From 1be5dcc06bf3653aafd0e6e0cd896677beae5db9 Mon Sep 17 00:00:00 2001
From: Hoan Luu Huu <110280845+xquanluu@users.noreply.github.com>
Date: Mon, 13 Feb 2023 20:37:32 +0700
Subject: [PATCH] feat: nvidia tts/stt speech (#201)

* feat: nvidia tts/stt speech

* fix: rename riva server uri

* fix: review comments

* fix: review comments

* sort vendors in alphabetic order

---------

Co-authored-by: Quan HL <quanluuhoang8@gmail.com>
Co-authored-by: Dave Horton <daveh@beachdognet.com>
---
 src/api/types.ts                              |  1 +
 .../internal/views/speech-services/form.tsx   | 28 +++++++++
 src/vendor/index.tsx                          | 31 +++++++---
 .../nvidia-speech-recognizer-lang.ts          | 62 +++++++++++++++++++
 .../nvidia-speech-synthesis-lang.ts           | 20 ++++++
 src/vendor/types.ts                           |  7 ++-
 6 files changed, 137 insertions(+), 12 deletions(-)
 create mode 100644 src/vendor/speech-recognizer/nvidia-speech-recognizer-lang.ts
 create mode 100644 src/vendor/speech-synthesis/nvidia-speech-synthesis-lang.ts

diff --git a/src/api/types.ts b/src/api/types.ts
index d460705..6cdc61e 100644
--- a/src/api/types.ts
+++ b/src/api/types.ts
@@ -307,6 +307,7 @@ export interface SpeechCredential {
   stt_api_key: null | string;
   stt_region: null | string;
   instance_id: null | string;
+  riva_server_uri: null | string;
 }
 
 export interface Alert {
diff --git a/src/containers/internal/views/speech-services/form.tsx b/src/containers/internal/views/speech-services/form.tsx
index eb2884d..12e0cad 100644
--- a/src/containers/internal/views/speech-services/form.tsx
+++ b/src/containers/internal/views/speech-services/form.tsx
@@ -26,6 +26,7 @@ import {
   VENDOR_WELLSAID,
   VENDOR_DEEPGRAM,
   VENDOR_IBM,
+  VENDOR_NVIDIA,
 } from "src/vendor";
 import { MSG_REQUIRED_FIELDS } from "src/constants";
 import {
@@ -75,6 +76,7 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
   const [tmpCustomTtsEndpoint, setTmpCustomTtsEndpoint] = useState("");
   const [customSttEndpoint, setCustomSttEndpoint] = useState("");
   const [tmpCustomSttEndpoint, setTmpCustomSttEndpoint] = useState("");
+  const [rivaServerUri, setRivaServerUri] = useState("");
 
   const handleFile = (file: File) => {
     const handleError = () => {
@@ -136,6 +138,9 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
           tts_api_key: ttsApiKey || null,
           tts_region: ttsRegion || null,
         }),
+        ...(vendor === VENDOR_NVIDIA && {
+          riva_server_uri: rivaServerUri || null,
+        }),
       };
 
       if (credential && credential.data) {
@@ -173,6 +178,7 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
               : null,
           client_id: vendor === VENDOR_NUANCE ? clientId : null,
           secret: vendor === VENDOR_NUANCE ? secretKey : null,
+          riva_server_uri: vendor == VENDOR_NVIDIA ? rivaServerUri : null,
         })
           .then(() => {
             toastSuccess("Speech credential created successfully");
@@ -261,6 +267,10 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
         setInstanceId(credential.data.instance_id);
       }
 
