def extract_features(file_path,
feature_extractors,
n=5,
numeric=True,
feature_list_dir_prefix='.'):
features = {}
for extractor in feature_extractors:
kwargs = feature_extractors[extractor]
'''
Special Case: ASM Extract returns two feature dicts - one for opcodes and one for numeric data.
When we want the opcode ngrams from ASM extractor, we take e.extract(kwargs=kwargs)[0] and when we
want numeric data, e.extract(kwargs=kwargs)[1]
'''
if extractor.__name__ == 'ASMExtractor' and numeric == False:
e = extractor(file_path, n)
features.update(e.extract(kwargs=kwargs)[0])
else:
e = extractor(file_path)
if extractor.__name__ == 'ASMExtractor' and numeric == True:
features.update(e.extract(kwargs=kwargs)[1])
else:
features.update(e.extract(kwargs=kwargs))
df = pd.DataFrame(data=[features],
columns=get_features_names(
feature_list_dir_prefix=feature_list_dir_prefix))
sparse_feature_vector = list(df.iloc[0, :])
return features, sparse_feature_vector
def extractWordFeatures(x):
"""
Chess Features
"""
features = {}
features = defaultdict(lambda: 0.0, features)
features.update(canCatchPawn(x))
if (features['cant catch pawn'] == 1):
return features
features.update(canBeCaptured(x))
# if (features['black_can_capture'] == 1):
# return {'black_can_capture':1}
features.update(isWhiteKingAhead(x))
features.update(isOpposition(x))
features.update(move_distances(x))
features.update(wrongSide(x))
features.update(ishPawn(x))
return features
def featureExtractor_3(x):
"""
Chess Features
"""
features = {}
features = Counter()
features.update(canCatchPawn(x))
if (features['cant catch pawn'] == 1):
return features
features.update(canBeCaptured(x))
# if (features['black_can_capture'] == 1):
# return {'black_can_capture':1}
features.update(isWhiteKingAhead(x))
features.update(isOpposition(x))
features.update(move_distances(x))
features.update(wrongSide(x))
features.update(ishPawn(x))
return features
def featureExtractor_3(x):
"""
Chess Features
"""
features = {}
features = Counter()
features.update(canCatchPawn(x))
if (features['cant catch pawn'] == 1):
return features
features.update(canBeCaptured(x))
# if (features['black_can_capture'] == 1):
# return {'black_can_capture':1}
features.update(isWhiteKingAhead(x))
features.update(isOpposition(x))
features.update(move_distances(x))
features.update(wrongSide(x))
features.update(ishPawn(x))
return features
def extractFeatures(x):
"""
Chess Features
"""
features = {}
features = defaultdict(lambda: 0.0, features)
features.update(canCatchPawn(x))
if (features['cant catch pawn'] == 1):
return features
features.update(canBeCaptured(x))
features.update(isWhiteKingAhead(x))
features.update(isOpposition(x))
features.update(move_distances(x))
features.update(wrongSide(x))
features.update(ishPawn(x))
return features
def text2feat_dict(text, extractors):
features = {}
for ename,extr in extractors.items():
features.update(extr.extract_from(text))
return features
def func_features(): y=update() return y
