"git.max-richter.dev/max/marka/parser"
+	"git.max-richter.dev/max/marka/parser/blocks"
+	"git.max-richter.dev/max/marka/registry"
 )
 
-func readFile(t *testing.T, fileName string) string {
-	path := filepath.Join("testdata", fileName)
-	data, err := os.ReadFile(path)
-	if err != nil {
-		t.Fatalf("failed to read test data file: %v", err)
-	}
-	return string(data)
-}
-
 func TestExtractBlocks(t *testing.T) {
-	src := readFile(t, "recipe.schema.md")
-	blocks := ExtractBlocks(src)
+	src, err := registry.GetTemplate("recipe")
+	if err != nil {
+		t.Errorf("Failed to extract blocks: %s", err.Error())
+		t.FailNow()
+	}
+
+	templateBlocks, err := parser.ExtractBlocks(src)
+	if err != nil {
+		t.Errorf("Failed to extract blocks: %s", err.Error())
+		t.FailNow()
+	}
 
 	expected := []struct {
-		Type    BlockType
+		Type    blocks.BlockType
 		Content string
 	}{
-		{BlockMatching, "---\\n"},
-		{BlockData, "{ . }"},
-		{BlockMatching, "\\n---\\n\\n# "},
-		{BlockData, "{ name | text,required }"},
-		{BlockMatching, "\\n\\n"},
-		{BlockData, "{ description | text,optional }"},
-		{BlockMatching, "\\n\\n## Ingredients\\n"},
-		{BlockData, "{\\n  path: recipeIngredient\\n  codec: list\\n  required: true\\n  item:\\n    template: \"- { . }\"\\n}"},
-		{BlockMatching, "\\n\\n## Steps\\n"},
-		{BlockData, "{\\n  path: recipeInstructions\\n  codec: list\\n  required: true\\n  item:\\n    template: \"{ @index }. { . }\"\\n}"},
+		{blocks.MatchingBlock, "---\n"},
+		{blocks.DataBlock, "{\n  path: .\n  codec: yaml\n  fields:\n    - path: name\n      codec: text\n      required: true\n    - path: image\n      codec: text\n      required: true\n    - path: author.@type\n      codec: const\n      value: Person\n    - path: author.name\n      codec: text\n    - path: datePublished\n      codec: text\n    - path: description\n      codec: text\n    - path: prepTime\n      codec: text\n    - path: cookTime\n      codec: text\n    - path: recipeYield\n      codec: text\n}"},
+		{blocks.MatchingBlock, "\n---\n\n# "},
+		{blocks.DataBlock, "{ name | text,required }"},
+		{blocks.MatchingBlock, "\n\n"},
+		{blocks.DataBlock, "{ description | text }"},
+		{blocks.MatchingBlock, "\n\n## Ingredients\n"},
+		{blocks.DataBlock, "{\n  path: recipeIngredient\n  codec: list\n  required: true\n  item:\n    template: \"- { . }\"\n}"},
+		{blocks.MatchingBlock, "\n\n## Steps\n"},
+		{blocks.DataBlock, "{\n  path: recipeInstructions\n  codec: list\n  required: true\n  item:\n    template: \"{ @index }. { . }\"\n}"},
 	}
 
-	if len(blocks) != len(expected) {
-		t.Fatalf("expected %d blocks, got %d", len(expected), len(blocks))
+	if len(templateBlocks) != len(expected) {
+		t.Fatalf("expected %d blocks, got %d", len(expected), len(templateBlocks))
 	}
 