+      if (credential.data.riva_server_uri) {
+        setRivaServerUri(credential.data.riva_server_uri);
+      }
+
       setUseCustomTts(credential.data.use_custom_tts > 0 ? true : false);
       setUseCustomStt(credential.data.use_custom_stt > 0 ? true : false);
       setCustomTtsEndpoint(credential.data.custom_tts_endpoint || "");
@@ -626,6 +636,24 @@ export const SpeechServiceForm = ({ credential }: SpeechServiceFormProps) => {
             </fieldset>
           </React.Fragment>
         )}
+        {vendor === VENDOR_NVIDIA && (
+          <React.Fragment>
+            <fieldset>
+              <label htmlFor="riva_server_uri">
+                Riva Server Uri<span>*</span>
+              </label>
+              <input
+                id="riva_server_uri"
+                type="text"
+                name="riva_server_uri"
+                placeholder="Riva Server Uri"
+                value={rivaServerUri}
+                onChange={(e) => setRivaServerUri(e.target.value)}
+              />
+            </fieldset>
+          </React.Fragment>
+        )}
+
         <fieldset>
           <ButtonGroup left>
             <Button small subStyle="grey" as={Link} to={ROUTE_INTERNAL_SPEECH}>
diff --git a/src/vendor/index.tsx b/src/vendor/index.tsx
index 3329e61..e4e067e 100644
--- a/src/vendor/index.tsx
+++ b/src/vendor/index.tsx
@@ -16,35 +16,40 @@ export const VENDOR_WELLSAID = "wellsaid";
 export const VENDOR_NUANCE = "nuance";
 export const VENDOR_DEEPGRAM = "deepgram";
 export const VENDOR_IBM = "ibm";
+export const VENDOR_NVIDIA = "nvidia";
 
 export const vendors: VendorOptions[] = [
+  {
+    name: "AWS",
+    value: VENDOR_AWS,
+  },
   {
     name: "Google",
     value: VENDOR_GOOGLE,
   },
   {
-    name: "AWS",
-    value: VENDOR_AWS,
+    name: "Deepgram",
+    value: VENDOR_DEEPGRAM,
+  },
+  {
+    name: "IBM",
+    value: VENDOR_IBM,
   },
   {
     name: "Microsoft",
     value: VENDOR_MICROSOFT,
   },
-  {
-    name: "WellSaid",
-    value: VENDOR_WELLSAID,
-  },
   {
     name: "Nuance",
     value: VENDOR_NUANCE,
   },
   {
-    name: "deepgram",
-    value: VENDOR_DEEPGRAM,
+    name: "Nvidia",
+    value: VENDOR_NVIDIA,
   },
   {
-    name: "IBM",
-    value: VENDOR_IBM,
+    name: "WellSaid",
+    value: VENDOR_WELLSAID,
   },
 ];
 
