# This is part of Kaylee # -- this code is licensed GPLv3 # Copyright 2015-2016 Clayton G. Hobbs # Portions Copyright 2013 Jezra import re import json import hashlib import os from argparse import ArgumentParser, Namespace import requests from gi.repository import GLib class Config: """Keep track of the configuration of Kaylee""" # Name of the program, for later use program_name = "kaylee" # Directories conf_dir = os.path.join(GLib.get_user_config_dir(), program_name) cache_dir = os.path.join(GLib.get_user_cache_dir(), program_name) data_dir = os.path.join(GLib.get_user_data_dir(), program_name) # Configuration files opt_file = os.path.join(conf_dir, "options.json") # Cache files history_file = os.path.join(cache_dir, program_name + "history") hash_file = os.path.join(cache_dir, "hash.json") # Data files strings_file = os.path.join(data_dir, "sentences.corpus") lang_file = os.path.join(data_dir, 'lm') dic_file = os.path.join(data_dir, 'dic') def __init__(self): # Ensure necessary directories exist self._make_dir(self.conf_dir) self._make_dir(self.cache_dir) self._make_dir(self.data_dir) # Set up the argument parser self.parser = ArgumentParser() self.parser.add_argument("-i", "--interface", type=str, dest="interface", action='store', help="Interface to use (if any). 'g' for GTK or 'gt' for GTK" + " system tray icon") self.parser.add_argument("-c", "--continuous", action="store_true", dest="continuous", default=False, help="Start interface with 'continuous' listen enabled") self.parser.add_argument("-p", "--pass-words", action="store_true", dest="pass_words", default=False, help="Pass the recognized words as arguments to the shell" + " command") self.parser.add_argument("-H", "--history", type=int, action="store", dest="history", help="Number of commands to store in history file") self.parser.add_argument("-m", "--microphone", type=int, action="store", dest="microphone", default=None, help="Audio input card to use (if other than system default)") self.parser.add_argument("--valid-sentence-command", type=str, dest="valid_sentence_command", action='store', help="Command to run when a valid sentence is detected") self.parser.add_argument("--invalid-sentence-command", type=str, dest="invalid_sentence_command", action='store', help="Command to run when an invalid sentence is detected") # Read the configuration file self._read_options_file() # Parse command-line arguments, overriding config file as appropriate self.parser.parse_args(namespace=self.options) def _make_dir(self, directory): if not os.path.exists(directory): os.makedirs(directory) def _read_options_file(self): try: with open(self.opt_file, 'r') as f: self.options = json.load(f) self.options = Namespace(**self.options) except FileNotFoundError: # Make an empty options namespace self.options = Namespace() class Hasher: """Keep track of hashes for Kaylee""" def __init__(self, config): self.config = config try: with open(self.config.hash_file, 'r') as f: self.hashes = json.load(f) except IOError: # No stored hash self.hashes = {} def __getitem__(self, hashname): try: return self.hashes[hashname] except (KeyError, TypeError): return None def __setitem__(self, hashname, value): self.hashes[hashname] = value def get_hash_object(self): """Returns an object to compute a new hash""" return hashlib.sha256() def store(self): """Store the current hashes into a the hash file""" with open(self.config.hash_file, 'w') as f: json.dump(self.hashes, f) class LanguageUpdater: """ Handles updating the language using the online lmtool. This class provides methods to check if the corpus has changed, and to update the language to match the new corpus using the lmtool. This allows us to automatically update the language if the corpus has changed, saving the user from having to do this manually. """ def __init__(self, config): self.config = config self.hasher = Hasher(config) def update_language_if_changed(self): """Test if the language has changed, and if it has, update it""" if self.language_has_changed(): self.update_language() self.save_language_hash() def language_has_changed(self): """Use hashes to test if the language has changed""" self.stored_hash = self.hasher['language'] # Calculate the hash the language file has right now hasher = self.hasher.get_hash_object() with open(self.config.strings_file, 'rb') as sfile: buf = sfile.read() hasher.update(buf) self.new_hash = hasher.hexdigest() return self.new_hash != self.stored_hash def update_language(self): """Update the language using the online lmtool""" print('Updating language using online lmtool') host = 'http://www.speech.cs.cmu.edu' url = host + '/cgi-bin/tools/lmtool/run' # Submit the corpus to the lmtool response_text = "" with open(self.config.strings_file, 'rb') as corpus: files = {'corpus': corpus} values = {'formtype': 'simple'} r = requests.post(url, files=files, data=values) response_text = r.text # Parse response to get URLs of the files we need path_re = r'.*Index of (.*?).*' number_re = r'.*TAR([0-9]*?)\.tgz.*' for line in response_text.split('\n'): # If we found the directory, keep it and don't break if re.search(path_re, line): path = host + re.sub(path_re, r'\1', line) # If we found the number, keep it and break elif re.search(number_re, line): number = re.sub(number_re, r'\1', line) break lm_url = path + '/' + number + '.lm' dic_url = path + '/' + number + '.dic' self._download_file(lm_url, self.config.lang_file) self._download_file(dic_url, self.config.dic_file) def save_language_hash(self): self.hasher['language'] = self.new_hash self.hasher.store() def _download_file(self, url, path): r = requests.get(url, stream=True) if r.status_code == 200: with open(path, 'wb') as f: for chunk in r: f.write(chunk)