[checkapi] use a YAML schema (#1148)

* [checkapi] use a YAML schema

* code review

---------

Co-authored-by: Paulo Janotti <[email protected]>

Author: atoulme

.chloggen/yaml_output.yaml

@@ -0,0 +1,16 @@
+# One of 'breaking', 'deprecation', 'new_component', 'enhancement', 'bug_fix'
+change_type: enhancement
+
+# The name of the component, or a single word describing the area of concern, (e.g. crosslink)
+component: checkapi
+
+# A brief description of the change.  Surround your text with quotes ("") if it needs to start with a backtick (`).
+note: Work with a JSON schema stored as yaml under metadata.yaml config key.
+
+# One or more tracking issues related to the change
+issues: [1148]
+
+# (Optional) One or more lines of additional information to render under the primary note.
+# These lines will be padded with 2 spaces and then inserted directly into the document.
+# Use pipe (|) for multiline entries.
+subtext:

checkapi/internal/config/receiver/badconfigreceiver/config.schema.json

@@ -1,4 +1,34 @@
 type: config
 
 status:
-  class: receiver
+  class: receiver
+
+config:
+  properties:
+    bar:
+      type: string
+    bool:
+      type: boolean
+    data_defined_elsewhere:
+      $ref: http://example.com/schemas/external_schema.json#/properties/data
+    foo:
+      type: string
+    my_special_embedded_field:
+      type: string
+    ptrStruct:
+      type: object
+    subconfig:
+      allOf:
+        - $ref: http://example.com/schemas/external_schema.json#/properties/data
+      properties:
+        bar:
+          type: boolean
+        foo:
+          type: string
+        sub_config2:
+          properties:
+            foobar:
+              type: string
+          type: object
+      type: object
+  type: object

checkapi/internal/config/receiver/configcallreceiver/config.schema.json

@@ -6,25 +6,12 @@ package internal
 import (
 	"errors"
 	"fmt"
-	"os"
-	"path"
 	"slices"
 	"strings"
 
 	"github.com/kaptinlin/jsonschema"
 )
 
-// ReadJSONSchema returns a JSON schema from the config.schema.json file under the folder
-func ReadJSONSchema(folder string) (*jsonschema.Schema, error) {
-	schemaFile := path.Join(folder, "config.schema.json")
-	b, err := os.ReadFile(schemaFile) //nolint: gosec
-	if err != nil {
-		return nil, err
-	}
-	compiler := jsonschema.NewCompiler()
-	return compiler.Compile(b)
-}
-
 // DeriveSchema interprets the config struct to return a valid JSON schema
 func DeriveSchema(cfgStruct APIstruct, structs []APIstruct, typeMappings map[string]string) (*jsonschema.Schema, error) {
 	return createObject(parseObject(cfgStruct, structs, typeMappings)), nil

checkapi/internal/config/receiver/configreceiver/config.schema.json

@@ -11,29 +11,29 @@ import (
 	"gopkg.in/yaml.v3"
 )
 
-type status struct {
+// Status of the component
+type Status struct {
 	Class string `yaml:"class"`
 }
 
-type metadata struct {
-	Status status `yaml:"status"`
+// Metadata of the component
+type Metadata struct {
+	Status Status `yaml:"status"`
+	Config any    `yaml:"config"`
 }
 
-// ReadComponentType reads the component type from the metadata.yaml file in the given folder.
-func ReadComponentType(folder string) (string, error) {
-	var componentType string
+// ReadMetadata reads from the metadata.yaml file in the given folder.
+func ReadMetadata(folder string) (Metadata, error) {
 	if _, err := os.Stat(filepath.Join(folder, "metadata.yaml")); errors.Is(err, os.ErrNotExist) {
-		componentType = "pkg"
-	} else {
-		m, err := os.ReadFile(filepath.Join(folder, "metadata.yaml")) // #nosec G304
-		if err != nil {
-			return "", err
-		}
-		var componentInfo metadata
-		if err = yaml.Unmarshal(m, &componentInfo); err != nil {
-			return "", err
-		}
-		componentType = componentInfo.Status.Class
+		return Metadata{Status: Status{Class: "pkg"}}, nil
 	}
-	return componentType, nil
+	m, err := os.ReadFile(filepath.Join(folder, "metadata.yaml")) // #nosec G304
+	if err != nil {
+		return Metadata{}, err
+	}
+	var componentInfo Metadata
+	if err = yaml.Unmarshal(m, &componentInfo); err != nil {
+		return Metadata{}, err
+	}
+	return componentInfo, nil
 }

checkapi/internal/config/receiver/configreceiver/metadata.yaml

@@ -18,6 +18,7 @@ import (
 	"strings"
 	"unicode"
 
