Added basic kaldi-gstreamer-server support

examples/audio_transcribe.py

@@ -68,3 +68,13 @@
     print("IBM Speech to Text could not understand audio")
 except sr.RequestError as e:
     print("Could not request results from IBM Speech to Text service; {0}".format(e))
+
+# recognize speech using kaldi-gstreamer-server at localhost
+KALDI_HOST = "localhost:8888"
+try:
+    print("Kaldi thinks you said " + str(r.recognize_kaldi(audio, host=KALDI_HOST)))
+except sr.UnknownValueError:
+    print("Kaldi could not understand audio")
+except sr.RequestError as e:
+    print("Could not request results from Kaldi; {0}".format(e))
+

examples/extended_results.py

@@ -76,3 +76,12 @@
     print("IBM Speech to Text could not understand audio")
 except sr.RequestError as e:
     print("Could not request results from IBM Speech to Text service; {0}".format(e))
+
+# recognize speech using kaldi-gstreamer-server at localhost                                 
+KALDI_HOST = "localhost:8888"
+try:
+    print("Kaldi thinks you said " + str(r.recognize_kaldi(audio, host=KALDI_HOST, show_all=True)))
+except sr.UnknownValueError:
+    print("Kaldi could not understand audio")
+except sr.RequestError as e:
+    print("Could not request results from Kaldi; {0}".format(e))

speech_recognition/__init__.py

@@ -675,6 +675,66 @@ def recognize_google(self, audio_data, key = None, language = "en-US", show_all
                 return entry["transcript"]
         raise UnknownValueError() # no transcriptions available
 
+    def recognize_kaldi(self, audio_data, host, show_all = False):
+        """
+        Performs speech recognition on ``audio_data`` (an ``AudioData`` instance),
+        using the http POST api for kaldi-gstreamer-server.
+
+        Returns the most likely transcription if ``show_all`` is false (the default).
+        Otherwise, returns the raw API response as a JSON dictionary.
+
+        Raises a ``speech_recognition.UnknownValueError`` exception if the speech
+        is unintelligible. Raises a ``speech_recognition.RequestError`` exception
+        if the speech recognition operation failed,  or if there is no internet connection.
+
+        """
+        assert isinstance(audio_data, AudioData), "`audio_data` must be audio data"
+        assert isinstance(host, str), "`host` must be a string"
+
+        response_status={
+                0: "Success.",
+                2: "Aborted.",
+                1: "No speech.",
+                5: "Internal data flow error.",
+                9: "Not available."
+        }
+
+        url = "http://{0}/client/dynamic/recognize".format(host)
+        content_type=("audio/x-raw-int;rate={0}").format(audio_data.sample_rate)
+
+        request = Request(url, data = audio_data.get_raw_data(), headers = {
+            "Content-Type": content_type })
+
+        # obtain audio transcription results
+
+        try:
+            response = urlopen(request)
+        except HTTPError as e:
+            # use getattr to be compatible with Python 2.6
+            raise RequestError("recognition request failed: {0}".format(getattr(e, "reason", "status {0}".format(e.code))))
+        except URLError as e:
+            raise RequestError("recognition connection failed: {0}".format(e.reason))
+        response_text = response.read().decode("utf-8")
+
+        actual_result = json.loads(response_text)
+
+        status = int(actual_result['status'])
+
+        if status not in (0,1):
+            raise RequestError("Server returned: {0} - {1} ".format(status,response_status[status]))
+
+        if "hypotheses" in actual_result:
+            result=actual_result["hypotheses"]
+
+        if show_all:
+            return actual_result
+
+        for entry in result:
+            if "utterance" in entry:
+                return entry["utterance"]
+
+        raise UnknownValueError() # no transcriptions available
+
     def recognize_wit(self, audio_data, key, show_all = False):
         """
         Performs speech recognition on ``audio_data`` (an ``AudioData`` instance), using the Wit.ai API.