-	for i, b := range blocks {
+	for i, b := range templateBlocks {
 		exp := expected[i]
-		content := strings.ReplaceAll(b.GetContent(), "\n", "\\n")
-		if b.Type != exp.Type || content != exp.Content {
-			t.Errorf("Block %d: expected %v, got Type: %v, Start: %d, End: %d, Content: %s", i, exp, b.Type, b.Start, b.End, content)
+		if b.Type != exp.Type {
+			t.Errorf("Block#%d Type '%s' did not match expected type '%s'", i, b.Type, exp.Type)
+		}
+		content := b.GetContent()
+		if content != exp.Content {
+			t.Errorf("Block#%d Content '%s' did not match expected Content: '%s'", i, content, exp.Content)
 		}
 	}
 

parser/go.mod

@@ -3,3 +3,5 @@ module git.max-richter.dev/max/marka/parser
 go 1.24.3
 
 require github.com/agext/levenshtein v1.2.3
+
+require go.yaml.in/yaml/v4 v4.0.0-rc.1 // indirect

parser/go.sum

@@ -1,2 +1,4 @@
 github.com/agext/levenshtein v1.2.3 h1:YB2fHEn0UJagG8T1rrWknE3ZQzWM06O8AMAatNn7lmo=
 github.com/agext/levenshtein v1.2.3/go.mod h1:JEDfjyjHDjOF/1e4FlBE/PkbqA9OfWu2ki2W0IB5558=
+go.yaml.in/yaml/v4 v4.0.0-rc.1 h1:4J1+yLKUIPGexM/Si+9d3pij4hdc7aGO04NhrElqXbY=
+go.yaml.in/yaml/v4 v4.0.0-rc.1/go.mod h1:CBdeces52/nUXndfQ5OY8GEQuNR9uEEOJPZj/Xq5IzU=

parser/main.go

@@ -0,0 +1,17 @@
+// Package parser provides functions for parsing Markdown templates into
+// structured JSON objects that conform to a JSON Schema.
+package parser
+
+func ParseFile(markdownContent string) (map[string]any, error) {
+
+	// _schema, err := registry.GetTemplate("Recipe")
+	// if err != nil {
+	// 	return nil, fmt.Errorf("could not get schema: %w", err)
+	// }
+
+	// Idea is to split the template into blocks, either "matching" blocks which are simple strings.
+	// Or "data" blocks which match the content. Then i want to soft match the "matching" blocks and "data" blocks to the template.
+	// The "matching" blocks should soft match with a levenshtein distance
+
+	return map[string]any{}, nil
+}

parser/main_test.go

@@ -0,0 +1,43 @@
+package parser_test
+
+import (
+	"encoding/json"
+	"os"
+	"path/filepath"
+	"reflect"
+	"testing"
+
+	"git.max-richter.dev/max/marka/parser"
+)
+
+func TestParseRecipe_Golden(t *testing.T) {
+	td := filepath.Join("testdata", "recipe_salad")
+	input := filepath.Join(td, "input.md")
+	output := filepath.Join(td, "output.json")
+
+	inputContent, err := os.ReadFile(input)
+	if err != nil {
+		t.Fatalf("read input.md: %v", err)
+	}
+
+	got, err := parser.ParseFile(string(inputContent))
+	if err != nil {
+		t.Fatalf("ParseFile: %v", err)
+	}
+
+	var want map[string]any
+	b, err := os.ReadFile(output)
+	if err != nil {
+		t.Fatalf("read expected.json: %v", err)
+	}
+	if err := json.Unmarshal(b, &want); err != nil {
+		t.Fatalf("unmarshal expected.json: %v", err)
+	}
+
+	// Deep structural compare
+	if !reflect.DeepEqual(want, got) {
+		gb, _ := json.MarshalIndent(got, "", "  ")
+		wb, _ := json.MarshalIndent(want, "", "  ")
+		t.Fatalf("parsed JSON mismatch\n--- got ---\n%s\n--- want ---\n%s", string(gb), string(wb))
+	}
+}

parser/matcher.go

@@ -3,12 +3,13 @@ package parser
 import (
 	"math"
 
+	"git.max-richter.dev/max/marka/parser/blocks"
 	"github.com/agext/levenshtein"
 )
 
 type MatchBlock struct {
 	Start, End int
-	Block      Block
+	Block      blocks.TemplateBlock
 	src        *string
 }
 
