def __init__(self, cache_file_path=None):
self.knowledge_tree = KnowledgeTree(cache_file_path=cache_file_path)
self.stats = KnowledgeBaseStats()
def __init__(self):
self.knowledge_tree = KnowledgeTree()
class KnowledgeBase(object):
"""An abstract class that implements a knowledge base.
This knowledge base stores all the query results in a tree.
>>> from pylstar.KnowledgeBase import KnowledgeBase
>>> from pylstar.OutputQuery import OutputQuery
>>> from pylstar.Word import Word
>>> from pylstar.Letter import Letter
>>> word1 = Word([Letter('a'), Letter('b')])
>>> query1 = OutputQuery(word1)
>>> kbase = KnowledgeBase()
>>> kbase.resolve_query(query1)
Traceback (most recent call last):
...
Exception: Passive inference process
>>> word2 = Word([Letter('a'), Letter('b'), Letter('c')])
>>> word3 = Word([Letter('1'), Letter('2'), Letter('3')])
>>> kbase.add_word(input_word = word2, output_word = word3)
>>> word4 = Word([Letter('a'), Letter('d'), Letter('e')])
>>> word5 = Word([Letter('1'), Letter('4'), Letter('5')])
>>> kbase.add_word(input_word = word4, output_word = word5)
>>> word6 = Word([Letter('f'), Letter('g')])
>>> word7 = Word([Letter('6'), Letter('7')])
>>> kbase.add_word(input_word = word6, output_word = word7)
>>> kbase.resolve_query(query1)
>>> print(query1.output_word)
[Letter('1'), Letter('2')]
>>> word8 = Word([Letter('a'), Letter('d'), Letter('e')])
>>> query2 = OutputQuery(word8)
>>> kbase.resolve_query(query2)
>>> print(query2.output_word)
[Letter('1'), Letter('4'), Letter('5')]
"""
__metaclass__ = abc.ABCMeta
def __init__(self, cache_file_path=None):
self.knowledge_tree = KnowledgeTree(cache_file_path=cache_file_path)
self.stats = KnowledgeBaseStats()
def load_cache(self, possible_letters):
self.knowledge_tree.load_cache(possible_letters)
def write_cache(self):
self.knowledge_tree.write_cache()
def __str__(self):
return str(self.knowledge_tree)
def resolve_query(self, query):
"""This method can be use to interogate the cache for
the output word associated with the specified query. If no
previous knowledge can be found for this query, the input word
is submitted to the target.
"""
if query is None:
raise Exception("Query cannot be None")
query.output_word = self._resolve_word(query.input_word)
def _resolve_word(self, word):
if word is None:
raise Exception("Word cannot be None")
try:
self.stats.nb_query += 1
self.stats.nb_letter += len(word.letters)
return self.knowledge_tree.get_output_word(word)
except Exception:
self._logger.debug(
"Knowledge base has no previous knowledge for '{}'".format(
word))
self.stats.nb_submited_query += 1
self.stats.nb_submited_letter += len(word.letters)
output = self._execute_word(word)
if output is not None:
self.knowledge_tree.add_word(input_word=word,
output_word=output)
return output
def _execute_word(self, word):
"""This method must be overwritten by subclasses that implements
an active learning process.
"""
raise Exception("Passive inference process")
def add_word(self, input_word, output_word):
"""This method stores in the knowledge base the relationship between
the specified input_word and output_word
"""
self._logger.debug("adding : {}".format(','.join(
[str(l) for l in input_word.letters])))
self.knowledge_tree.add_word(input_word, output_word)
class KnowledgeBase(object):
"""An abstract class that implements a knowledge base.
This knowledge base stores all the query results in a tree.
>>> from pylstar.KnowledgeBase import KnowledgeBase
>>> from pylstar.OutputQuery import OutputQuery
>>> from pylstar.Word import Word
>>> from pylstar.Letter import Letter
>>> word1 = Word([Letter('a'), Letter('b')])
>>> query1 = OutputQuery(word1)
>>> kbase = KnowledgeBase()
>>> kbase.resolve_query(query1)
Traceback (most recent call last):
...
Exception: Passive inference process
>>> word2 = Word([Letter('a'), Letter('b'), Letter('c')])
>>> word3 = Word([Letter('1'), Letter('2'), Letter('3')])
>>> kbase.add_word(input_word = word2, output_word = word3)
>>> word4 = Word([Letter('a'), Letter('d'), Letter('e')])
>>> word5 = Word([Letter('1'), Letter('4'), Letter('5')])
>>> kbase.add_word(input_word = word4, output_word = word5)
>>> word6 = Word([Letter('f'), Letter('g')])
>>> word7 = Word([Letter('6'), Letter('7')])
>>> kbase.add_word(input_word = word6, output_word = word7)
>>> kbase.resolve_query(query1)
>>> print query1.output_word
[Letter('1'), Letter('2')]
>>> word8 = Word([Letter('a'), Letter('d'), Letter('e')])
>>> query2 = OutputQuery(word8)
>>> kbase.resolve_query(query2)
>>> print query2.output_word
[Letter('1'), Letter('4'), Letter('5')]
"""
__metaclass__ = abc.ABCMeta
def __init__(self):
self.knowledge_tree = KnowledgeTree()
def __str__(self):
return str(self.knowledge_tree)
def resolve_query(self, query):
"""This method can be use to interogate the cache for
the output word associated with the specified query. If no
previous knowledge can be found for this query, the input word
is submitted to the target.
"""
if query is None:
raise Exception("Query cannot be None")
query.output_word = self._resolve_word(query.input_word)
def _resolve_word(self, word):
if word is None:
raise Exception("Word cannot be None")
try:
return self.knowledge_tree.get_output_word(word)
except Exception:
self._logger.debug("Knowledge base has no previous knowledge for '{}'".format(word))
output = self._execute_word(word)
if output is not None:
self.knowledge_tree.add_word(input_word = word, output_word = output)
return output
def _execute_word(self, word):
"""This method must be overwritten by subclasses that implements
an active learning process.
"""
raise Exception("Passive inference process")
def add_word(self, input_word, output_word):
"""This method stores in the knowledge base the relationship between
the specified input_word and output_word
"""
self._logger.debug("adding : {}".format(','.join([str(l) for l in input_word.letters])))
self.knowledge_tree.add_word(input_word, output_word)
