out.print_verbose('Loading input')
source1, source2 = load_input()
out.print_verbose('Sizes of data sets: ', len(source1), len(source2))
out.print_verbose('Words: ', struct.word_count(source1), struct.word_count(source2))
'''
/source.../ are structures that contain
{
0: { 'opinions': [('CÂMERA', 80.0)],
'verbatim': 'Câmera boa.'}
}
'''
evaluate.reset() # To start evaluating summaries of the current sources.
output_files.new_source(SOURCE1, SOURCE2, source1, source2, METHOD_NAME) # Prepare output files for the current sources.
map_scores_summary = {}
distinct_summaries = set()
time_total = 0
out.print_verbose('Making summaries\n\n')
print(' %5s %5s %5s %5s\n' % ('R', 'C', 'D', 'H'))
for repeat in range(REPEAT_TESTS):
time_initial = time()
# Make summary
if sum([source1[i]['sent'][a] for a in source1[i]['sent']]) < 0:
e1_neg[i] = source1[i]
for i in source2:
if sum([source2[i]['sent'][a] for a in source2[i]['sent']]) > 0:
e2_pos[i] = source2[i]
if sum([source2[i]['sent'][a] for a in source2[i]['sent']]) < 0:
e2_neg[i] = source2[i]
stats_e1_pos = struct.aspects_stats_SIMILARITY(e1_pos)
stats_e1_neg = struct.aspects_stats_SIMILARITY(e1_neg)
stats_e2_pos = struct.aspects_stats_SIMILARITY(e2_pos)
stats_e2_neg = struct.aspects_stats_SIMILARITY(e2_neg)
evaluate.reset() # To start evaluating summaries of the current sources.
output_files.new_source(SOURCE1, SOURCE2, source1, source2, 'Similarity') # Prepare output files for the current sources.
map_scores_summary = {}
distinct_summaries = set()
time_total = 0
out.print_verbose('Making summaries\n\n')
print(' %5s %5s %5s %5s\n' % ('R', 'C', 'D', 'H'))
w_e1_pos = word_count(e1_pos)
w_e1_neg = word_count(e1_neg)
w_e2_pos = word_count(e2_pos)
w_e2_neg = word_count(e2_neg)
'word_count': 21}
}
'''
# Get statistics about aspects in the source (mean, standard deviation, probability)
stats_source_1 = struct.aspects_stats(source1)
stats_source_2 = struct.aspects_stats(source2)
'''
/stats_.../ are structures of the form:
{'tela': {'mean': 83, 'prob': 0.07, 'std': 0},
'cor': {'mean': -87, 'prob': 0.21, 'std': 1.73}}
'''
evaluate.reset() # To start evaluating summaries of the current sources.
output_files.new_source(
SOURCE1, SOURCE2, source1, source2,
'Statistic') # Prepare output files for the current sources.
map_scores_summary = {}
distinct_summaries = set()
time_total = 0
out.print_verbose('Making summaries\n\n')
print(' %5s %5s %5s %5s\n' % ('R', 'C', 'D', 'H'))
for repeat in range(REPEAT_TESTS):
time_initial = time()
'text_info': ['bom', 'maquinar', 'começar', 'fotografar'],
'opinions': [['PRODUTO', 100]],
'word_count': 8}
}
'''
source1_proc = preprocess_CLUSTERING(source1)
source2_proc = preprocess_CLUSTERING(source2)
set1_pos = source1_proc['+']
set1_neg = source1_proc['-']
set2_pos = source2_proc['+']
set2_neg = source2_proc['-']
evaluate.reset() # To start evaluating summaries of the current sources.
output_files.new_source(SOURCE1, SOURCE2, source1, source2, 'Clustering') # Prepare output files for the current sources.
map_scores_summary = {}
distinct_summaries = set()
time_total = 0
out.print_verbose('Making summaries\n\n')
print(' %5s %5s %5s %5s\n' % ('R', 'C', 'D', 'H'))
repeat = 0
discarded = 0
while repeat < REPEAT_TESTS: