fix: avro parser now handles record reuse in definitions

package.json

@@ -6,7 +6,7 @@
   "scripts": {
     "test": "jest",
     "release": "semantic-release",
-    "lint": "eslint --max-warnings 0 --config .eslintrc.yaml .",
+    "lint": "eslint --max-warnings 1 --config .eslintrc.yaml .",
     "generate:assets": "echo 'No additional assets need to be generated at the moment'",
     "bump:version": "npm --no-git-tag-version --allow-same-version version $VERSION"
   },

tests/asyncapi-avro-111-1.9.0.yaml

@@ -0,0 +1,11 @@
+asyncapi: 2.0.0
+info:
+  title: My API
+  version: '1.0.0'
+channels:
+  mychannel:
+    publish:
+      message:
+        schemaFormat: application/vnd.apache.avro;version=1.9.0
+        payload:
+          $ref: 'schemas/issue-111-testcase.avsc'

tests/asyncapi-avro-113-1.9.0.yaml

@@ -0,0 +1,11 @@
+asyncapi: 2.0.0
+info:
+  title: My API
+  version: '1.0.0'
+channels:
+  mychannel:
+    publish:
+      message:
+        schemaFormat: application/vnd.apache.avro;version=1.9.0
+        payload:
+          $ref: 'schemas/issue-113-testcase.avsc'

tests/schemas/issue-111-testcase.avsc

@@ -0,0 +1,103 @@
+{
+  "type": "record",
+  "name": "ConnectionRequested",
+  "namespace": "com.foo.connections",
+  "doc": "An example schema to illustrate the issue",
+  "fields": [
+    {
+      "name": "metadata",
+      "type": {
+        "type": "record",
+        "name": "EventMetadata",
+        "namespace": "com.foo",
+        "doc": "Metadata to be associated with every published event",
+        "fields": [
+          {
+            "name": "id",
+            "type": {
+              "type": "string",
+              "logicalType": "uuid"
+            },
+            "doc": "Unique identifier for this specific event"
+          },
+          {
+            "name": "timestamp",
+            "type": {
+              "type": "long",
+              "logicalType": "timestamp-millis"
+            },
+            "doc": "Instant the event took place (not necessary when it was published)"
+          },
+          {
+            "name": "correlation_id",
+            "type": [
+              "null",
+              {
+                "type": "string",
+                "logicalType": "uuid"
+              }
+            ],
+            "doc": "id of the event that resulted in this\nevent being published (optional)",
+            "default": null
+          },
+          {
+            "name": "publisher_context",
+            "type": [
+              "null",
+              {
+                "type": "map",
+                "values": {
+                  "type": "string",
+                  "avro.java.string": "String"
+                },
+                "avro.java.string": "String"
+              }
+            ],
+            "doc": "optional set of key-value pairs of context to be echoed back\nin any resulting message (like a richer\ncorrelationId.\n\nThese values are likely only meaningful to the publisher\nof the correlated event",
+            "default": null
+          }
+        ]
+      }
+    },
+    {
+      "name": "auth_code",
+      "type": {
+        "type": "record",
+        "name": "EncryptedString",
+        "namespace": "com.foo",
+        "doc": "A string that was encrypted with AES (using CTR mode), its key encrypted with RSA, and the nonce used for the encryption.",
+        "fields": [
+          {
+            "name": "value",
+            "type": "string",
+            "doc": "A sequence of bytes that has been AES encrypted in CTR mode."
+          },
+          {
+            "name": "nonce",
+            "type": "string",
+            "doc": "A nonce, used by the CTR encryption mode for our encrypted value. Not encrypted, not a secret."
+          },
+          {
+            "name": "key",
+            "type": "string",
+            "doc": "An AES key, used to encrypt the value field, that has itself been encrypted using RSA."
+          }
+        ]
+      },
+      "doc": "Encrypted auth_code received when user authorizes the app."
+    },
+    {
+      "name": "refresh_token",
+      "type": "com.foo.EncryptedString",
+      "doc": "Encrypted refresh_token generated by using clientId and clientSecret."
+    },
+    {
+      "name": "triggered_by",
+      "type": {
+        "type": "string",
+        "logicalType": "uuid"
+      },
+      "doc": "ID of the user who triggered this event."
+    }
+  ]
+}

