Support non-ascii (unicode) characters in service name, operation nam…

…e and parameter name. This replaces regexp patterns that only worked with ascii characters with more proper matching that supports unicode identifiers in typescript/javascript. The platform must support "unicode-aware mode" (the u flag) for this to work. https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/RegExp/unicode
hey-api · Feb 16, 2024 · 7588831 · 7588831
1 parent 308a180
commit 7588831
Show file tree

Hide file tree

Showing 13 changed files with 172 additions and 29 deletions.
diff --git a/src/openApi/v2/parser/getOperationName.ts b/src/openApi/v2/parser/getOperationName.ts
@@ -1,18 +1,15 @@
 import camelCase from 'camelcase';
 
+import sanitizeOperationName from '../../../utils/sanitizeOperationName';
+
 /**
  * Convert the input value to a correct operation (method) classname.
  * This will use the operation ID - if available - and otherwise fallback
  * on a generated name from the URL
  */
 export const getOperationName = (url: string, method: string, operationId?: string): string => {
     if (operationId) {
-        return camelCase(
-            operationId
-                .replace(/^[^a-zA-Z]+/g, '')
-                .replace(/[^\w\-]+/g, '-')
-                .trim()
-        );
+        return camelCase(sanitizeOperationName(operationId).trim());
     }
 
     const urlWithoutPlaceholders = url

diff --git a/src/openApi/v2/parser/getOperationParameterName.ts b/src/openApi/v2/parser/getOperationParameterName.ts
@@ -1,15 +1,13 @@
 import camelCase from 'camelcase';
 
 import { reservedWords } from '../../../utils/reservedWords';
+import sanitizeOperationParameterName from '../../../utils/sanitizeOperationParameterName';
 
 /**
  * Replaces any invalid characters from a parameter name.
  * For example: 'filter.someProperty' becomes 'filterSomeProperty'.
  */
 export const getOperationParameterName = (value: string): string => {
-    const clean = value
-        .replace(/^[^a-zA-Z]+/g, '')
-        .replace(/[^\w\-]+/g, '-')
-        .trim();
+    const clean = sanitizeOperationParameterName(value).trim();
     return camelCase(clean).replace(reservedWords, '_$1');
 };
diff --git a/src/openApi/v2/parser/getServiceName.ts b/src/openApi/v2/parser/getServiceName.ts
@@ -1,13 +1,12 @@
 import camelCase from 'camelcase';
 
+import sanitizeServiceName from '../../../utils/sanitizeServiceName';
+
 /**
  * Convert the input value to a correct service name. This converts
  * the input string to PascalCase.
  */
 export const getServiceName = (value: string): string => {
-    const clean = value
-        .replace(/^[^a-zA-Z]+/g, '')
-        .replace(/[^\w\-]+/g, '-')
-        .trim();
+    const clean = sanitizeServiceName(value).trim();
     return camelCase(clean, { pascalCase: true });
 };
diff --git a/src/openApi/v3/parser/getOperationName.ts b/src/openApi/v3/parser/getOperationName.ts
@@ -1,18 +1,15 @@
 import camelCase from 'camelcase';
 
+import sanitizeOperationName from '../../../utils/sanitizeOperationName';
+
 /**
  * Convert the input value to a correct operation (method) classname.
  * This will use the operation ID - if available - and otherwise fallback
  * on a generated name from the URL
  */
 export const getOperationName = (url: string, method: string, operationId?: string): string => {
     if (operationId) {
-        return camelCase(
-            operationId
-                .replace(/^[^a-zA-Z]+/g, '')
-                .replace(/[^\w\-]+/g, '-')
-                .trim()
-        );
+        return camelCase(sanitizeOperationName(operationId).trim());
     }
 
     const urlWithoutPlaceholders = url

diff --git a/src/openApi/v3/parser/getOperationParameterName.ts b/src/openApi/v3/parser/getOperationParameterName.ts
@@ -1,16 +1,13 @@
 import camelCase from 'camelcase';
 
 import { reservedWords } from '../../../utils/reservedWords';
+import sanitizeOperationParameterName from '../../../utils/sanitizeOperationParameterName';
 
 /**
  * Replaces any invalid characters from a parameter name.
  * For example: 'filter.someProperty' becomes 'filterSomeProperty'.
  */
 export const getOperationParameterName = (value: string): string => {
-    const clean = value
-        .replace(/^[^a-zA-Z]+/g, '')
-        .replace('[]', 'Array')
-        .replace(/[^\w\-]+/g, '-')
-        .trim();
+    const clean = sanitizeOperationParameterName(value).trim();
     return camelCase(clean).replace(reservedWords, '_$1');
 };
diff --git a/src/openApi/v3/parser/getServiceName.spec.ts b/src/openApi/v3/parser/getServiceName.spec.ts
@@ -9,5 +9,6 @@ describe('getServiceName', () => {
         expect(getServiceName('@fooBar')).toEqual('FooBar');
         expect(getServiceName('$fooBar')).toEqual('FooBar');
         expect(getServiceName('123fooBar')).toEqual('FooBar');
+        expect(getServiceName('non-ascii-æøåÆØÅöôêÊ字符串')).toEqual('NonAsciiÆøåÆøÅöôêÊ字符串');
     });
 });
diff --git a/src/openApi/v3/parser/getServiceName.ts b/src/openApi/v3/parser/getServiceName.ts
@@ -1,13 +1,12 @@
 import camelCase from 'camelcase';
 