@@ -98,12 +103,14 @@ export const useSpeechVendors = () => {
       import("./speech-recognizer/nuance-speech-recognizer-lang"),
       import("./speech-recognizer/deepgram-speech-recognizer-lang"),
       import("./speech-recognizer/ibm-speech-recognizer-lang"),
+      import("./speech-recognizer/nvidia-speech-recognizer-lang"),
       import("./speech-synthesis/aws-speech-synthesis-lang"),
       import("./speech-synthesis/google-speech-synthesis-lang"),
       import("./speech-synthesis/ms-speech-synthesis-lang"),
       import("./speech-synthesis/wellsaid-speech-synthesis-lang"),
       import("./speech-synthesis/nuance-speech-synthesis-lang"),
       import("./speech-synthesis/ibm-speech-synthesis-lang"),
+      import("./speech-synthesis/nvidia-speech-synthesis-lang"),
     ]).then(
       ([
         { default: awsRecognizer },
@@ -112,12 +119,14 @@ export const useSpeechVendors = () => {
         { default: nuanceRecognizer },
         { default: deepgramRecognizer },
         { default: ibmRecognizer },
+        { default: nvidiaRecognizer },
         { default: awsSynthesis },
         { default: googleSynthesis },
         { default: msSynthesis },
         { default: wellsaidSynthesis },
         { default: nuanceSynthesis },
         { default: ibmSynthesis },
+        { default: nvidiaynthesis },
       ]) => {
         if (!ignore) {
           setSpeech({
@@ -128,6 +137,7 @@ export const useSpeechVendors = () => {
               wellsaid: wellsaidSynthesis,
               nuance: nuanceSynthesis,
               ibm: ibmSynthesis,
+              nvidia: nvidiaynthesis,
             },
             recognizers: {
               aws: awsRecognizer,
@@ -136,6 +146,7 @@ export const useSpeechVendors = () => {
               nuance: nuanceRecognizer,
               deepgram: deepgramRecognizer,
               ibm: ibmRecognizer,
+              nvidia: nvidiaRecognizer,
             },
           });
         }
diff --git a/src/vendor/speech-recognizer/nvidia-speech-recognizer-lang.ts b/src/vendor/speech-recognizer/nvidia-speech-recognizer-lang.ts
new file mode 100644
index 0000000..793e798
--- /dev/null
+++ b/src/vendor/speech-recognizer/nvidia-speech-recognizer-lang.ts
@@ -0,0 +1,62 @@
+import type { Language } from "../types";
+
+export const languages: Language[] = [
+  {
+    name: "Arabic",
+    code: "ar-AR",
+  },
+  {
+    name: "English",
+    code: "en-US",
+  },
+  {
+    name: "English - GB",
+    code: "en-GB",
+  },
+  {
+    name: "Spanish - US",
+    code: "es-US",
+  },
+  {
+    name: "Spanish",
+    code: "es-ES",
+  },
+  {
+    name: "German",
+    code: "de-DE",
+  },
+  {
+    name: "French",
+    code: "fr-FR",
+  },
+  {
+    name: "Hindi",
+    code: "hi-IN",
+  },
+  {
+    name: "Russian",
+    code: "ru-RU",
+  },
+  {
+    name: "Korean",
+    code: "ko-KR",
+  },
+  {
+    name: "Brazilian-Portuguese",
+    code: "pt-BR",
+  },
+  {
+    name: "Japanese",
+    code: "ja-JP",
+  },
+  {
+    name: "Italian",
+    code: "it-IT",
+  },
+  {
+    name: "Mandarin",
+    code: "zh-CN",
+  },
+];
+
+export default languages;
diff --git a/src/vendor/speech-synthesis/nvidia-speech-synthesis-lang.ts b/src/vendor/speech-synthesis/nvidia-speech-synthesis-lang.ts
new file mode 100644
index 0000000..3ff7d7d
--- /dev/null
+++ b/src/vendor/speech-synthesis/nvidia-speech-synthesis-lang.ts
@@ -0,0 +1,20 @@
+import type { VoiceLanguage } from "../types";
+
+export const languages: VoiceLanguage[] = [
+  {
+    code: "en-US",
+    name: "English",
+    voices: [
+      {
+        value: "English-US.Female-1",
+        name: "Female",
+      },
+      {
+        value: "English-US.Male-1",
+        name: "Male",
+      },
+    ],
+  },
+];
+
+export default languages;
diff --git a/src/vendor/types.ts b/src/vendor/types.ts
index c0e8841..6916da1 100644
--- a/src/vendor/types.ts
+++ b/src/vendor/types.ts
@@ -4,8 +4,9 @@ export type Vendor =
   | "Microsoft"
   | "WellSaid"
   | "Nuance"
-  | "deepgram"
-  | "IBM";
+  | "Deepgram"
+  | "IBM"
+  | "Nvidia";
 
 export interface VendorOptions {
   name: Vendor;
@@ -60,6 +61,7 @@ export interface RecognizerVendors {
   nuance: Language[];
   deepgram: Language[];
   ibm: Language[];
+  nvidia: Language[];
 }
 
 export interface SynthesisVendors {
@@ -69,6 +71,7 @@ export interface SynthesisVendors {
   wellsaid: VoiceLanguage[];
   nuance: VoiceLanguage[];
   ibm: VoiceLanguage[];
+  nvidia: VoiceLanguage[];
 }
 
 export interface MSRawSpeech {