@@ -22,17 +23,17 @@ func (m MatchBlock) GetContent() string {
 // MatchBlocksFuzzy finds anchor positions for all BlockMatching blocks using
 // Levenshtein distance (tolerant matching), then returns ONLY the BlockData
 // segments as gaps between those anchors.
-func MatchBlocksFuzzy(markdown string, blocks []Block, maxDist float64) []MatchBlock {
+func MatchBlocksFuzzy(markdown string, templateBlocks []blocks.TemplateBlock, maxDist float64) []MatchBlock {
 	var out []MatchBlock
 
 	var lastIndex = 0
-	for i, b := range blocks {
-		if b.Type == BlockMatching {
+	for i, b := range templateBlocks {
+		if b.Type == blocks.MatchingBlock {
 			start, end := FuzzyFind(markdown, lastIndex, b.GetContent(), 0.3)
 			if end != -1 {
 				if i > 0 {
-					previousBlock := blocks[i-1]
-					if previousBlock.Type == BlockData {
+					previousBlock := templateBlocks[i-1]
+					if previousBlock.Type == blocks.DataBlock {
 						out = append(out, MatchBlock{
 							Start: lastIndex,
 							End:   start,
@@ -47,8 +48,8 @@ func MatchBlocksFuzzy(markdown string, blocks []Block, maxDist float64) []MatchB
 	}
 
 	// Handle the last block
-	lastBlock := blocks[len(blocks)-1]
-	if lastBlock.Type == BlockData {
+	lastBlock := templateBlocks[len(templateBlocks)-1]
+	if lastBlock.Type == blocks.DataBlock {
 		out = append(out, MatchBlock{
 			Start: lastIndex,
 			End:   len(markdown),

parser/matcher_test.go

@@ -1,37 +1,27 @@
 package parser_test
 
 import (
-	"os"
-	"path/filepath"
 	"testing"
 
 	"git.max-richter.dev/max/marka/parser"
+	"git.max-richter.dev/max/marka/registry"
 )
 
-func readFile(t *testing.T, fileName string) string {
-	path := filepath.Join("testdata", fileName)
-	data, err := os.ReadFile(path)
-	if err != nil {
-		t.Fatalf("failed to read test data file: %v", err)
-	}
-	return string(data)
-}
-
 func TestFuzzyFindAll(t *testing.T) {
-	recipeMd := readFile(t, "baguette.md")
+	recipeMd := readTestDataFile(t, "baguette.md")
 
 	tests := []struct {
 		Needle                 string
 		Start, End, StartIndex int
 	}{
-		{StartIndex: 0, Needle: "# Ingredients\n", Start: 72, End: 86},
-		{StartIndex: 0, Needle: "# Ingrdients\n", Start: 72, End: 86},
-		{StartIndex: 0, Needle: "# Inrdients\n", Start: 72, End: 86},
+		{StartIndex: 0, Needle: "# Ingredients\n", Start: 77, End: 91},
+		{StartIndex: 0, Needle: "# Ingrdients\n", Start: 77, End: 91},
+		{StartIndex: 0, Needle: "# Inrdients\n", Start: 77, End: 91},
 		{StartIndex: 0, Needle: "---\n", Start: 0, End: 4},
-		{StartIndex: 4, Needle: "---\n", Start: 24, End: 28},
-		{StartIndex: 0, Needle: "# Steps\n", Start: 111, End: 119},
-		{StartIndex: 0, Needle: "# Stps\n", Start: 111, End: 119},
-		{StartIndex: 0, Needle: "# Step\n", Start: 111, End: 119},
+		{StartIndex: 4, Needle: "---\n", Start: 29, End: 33},
+		{StartIndex: 0, Needle: "# Steps\n", Start: 116, End: 124},
+		{StartIndex: 0, Needle: "# Stps\n", Start: 116, End: 124},
+		{StartIndex: 0, Needle: "# Step\n", Start: 116, End: 124},
 	}
 