+import sanitizeServiceName from '../../../utils/sanitizeServiceName';
+
 /**
  * Convert the input value to a correct service name. This converts
  * the input string to PascalCase.
  */
 export const getServiceName = (value: string): string => {
-    const clean = value
-        .replace(/^[^a-zA-Z]+/g, '')
-        .replace(/[^\w\-]+/g, '-')
-        .trim();
+    const clean = sanitizeServiceName(value).trim();
     return camelCase(clean, { pascalCase: true });
 };
diff --git a/src/utils/sanitizeOperationName.ts b/src/utils/sanitizeOperationName.ts
@@ -0,0 +1,7 @@
+import sanitizeServiceName from './sanitizeServiceName';
+
+/**
+ * sanitizeOperationName does the same as sanitizeServiceName.
+ */
+const sanitizeOperationName = sanitizeServiceName;
+export default sanitizeOperationName;
diff --git a/src/utils/sanitizeOperationParameterName.ts b/src/utils/sanitizeOperationParameterName.ts
@@ -0,0 +1,7 @@
+import sanitizeOperationName from './sanitizeOperationName';
+
+const sanitizeOperationParameterName = (name: string): string => {
+    const withoutBrackets = name.replace('[]', 'Array');
+    return sanitizeOperationName(withoutBrackets);
+};
+export default sanitizeOperationParameterName;
diff --git a/src/utils/sanitizeServiceName.ts b/src/utils/sanitizeServiceName.ts
@@ -0,0 +1,18 @@
+/**
+ * Sanitizes service names, so they are valid typescript identifiers of a certain form.
+ *
+ * 1: Remove any leading characters that are illegal as starting character of a typescript identifier.
+ * 2: Replace illegal characters in remaining part of type name with underscore (-).
+ *
+ * Step 1 should perhaps instead also replace illegal characters with underscore, or prefix with it, like sanitizeEnumName
+ * does. The way this is now one could perhaps end up removing all characters, if all are illegal start characters. It
+ * would be sort of a breaking change to do so, though, previously generated code might change then.
+ *
+ * Javascript identifier regexp pattern retrieved from https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Lexical_grammar#identifiers
+ *
+ * The output of this is expected to be converted to PascalCase
+ */
+const sanitizeServiceName = (name: string) =>
+    name.replace(/^[^\p{ID_Start}]+/u, '').replace(/[^$\u200c\u200d\p{ID_Continue}]/gu, '-');
+
+export default sanitizeServiceName;
diff --git a/test/__snapshots__/index.spec.ts.snap b/test/__snapshots__/index.spec.ts.snap
@@ -684,6 +684,7 @@ export { MultipleTags1Service } from './services/MultipleTags1Service';
 export { MultipleTags2Service } from './services/MultipleTags2Service';
 export { MultipleTags3Service } from './services/MultipleTags3Service';
 export { NoContentService } from './services/NoContentService';
+export { NonAsciiÆøåÆøÅöôêÊService } from './services/NonAsciiÆøåÆøÅöôêÊService';
 export { ParametersService } from './services/ParametersService';
 export { ResponseService } from './services/ResponseService';
 export { SimpleService } from './services/SimpleService';
@@ -2841,6 +2842,36 @@ export class NoContentService {
 "
 `;
 
+exports[`v2 should generate: test/generated/v2/services/NonAsciiÆøåÆøÅöôêÊService.ts 1`] = `
+"/* generated using openapi-typescript-codegen -- do no edit */
+/* istanbul ignore file */
+/* tslint:disable */
+/* eslint-disable */
+import type { NonAsciiStringæøåÆØÅöôêÊ字符串 } from '../models/NonAsciiStringæøåÆØÅöôêÊ字符串';
+import type { CancelablePromise } from '../core/CancelablePromise';
+import { OpenAPI } from '../core/OpenAPI';
+import { request as __request } from '../core/request';
+export class NonAsciiÆøåÆøÅöôêÊService {
+    /**
+     * @param nonAsciiParamæøåÆøÅöôêÊ Dummy input param
+     * @returns NonAsciiStringæøåÆØÅöôêÊ字符串 Successful response
+     * @throws ApiError
+     */
+    public static nonAsciiæøåÆøÅöôêÊ字符串(
+        nonAsciiParamæøåÆøÅöôêÊ: number,
+    ): CancelablePromise<NonAsciiStringæøåÆØÅöôêÊ字符串> {
+        return __request(OpenAPI, {
+            method: 'POST',
+            url: '/api/v{api-version}/non-ascii-æøåÆØÅöôêÊ字符串',
+            query: {
+                'nonAsciiParamæøåÆØÅöôêÊ': nonAsciiParamæøåÆøÅöôêÊ,
+            },
+        });
+    }
+}
+"
+`;
+
 exports[`v2 should generate: test/generated/v2/services/ParametersService.ts 1`] = `
 "/* generated using openapi-typescript-codegen -- do no edit */
 /* istanbul ignore file */
@@ -3852,6 +3883,7 @@ export { MultipleTags1Service } from './services/MultipleTags1Service';
 export { MultipleTags2Service } from './services/MultipleTags2Service';
 export { MultipleTags3Service } from './services/MultipleTags3Service';
 export { NoContentService } from './services/NoContentService';
+export { NonAsciiÆøåÆøÅöôêÊService } from './services/NonAsciiÆøåÆøÅöôêÊService';
 export { ParametersService } from './services/ParametersService';
 export { RequestBodyService } from './services/RequestBodyService';
 export { ResponseService } from './services/ResponseService';
@@ -7045,6 +7077,36 @@ export class NoContentService {
 "
 `;
 
