def main(words_path, vectors_path, weight_path, fpc_name, test_name):
# Loading preprocessed words, vectors and weight4ind files.
print("loading words file...")
words = pickle.load(open(words_path, 'rb'))
print("loading vectors file...")
vectors = pickle.load(open(vectors_path, 'rb'))
print("loading weight4ind file...")
weight4ind = pickle.load(open(weight_path, 'rb'))
rmpc = 1
params = params.params()
params.rmpc = rmpc
fpc_file = fpc_name
test_dataset = test_name
print("calculating sentence similarity scores, use fpc file: {}.".format(fpc_file))
pearson, mse = eval.sim_evaluate_one(vectors, words, weight4ind, sim_algo.weighted_average_sim_rmpc, params, fpc_file, test_dataset)
print("loading words file...")
words = pickle.load(open(words_path, 'rb'))
print("loading vectors file...")
vectors = pickle.load(open(vectors_path, 'rb'))
print("loading weight4ind file...")
weight4ind = pickle.load(open(weight_path, 'rb'))
# Using a list of datasets to generate the corresponding fpc files.
dataset_dir = "../data/"
dataset_list = [f for f in listdir(dataset_dir) if isfile(join(dataset_dir, f))]
for dataset_file in dataset_list:
print("preparing the first principle component based on {}.".format(str(dataset_file)))
eval.prepare_first_pc(vectors, words, weight4ind, sim_algo.get_first_pc, params, dataset_file)
test_dataset = 'sicktest' # name of the test dataset
pearson_list = []
mse_list = []
index = [fpc for fpc in fpc_list]
# Using a list of fpc files to evaluate on datasets.
fpc_dir = "../first_principle_component/"
fpc_list = [f for f in listdir(fpc_dir) if isfile(join(fpc_dir, f))]
for fpc_file in fpc_list:
print("calculating sentence similarity scores, use fpc file: {}.".format(fpc_file))
pearson, mse = eval.sim_evaluate_one(vectors, words, weight4ind, sim_algo.weighted_average_sim_rmpc, params, fpc_file, test_dataset)
pearson_list.append(pearson)
mse_list.append(mse)
if __name__ == '__main__':
plac.call(main)
parr4para = {}
sarr4para = {}
for wordfile in wordfiles:
(words, We) = data_io.getWordmap(wordfile)
for weightpara in weightparas:
word2weight = data_io.getWordWeight(weightfile, weightpara)
weight4ind = data_io.getWeight(words, word2weight)
for rmpc in rmpcs:
print('word vectors loaded from %s' % wordfile)
print('word weights computed from %s using parameter a=%f' %
(weightfile, weightpara))
params.rmpc = rmpc
print('remove the first %d principal components' % rmpc)
## eval just one example dataset
parr, sarr = eval.sim_evaluate_one(
We, words, weight4ind, sim_algo.weighted_average_sim_rmpc,
params)
## eval all datasets; need to obtained datasets from John Wieting (https://github.com/jwieting/iclr2016)
# parr, sarr = eval.sim_evaluate_all(We, words, weight4ind, sim_algo.weighted_average_sim_rmpc, params)
paras = (wordfile, weightfile, weightpara, rmpc)
parr4para[paras] = parr
sarr4para[paras] = sarr
## save results
save_result = False #True
result_file = 'result/sim_sif.result'
comment4para = [ # need to align with the following loop
['word vector files', wordfiles], # comments and values,
['weight parameters', weightparas],
['remove principal component or not', rmpcs]
]