tests/schemas/issue-113-testcase.avsc

@@ -0,0 +1,36 @@
+[
+  {
+    "type": "record",
+    "name": "Address",
+    "namespace": "com.example",
+    "fields": [
+      {
+        "name": "streetaddress",
+        "type": "string"
+      },
+      {
+        "name": "city",
+        "type": "string"
+      }
+    ]
+  },
+  {
+    "type": "record",
+    "name": "Person",
+    "namespace": "com.example",
+    "fields": [
+      {
+        "name": "firstname",
+        "type": "string"
+      },
+      {
+        "name": "lastname",
+        "type": "string"
+      },
+      {
+        "name": "address",
+        "type": "com.example.Address"
+      }
+    ]
+  }
+]

to-json-schema.js

@@ -136,21 +136,39 @@ function validateAvroSchema(avroDefinition) {
   avsc.Type.forSchema(avroDefinition);
 }
 
-async function convertAvroToJsonSchema(avroDefinition, isTopLevel) {
+/**
+ * Cache the passed value under the given key. If the key is undefined the value will not be cached. This function
+ * uses mutation of the passed cache object rather than a copy on write cache strategy.
+ *
+ * @param cache Map<String, JsonSchema> the cache to store the JsonSchema
+ * @param key String | Undefined - the fully qualified name of an avro record
+ * @param value JsonSchema - The json schema from the avro record
+ */
+function cacheAvroRecordDef(cache, key, value) {
+  if (key) {
+    cache[key] = value;
+  }
+}
+
+async function convertAvroToJsonSchema(avroDefinition, isTopLevel, recordCache = {}) {
   const jsonSchema = {};
   const isUnion = Array.isArray(avroDefinition);
 
-  validateAvroSchema(avroDefinition);
-
   if (isUnion) {
     jsonSchema.oneOf = [];
     let nullDef = null;
     for (const avroDef of avroDefinition) {
-      const def = await convertAvroToJsonSchema(avroDef, isTopLevel);
+      const def = await convertAvroToJsonSchema(avroDef, isTopLevel, recordCache);
       // avroDef can be { type: 'int', default: 1 } and this is why avroDef.type has priority here
       const defType = avroDef.type || avroDef;
       // To prefer non-null values in the examples skip null definition here and push it as the last element after loop
-      if (defType === 'null') nullDef = def; else jsonSchema.oneOf.push(def);
+      if (defType === 'null') {
+        nullDef = def;
+      } else {
+        jsonSchema.oneOf.push(def);
+        const qualifiedName = getFullyQualifiedName(avroDef);
+        cacheAvroRecordDef(recordCache, qualifiedName, def);
+      }
     }
     if (nullDef) jsonSchema.oneOf.push(nullDef);
 
@@ -195,13 +213,21 @@ async function convertAvroToJsonSchema(avroDefinition, isTopLevel) {
   case 'record':
     const propsMap = new Map();
     for (const field of avroDefinition.fields) {
-      const def = await convertAvroToJsonSchema(field.type, false);
-
-      requiredAttributesMapping(field, jsonSchema, field.default !== undefined);
-      commonAttributesMapping(field, def, false);
-      additionalAttributesMapping(field.type, field, def);
-
-      propsMap.set(field.name, def);
+      // If the type is a sub schema it will have been stored in the cache.
+      if (recordCache[field.type]) {
+        propsMap.set(field.name, recordCache[field.type]);
+      } else {
+        const def = await convertAvroToJsonSchema(field.type, false, recordCache);
+
+        requiredAttributesMapping(field, jsonSchema, field.default !== undefined);
+        commonAttributesMapping(field, def, false);
+        additionalAttributesMapping(field.type, field, def);
+
+        propsMap.set(field.name, def);
+        // If there is a name for the sub record cache it under the name.
+        const qualifiedFieldName = getFullyQualifiedName(field.type);
+        cacheAvroRecordDef(recordCache, qualifiedFieldName, def);
+      }
     }
     jsonSchema.properties = Object.fromEntries(propsMap.entries());
     break;
@@ -214,5 +240,6 @@ async function convertAvroToJsonSchema(avroDefinition, isTopLevel) {
 }
 
 module.exports.avroToJsonSchema = async function avroToJsonSchema(avroDefinition) {
+  validateAvroSchema(avroDefinition);
   return convertAvroToJsonSchema(avroDefinition, true);
 };