+exports[`v3 should generate: test/generated/v3/services/NonAsciiÆøåÆøÅöôêÊService.ts 1`] = `
+"/* generated using openapi-typescript-codegen -- do no edit */
+/* istanbul ignore file */
+/* tslint:disable */
+/* eslint-disable */
+import type { NonAsciiStringæøåÆØÅöôêÊ字符串 } from '../models/NonAsciiStringæøåÆØÅöôêÊ字符串';
+import type { CancelablePromise } from '../core/CancelablePromise';
+import { OpenAPI } from '../core/OpenAPI';
+import { request as __request } from '../core/request';
+export class NonAsciiÆøåÆøÅöôêÊService {
+    /**
+     * @param nonAsciiParamæøåÆøÅöôêÊ Dummy input param
+     * @returns NonAsciiStringæøåÆØÅöôêÊ字符串 Successful response
+     * @throws ApiError
+     */
+    public static nonAsciiæøåÆøÅöôêÊ字符串(
+        nonAsciiParamæøåÆøÅöôêÊ: number,
+    ): CancelablePromise<Array<NonAsciiStringæøåÆØÅöôêÊ字符串>> {
+        return __request(OpenAPI, {
+            method: 'POST',
+            url: '/api/v{api-version}/non-ascii-æøåÆØÅöôêÊ字符串',
+            query: {
+                'nonAsciiParamæøåÆØÅöôêÊ': nonAsciiParamæøåÆøÅöôêÊ,
+            },
+        });
+    }
+}
+"
+`;
+
 exports[`v3 should generate: test/generated/v3/services/ParametersService.ts 1`] = `
 "/* generated using openapi-typescript-codegen -- do no edit */
 /* istanbul ignore file */

diff --git a/test/spec/v2.json b/test/spec/v2.json
@@ -905,6 +905,33 @@
                     }
                 }
             }
+        },
+        "/api/v{api-version}/non-ascii-æøåÆØÅöôêÊ字符串": {
+            "post": {
+                "tags": [
+                    "Non-Ascii-æøåÆØÅöôêÊ"
+                ],
+                "operationId": "nonAsciiæøåÆØÅöôêÊ字符串",
+                "parameters": [
+                    {
+                        "description": "Dummy input param",
+                        "name": "nonAsciiParamæøåÆØÅöôêÊ",
+                        "in": "query",
+                        "required": true,
+                        "schema": {
+                            "type": "integer"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Successful response",
+                        "schema": {
+                            "$ref": "#/definitions/NonAsciiStringæøåÆØÅöôêÊ字符串"
+                        }
+                    }
+                }
+            }
         }
     },
     "definitions": {

diff --git a/test/spec/v3.json b/test/spec/v3.json
@@ -1464,6 +1464,40 @@
                     }
                 }
             }
+        },
+        "/api/v{api-version}/non-ascii-æøåÆØÅöôêÊ字符串": {
+            "post": {
+                "tags": [
+                    "Non-Ascii-æøåÆØÅöôêÊ"
+                ],
+                "operationId": "nonAsciiæøåÆØÅöôêÊ字符串",
+                "parameters": [
+                    {
+                        "description": "Dummy input param",
+                        "name": "nonAsciiParamæøåÆØÅöôêÊ",
+                        "in": "query",
+                        "required": true,
+                        "schema": {
+                            "type": "integer"
+                        }
+                    }
+                ],
+                "responses": {
+                    "200": {
+                        "description": "Successful response",
+                        "content": {
+                            "application/json": {
+                                "schema": {
+                                    "type": "array",
+                                    "items": {
+                                        "$ref": "#/components/schemas/NonAsciiStringæøåÆØÅöôêÊ字符串"
+                                    }
+                                }
+                            }
+                        }
+                    }
+                }
+            }
         }
     },
     "components": {