+	"github.com/goccy/go-json"
 	"github.com/kaptinlin/jsonschema"
 
 	"go.opentelemetry.io/build-tools/checkapi/internal"
@@ -63,11 +64,11 @@ func run(folder string, configPath string) error {
 					return nil
 				}
 			}
-			componentType, err3 := internal.ReadComponentType(base)
+			metadata, err3 := internal.ReadMetadata(base)
 			if err3 != nil {
 				return err3
 			}
-			if err = walkFolder(cfg, base, componentType); err != nil {
+			if err = walkFolder(cfg, base, metadata); err != nil {
 				errs = append(errs, err)
 			}
 		}
@@ -82,7 +83,7 @@ func run(folder string, configPath string) error {
 	return nil
 }
 
-func walkFolder(cfg internal.Config, folder string, componentType string) error {
+func walkFolder(cfg internal.Config, folder string, metadata internal.Metadata) error {
 	result, err := internal.Read(folder, cfg.IgnoredFunctions, cfg.ExcludedFiles)
 	if err != nil {
 		return err
@@ -111,7 +112,7 @@ func walkFolder(cfg internal.Config, folder string, componentType string) error
 		functionsPresent := map[string]struct{}{}
 	OUTER:
 		for _, fnDesc := range cfg.AllowedFunctions {
-			if !slices.Contains(fnDesc.Classes, componentType) {
+			if !slices.Contains(fnDesc.Classes, metadata.Status.Class) {
 				continue
 			}
 			for _, fn := range result.Functions {
@@ -131,12 +132,13 @@ func walkFolder(cfg internal.Config, folder string, componentType string) error
 
 	if cfg.UnkeyedLiteral.Enabled {
 		for _, s := range result.Structs {
-			if err := checkStructDisallowUnkeyedLiteral(cfg, s, folder); err != nil {
+			if err = checkStructDisallowUnkeyedLiteral(cfg, s, folder); err != nil {
 				errs = append(errs, err)
 			}
 		}
 	}
 
+	componentType := metadata.Status.Class
 	if (!cfg.JSONSchema.CheckPresent && !cfg.JSONSchema.CheckValid && !cfg.ComponentAPI && !cfg.ComponentAPIStrict) || (componentType != "connector" && componentType != "exporter" && componentType != "extension" && componentType != "processor" && componentType != "receiver") {
 		return errors.Join(errs...)
 	}
@@ -165,24 +167,31 @@ func walkFolder(cfg internal.Config, folder string, componentType string) error
 		return errors.Join(errs...)
 	}
 
-	if _, err := os.Stat(filepath.Join(folder, "config.schema.json")); errors.Is(err, os.ErrNotExist) {
+	if metadata.Config == nil {
 		if cfg.JSONSchema.CheckPresent {
 			errs = append(errs, err)
 		}
 	} else {
-		jsonSchema, err := internal.ReadJSONSchema(folder)
+		configSchemaBytes, err := json.Marshal(metadata.Config)
 		if err != nil {
 			errs = append(errs, err)
 		} else {
-			var after *jsonschema.Schema
-			if after, err = internal.DeriveSchema(*cfgStruct, result.Structs, cfg.JSONSchema.TypeMappings); err != nil {
+			configSchema, err := jsonschema.NewCompiler().Compile(configSchemaBytes)
+			if err != nil {
 				errs = append(errs, err)
-			} else if err := internal.CompareJSONSchema(folder, jsonSchema, after); err != nil {
-				errs = append(errs, err)
-				b, _ := after.MarshalJSON()
-				errs = append(errs, fmt.Errorf("[%s] new JSON schema: %s", folder, string(b)))
+			} else {
+				var structDerivedSchema *jsonschema.Schema
+				if structDerivedSchema, err = internal.DeriveSchema(*cfgStruct, result.Structs, cfg.JSONSchema.TypeMappings); err != nil {
+					errs = append(errs, err)
+				} else if err := internal.CompareJSONSchema(folder, configSchema, structDerivedSchema); err != nil {
+					errs = append(errs, err)
+					configSchemaBytes, _ := structDerivedSchema.MarshalJSON()
+					rawJSON := map[string]any{}
+					_ = json.Unmarshal(configSchemaBytes, &rawJSON)
+					configSchemaYAML, _ := yaml.Marshal(rawJSON)
+					errs = append(errs, fmt.Errorf("[%s] new JSON schema: %s", folder, string(configSchemaYAML)))
+				}
 			}
-
 		}
 	}