def _get_available_models(cls, recognizer):
available_models = SortedDict()
for directory in cls._get_search_path("models"):
directory = os.path.join(directory, recognizer)
if not os.path.exists(directory):
continue
meta_files = glob.glob(os.path.join(directory, "*.meta"))
for meta_file in meta_files:
meta = Recognizer.read_meta_file(meta_file)
if not meta.has_key("name") or \
not meta.has_key("shortname"):
continue
model_file = meta_file.replace(".meta", ".model")
if meta.has_key("path") and not os.path.exists(meta["path"]):
# skip model if specified path is incorrect
continue
elif not meta.has_key("path") and os.path.exists(model_file):
# if path option is missing, assume the .model file
# is in the same directory
meta["path"] = model_file
available_models[meta["name"]] = meta
return available_models
def read_meta_file(cls, meta_file):
"""
Read a .meta file.
@type meta_file: str
@param meta_file: meta file file to read
@rtype: dict
"""
f = open(meta_file)
ret = SortedDict()
for line in f.readlines():
try:
key, value = [s.strip() for s in line.strip().split("=")]
ret[key] = value
except ValueError:
continue
f.close()
return ret
def _load_available_recognizers(cls):
cls.available_recognizers = SortedDict()
for directory in cls._get_search_path("engines"):
if not os.path.exists(directory):
continue
for f in glob.glob(os.path.join(directory, "*.py")):
if f.endswith("__init__.py") or f.endswith("setup.py"):
continue
module_name = os.path.basename(f).replace(".py", "")
module_name += "recognizer"
module = imp.load_source(module_name, f)
try:
name = module.RECOGNIZER_CLASS.RECOGNIZER_NAME
cls.available_recognizers[name] = module.RECOGNIZER_CLASS
except AttributeError:
pass
def get_char_dict(self, directory, corpora):
"""
Returns a dictionary with xml file list.
keys are character codes.
values are arrays of xml files.
directory: root directory
corpora: corpora list to restrict to
"""
charcol = CharacterCollection()
for file in glob.glob(os.path.join(directory, "*", "*")):
corpus_name = file.split("/")[-2]
# exclude data which are not in the wanted corpora
if corpus_name not in corpora:
continue
if os.path.isdir(file):
self.print_verbose("Loading dir %s" % file)
charcol += CharacterCollection.from_character_directory(file)
elif ".charcol" in file:
self.print_verbose("Loading charcol %s" % file)
gzip = False
bz2 = False
if file.endswith(".gz"): gzip = True
if file.endswith(".bz2"): bz2 = True
charcol2 = CharacterCollection()
charcol2.read(file, gzip=gzip, bz2=bz2)
charcol += charcol2
self.print_verbose("Grouping characters together...")
dic = SortedDict()
for set_name in charcol.get_set_list():
for char in charcol.get_characters(set_name):
charcode = ord(char.get_unicode())
if not charcode in dic: dic[charcode] = []
dic[charcode].append(char)
return dic
def _update_set_ids(self):
self._SETIDS = SortedDict()
for row in self._efa("SELECT * FROM character_sets ORDER BY setid"):
self._SETIDS[row['name'].encode("utf8")] = row['setid']
