Vosk api: allow selecting different models and automatic model download

README.rst

@@ -57,8 +57,6 @@ The `library reference <https://github.com/Uberi/speech_recognition/blob/master/
 
 See `Notes on using PocketSphinx <https://github.com/Uberi/speech_recognition/blob/master/reference/pocketsphinx.rst>`__ for information about installing languages, compiling PocketSphinx, and building language packs from online resources. This document is also included under ``reference/pocketsphinx.rst``.
 
-You have to install Vosk models for using Vosk. `Here <https://alphacephei.com/vosk/models>`__ are models avaiable. You have to place them in models folder of your project, like "your-project-folder/models/your-vosk-model"
-
 Examples
 --------
 
@@ -143,9 +141,14 @@ Vosk API is **required if and only if you want to use Vosk recognizer** (``recog
 
 You can install it with ``python3 -m pip install vosk``.
 
-You also have to install Vosk Models:
+Languages can be selected with the language parameter e.g. ``recognizer_instance.recognize_vosk(language='de')``.
+Vosk will attempt to download the respective model from https://alphacephei.com/vosk/models automatically.
+Language defaults to english ``'en-us'``.
+
+It is possible to manually download a model and place it in a directory in your project folder.
+Reference this folder with the model parameter ``model='folder-name'``. This will take precedence over the language parameter.
 
-`Here <https://alphacephei.com/vosk/models>`__ are models avaiable for download. You have to place them in models folder of your project, like "your-project-folder/models/your-vosk-model"
+Models are avaiable for download `here <https://alphacephei.com/vosk/models>`__.
 
 Google Cloud Speech Library for Python (for Google Cloud Speech API users)
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

speech_recognition/__init__.py

@@ -1684,16 +1684,31 @@ def recognize_whisper(self, audio_data, model="base", show_dict=False, load_opti
             return result["text"]
 
 
-    def recognize_vosk(self, audio_data, language='en'):
+    def recognize_vosk(self, audio_data, model='', language='en-us'):
         from vosk import Model, KaldiRecognizer
-        
+
         assert isinstance(audio_data, AudioData), "Data must be audio data"
-        
+
         if not hasattr(self, 'vosk_model'):
-            if not os.path.exists("model"):
-                return "Please download the model from https://github.com/alphacep/vosk-api/blob/master/doc/models.md and unpack as 'model' in the current folder."
-                exit (1)
-            self.vosk_model = Model("model")
+            if model:
+                if not os.path.exists(model):
+                    raise RequestError(f"Please download the model from https://github.com/alphacep/vosk-api/blob/master/doc/models.md and unpack as '{model}' in the current folder.")
+                self.vosk_model = Model(model)
+            else:
+                try:
+                    import requests
+                except ImportError:
+                    raise RequestError("requests module is required to download model data")
+                # verify this language is available via api
+                response = requests.get('https://alphacephei.com/vosk/models/model-list.json', timeout=10)
+                # raise error if bad response
+                response.raise_for_status()
+
+                models = response.json()
+                languages = { m["lang"] for m in models }
+                if language not in languages:
+                    raise RequestError(f"Language '{language}' not available. Available language codes are: {languages}")
+                self.vosk_model = Model(lang=language)
 
         rec = KaldiRecognizer(self.vosk_model, 16000);