 	for _, test := range tests {
@@ -45,16 +35,20 @@ func TestFuzzyFindAll(t *testing.T) {
 }
 
 func TestFuzzyBlockMatch(t *testing.T) {
-	recipeMd := readFile(t, "baguette.md")
-	schemaMd := readFile(t, "recipe.schema.md")
-	blocks := parser.ExtractBlocks(schemaMd)
+	recipeMd := readTestDataFile(t, "baguette.md")
+	schemaMd, err := registry.GetTemplate("recipe")
+	if err != nil {
+		t.Errorf("Failed to load template: %s", err.Error())
+		t.FailNow()
+	}
+	blocks, _ := parser.ExtractBlocks(schemaMd)
 	matches := parser.MatchBlocksFuzzy(recipeMd, blocks, 0.3)
 
 	expected := []struct {
 		value string
 	}{
 		{
-			value: "author: Max Richter",
+			value: "author.name: Max Richter",
 		},
 		{
 			value: "Baguette",
@@ -66,7 +60,7 @@ func TestFuzzyBlockMatch(t *testing.T) {
 			value: "- Flour\n- Water\n- Salt",
 		},
 		{
-			value: "1. Mix Flour Water and Salt\n2. Bake the bread",
+			value: "1. Mix Flour Water and Salt\n2. Bake the bread\n",
 		},
 	}
 

parser/parser.go

@@ -1,17 +1,15 @@
-// Package parser provides functions for parsing Markdown templates into
-// structured JSON objects that conform to a JSON Schema.
 package parser
 
-func ParseFile(markdownContent string) (map[string]any, error) {
+func Parse(blocks []MatchBlock) map[string]any {
 
-	// _schema, err := registry.GetTemplate("Recipe")
-	// if err != nil {
-	// 	return nil, fmt.Errorf("could not get schema: %w", err)
-	// }
+	result := make(map[string]any)
 
-	// Idea is to split the template into blocks, either "matching" blocks which are simple strings.
-	// Or "data" blocks which match the content. Then i want to soft match the "matching" blocks and "data" blocks to the template.
-	// The "matching" blocks should soft match with a levenshtein distance
+	for _, b := range blocks {
+		input := b.GetContent()
 
-	return map[string]any{}, nil
+		key, value, _ := b.Block.Parse(input)
+		result[key] = value
+	}
+
+	return result
 }

parser/parser_test.go

@@ -1,43 +1,42 @@
 package parser_test
 
 import (
-	"encoding/json"
-	"os"
-	"path/filepath"
-	"reflect"
+	"fmt"
 	"testing"
 
 	"git.max-richter.dev/max/marka/parser"
+	"git.max-richter.dev/max/marka/registry"
 )
 
-func TestParseRecipe_Golden(t *testing.T) {
-	td := filepath.Join("testdata", "recipe_salad")
-	input := filepath.Join(td, "input.md")
-	output := filepath.Join(td, "output.json")
+func TestParseBaguette(t *testing.T) {
+	recipeMd := readTestDataFile(t, "baguette.md")
 
-	inputContent, err := os.ReadFile(input)
+	template, err := registry.GetTemplate("recipe")
 	if err != nil {
-		t.Fatalf("read input.md: %v", err)
+		t.Fatalf("Err: %s", err)
 	}
 
-	got, err := parser.ParseFile(string(inputContent))
+	blocks, err := parser.ExtractBlocks(template)
 	if err != nil {
-		t.Fatalf("ParseFile: %v", err)
+		t.Fatalf("Err: %s", err)
 	}
 
