From 9fff4f49a4a4b22ed174f65344f61d73cc55d688 Mon Sep 17 00:00:00 2001
From: Julien Odent <jodent@fb.com>
Date: Thu, 17 Mar 2022 10:19:26 -0700
Subject: [PATCH] Live transcription support

Summary:
* emits `partialTranscription` and `fullTranscription` events for `speech()`
* `interactive` integration
* for simplicity, dup the response handler into 2 promises so we don't have to set up timeouts waiting for a final payload before resolving
* showing microphone input feedback in `interactive`
* including proxy support for `speech()`

Reviewed By: ruoyipu

Differential Revision: D34938698

fbshipit-source-id: ec0ac55c8778be8aff4de8b38c6a840f663ca28b
---
 CHANGES.md         |  6 +++++
 README.md          |  3 +++
 lib/interactive.js |  8 +++++++
 lib/wit.js         | 58 +++++++++++++++++++++++++++++++++++++---------
 package.json       |  2 +-
 tests/shared.js    |  1 -
 6 files changed, 65 insertions(+), 13 deletions(-)

diff --git a/CHANGES.md b/CHANGES.md
index 3462ece..44bf101 100644
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -1,3 +1,9 @@
+## v6.2.1
+
+- Emits `partialTranscription` and `fullTranscription` events.
+- Shows microphone input feedback for `interactive`.
+- Includes `proxy` support for `speech()`.
+
 ## v6.2.0
 
 Requires Node.js >= 6.17.1 to support ES6 directly.
diff --git a/README.md b/README.md
index 391bfd2..5f30af8 100644
--- a/README.md
+++ b/README.md
@@ -91,6 +91,9 @@ Takes the following paramters:
 - `context` - (optional) the [Context](https://wit.ai/docs/http/#context_link) object
 - `n` - (optional) the max number of intents and traits to get back
 
+Emits `partialTranscription` and `fullTranscription` events.
+The Promise returns the final JSON payload.
+
 See `lib/interactive.js` for an example.
 
 ### interactive
diff --git a/lib/interactive.js b/lib/interactive.js
index fe3324d..e7db41b 100644
--- a/lib/interactive.js
+++ b/lib/interactive.js
@@ -33,6 +33,13 @@ module.exports = (wit, handleResponse, context) => {
     prompt();
   };
 
+  wit.on('partialTranscription', text => {
+    console.log(text + '...');
+  });
+  wit.on('fullTranscription', text => {
+    console.log(text + ' (final)');
+  });
+
   rl.on('line', line => {
     line = line.trim();
     if (!line) {
@@ -72,6 +79,7 @@ module.exports = (wit, handleResponse, context) => {
       });
 
       microphone.start();
+      console.log('🎤 Listening...');
 
       return;
     }
diff --git a/lib/wit.js b/lib/wit.js
index 729446f..e4f1d2c 100644
--- a/lib/wit.js
+++ b/lib/wit.js
@@ -7,12 +7,14 @@
 const {DEFAULT_API_VERSION, DEFAULT_WIT_URL} = require('./config');
 const log = require('./log');
 const fetch = require('isomorphic-fetch');
-const Url = require('url');
+const EventEmitter = require('events');
 const HttpsProxyAgent = require('https-proxy-agent');
 const {Readable} = require('stream');
+const Url = require('url');
 
-class Wit {
+class Wit extends EventEmitter {
   constructor(opts) {
+    super();
     this.config = Object.freeze(validate(opts));
   }
 
@@ -76,23 +78,51 @@ class Wit {
     const fullURL = witURL + '/speech?' + encodeURIParams(params);
     logger.debug(method, fullURL);
 
-    return fetch(fullURL, {
+    const req = fetch(fullURL, {
       body,
       method,
+      proxy,
       headers: {
         ...headers,
         'Content-Type': contentType,
         'Transfer-Encoding': 'chunked',
       },
-    })
+    });
+
+    const _partialResponses = req
+      .then(
+        response =>
+          new Promise((resolve, reject) => {
+            logger.debug('status', response.status);
+            const bodyStream = response.body;
+            bodyStream.on('readable', () => {
+              let chunk;
+              let contents = '';
+              while (null !== (chunk = bodyStream.read())) {
+                contents += chunk.toString();
+              }
+              for (const {error, intents, text} of splitHttpChunks(
+                contents,
+              ).map(x => JSON.parse(x))) {
+                if (!(error || intents)) {
+                  logger.debug('[speech] partialTranscription:', text);
+                  this.emit('partialTranscription', text);
+                } else if (text) {
+                  logger.debug('[speech] fullTranscription:', text);
+                  this.emit('fullTranscription', text);
+                }
+              }
+            });
+          }),
+      )
+      .catch(e => logger.error('[speech] could not parse partial response', e));
+
+    return req
       .then(response => Promise.all([response.text(), response.status]))
-      .then(([contents, status]) => {
-        const chunks = contents
-          .split('\r\n')
-          .map(x => x.trim())
-          .filter(x => x.length > 0);
-        return [JSON.parse(chunks[chunks.length - 1]), status];
-      })
+      .then(([contents, status]) => [
+        JSON.parse(splitHttpChunks(contents).pop()),
+        status,
+      ])
       .catch(e => e)
       .then(makeWitResponseHandler(logger, 'speech'));
   }
@@ -147,6 +177,12 @@ const encodeURIParams = params =>
     .map(([key, value]) => key + '=' + encodeURIComponent(value))
     .join('&');
 
+const splitHttpChunks = response =>
+  response
+    .split('\r\n')
+    .map(x => x.trim())
+    .filter(x => x.length > 0);
+
 const validate = opts => {
   if (!opts.accessToken) {
     throw new Error(
diff --git a/package.json b/package.json
index e744dca..04ddf74 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "node-wit",
-  "version": "6.2.0",
+  "version": "6.2.1",
   "description": "Wit.ai Node.js SDK",
   "keywords": [
     "wit",
diff --git a/tests/shared.js b/tests/shared.js
index d32f17e..06c2219 100644
--- a/tests/shared.js
+++ b/tests/shared.js
@@ -65,7 +65,6 @@ module.exports.runTests = wit => {
     });
 
     it('tests that Wit has correct functions', () => {
-      expect(Object.keys(client)).to.eql(['config']);
       expect(typeof client.message).to.eql('function');
       expect(typeof client.speech).to.eql('function');
     });