-	var want map[string]any
-	b, err := os.ReadFile(output)
-	if err != nil {
-		t.Fatalf("read expected.json: %v", err)
-	}
-	if err := json.Unmarshal(b, &want); err != nil {
-		t.Fatalf("unmarshal expected.json: %v", err)
+	matches := parser.MatchBlocksFuzzy(recipeMd, blocks, 0.3)
+	parsed := parser.Parse(matches)
+	expected := map[string]any{
+		"name":             "Baguette",
+		"description":      "My favourite baguette recipe",
+		"recipeIngredient": []string{"Flour", "Water", "Salt"},
+		// "recipeInstructions": []string{
+		// 	"Mix Flour Water and Salt",
+		// 	"Bake the bread",
+		// },
 	}
 
-	// Deep structural compare
-	if !reflect.DeepEqual(want, got) {
-		gb, _ := json.MarshalIndent(got, "", "  ")
-		wb, _ := json.MarshalIndent(want, "", "  ")
-		t.Fatalf("parsed JSON mismatch\n--- got ---\n%s\n--- want ---\n%s", string(gb), string(wb))
+	for k, v := range expected {
+		if fmt.Sprintf("%v", parsed[k]) != fmt.Sprintf("%v", v) {
+			t.Errorf("Expected %v but got %v", v, parsed[k])
+		}
 	}
+
 }

parser/testdata/baguette.md

@@ -1,5 +1,5 @@
 ---
-author: Max Richter
+author.name: Max Richter
 ---
 
 # Baguette
@@ -13,4 +13,4 @@ My favourite baguette recipe
 
 ## Steps
 1. Mix Flour Water and Salt
-2. Bake the bread
+2. Bake the bread

parser/testdata/recipe.schema.md

@@ -1,25 +0,0 @@
----
-{ . }
----
-
-# { name | text,required }
-
-{ description | text,optional }
-
-## Ingredients
-{
-  path: recipeIngredient
-  codec: list
-  required: true
-  item:
-    template: "- { . }"
-}
-
-## Steps
-{
-  path: recipeInstructions
-  codec: list
-  required: true
-  item:
-    template: "{ @index }. { . }"
-}

parser/testutils_test.go

@@ -0,0 +1,16 @@
+package parser_test
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+func readTestDataFile(t *testing.T, fileName string) string {
+	path := filepath.Join("testdata", fileName)
+	data, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatalf("failed to read test data file: %v", err)
+	}
+	return string(data)
+}

registry/aliases/default.yaml

@@ -1,6 +0,0 @@
-description: "Core capture aliases for Marka"
-patterns:
-  text:        ".+"
-  word:        "\\S+"
-  num:         "(?:\\d+(?:[.,]\\d+)?(?:\\s?\\d+/\\d+)?)"    # 3 | 1.5 | 1 1/2
-  indexMarker: "\\d+[.)]"                                   # 1. / 1)

registry/registry.go

@@ -28,9 +28,6 @@ var templates embed.FS
 //go:embed schema-org/*
 var schemas embed.FS
 
-//go:embed aliases/*
-var aliases embed.FS
-
 func GetTemplates() Source {
 	return src{fsys: templates}
 }
@@ -49,11 +46,3 @@ func GetTemplate(name string) (string, error) {
 
 	return string(templateBytes), nil
 }
-
-func GetSchemas() Source {
-	return src{fsys: schemas}
-}
-
-func GetAliases() Source {
-	return src{fsys: aliases}
-}

registry/templates/recipe.marka

@@ -2,10 +2,6 @@
 {
   path: .
   codec: yaml
-  required: true
-  assert:
-    "@context": https://schema.org/
-    "@type": Recipe
   fields:
     - path: name
       codec: text
@@ -18,28 +14,22 @@
       value: Person
     - path: author.name
       codec: text
-      required: true
     - path: datePublished
       codec: text
-      optional: true
     - path: description
       codec: text
-      optional: true
     - path: prepTime
       codec: text
-      optional: true
     - path: cookTime
       codec: text
-      optional: true
     - path: recipeYield
       codec: text
-      optional: true
 }
 ---
 
 # { name | text,required }
 
-{ description | text,optional }
+{ description | text }
 
 ## Ingredients
 {