def keyword_statistics(relationship_file, subtitle_file):
relation_list = csv_io.read_csv(relationship_file)
subtitle = read_subtitle_file(subtitle_file)
relation_patterns = {}
for relation in relation_list:
relation_patterns[relation] = '(^[->]*' + relation.lower() + '[,|\.|\?|\!].*)' + '|' + \
'(?<!(her|his|our|eir|our|\smy|.\sa))\s+' + relation.lower() + '[\.|,|\?|!|>]'
subtitle_interval = []
time_to_keyword = []
keyword_list = []
for line in subtitle:
if line.strip():
subtitle_interval.append(line)
if len(subtitle_interval) < 2:
continue
if len(subtitle_interval) == 2:
subtitle_time = line[:-2]
continue
time_to_keyword, keyword_list = keyword_matching(relation_patterns, line, subtitle_time,\
time_to_keyword, keyword_list)
else:
subtitle_interval=[]
frame_to_keyword = to_frame_keyword(time_to_keyword)
csv_io.write_csv(OUTPUT_ROOT_PATH + 'statistics_result.csv', frame_to_keyword)
csv_io.write_csv(OUTPUT_ROOT_PATH + 'keyword_list.csv', [keyword_list])
def social_reconstruction(keyword_list_file, relations_file):
keyword_list = csv_io.read_csv(keyword_list_file)
relations = json_io.read_json(relations_file)
relation_graph = {'nodes': [], 'links': []}
node_index = {}
index = 0
for keyword in keyword_list:
if keyword not in node_index:
relation_graph['nodes'].append({'name': keyword, 'group': index, 'ID': index})
node_index[keyword] = index
index += 1
for name, relation in relations.iteritems():
#total = sum(relation.values())
for person in relation:
#if total != 0 and (float(relation[person]) / total > (1.0/len(relation)) - 0.03 ):
relation_graph['links'].append({'source': node_index[name], 'target': node_index[person],
'value': relation[person], 'label': person })
relation_graph['links'].append({'source': node_index[person], 'target': node_index[name],
'value': relation[person], 'label': name })
print relation_graph
json_io.write_json('output/result/relation_graph.json', relation_graph)
def keyword_statistics(relationship_file, subtitle_file):
relation_list = csv_io.read_csv(relationship_file)
subtitle = read_subtitle_file(subtitle_file)
relation_patterns = {}
for relation in relation_list:
relation_patterns[relation] = '(^[->]*' + relation.lower() + '[,|\.|\?|\!].*)' + '|' + \
'(?<!(her|his|our|eir|our|\smy|.\sa))\s+' + relation.lower() + '[\.|,|\?|!|>]'
subtitle_interval = []
time_to_keyword = []
keyword_list = []
for line in subtitle:
if line.strip():
subtitle_interval.append(line)
if len(subtitle_interval) < 2:
continue
if len(subtitle_interval) == 2:
subtitle_time = line[:-2]
continue
time_to_keyword, keyword_list = keyword_matching(relation_patterns, line, subtitle_time,\
time_to_keyword, keyword_list)
else:
subtitle_interval = []
frame_to_keyword = to_frame_keyword(time_to_keyword)
csv_io.write_csv(OUTPUT_ROOT_PATH + 'statistics_result.csv',
frame_to_keyword)
csv_io.write_csv(OUTPUT_ROOT_PATH + 'keyword_list.csv', [keyword_list])
def reconstruct_role(recongition_merge_file, keword_list_file):
keyword_to_frame = json_io.read_json(recongition_merge_file)
keword_list = csv_io.read_csv(keword_list_file)
leading_keyword = keword_list[0]
for keyword, frame_list in keyword_to_frame.iteritems():
for frame in frame_list:
for face in frame_list[frame]:
name = keyword + str(face['frame_position']) + '.jpg'
face['img'] = cv2.imread(OUTPUT_PATH + '/img/' + name)
detector, matcher = cv_face.init_feature('orb')
# Find other characters
face_list = {}
character_list = {}
for keyword, frame_list in keyword_to_frame.iteritems():
print keyword
for frame in frame_list:
for face in frame_list[frame]:
if face and face['face_id'] not in face_list:
face_list[face['face_id']] = []
if face:
face_list[face['face_id']].append(face)
rank = sorted(face_list, key=lambda k: len(face_list[k]), reverse=True)
character_list[keyword] = [face_list[rank[0]]]
i=0
def movie_prosessing(movie_file, two_entity_file, search_result_file):
two_entity_set = json_io.read_json(two_entity_file)
keyword_search_result = csv_io.read_csv(search_result_file)
# load video
videoInput = cv2.VideoCapture(movie_file)
# crate a start_frame to end_frame dictionary for two_entity_set look up
start_end = {}
for row in keyword_search_result:
start_frame, end_frame = time_format.to_frame(row)
while start_frame in start_end:
start_frame = start_frame + 0.001
while end_frame in start_end:
end_frame = end_frame + 0.001
start_end[start_frame] = end_frame
frame = {}
face_count = 0
for keyword in two_entity_set:
for start_frame in two_entity_set[keyword]:
frame_position = int(start_frame) - 24 * 10
finish_frame = start_end[start_frame] + 24 * 10
while frame_position <= finish_frame:
print keyword
videoInput.set(cv2.cv.CV_CAP_PROP_POS_FRAMES, frame_position)
flag, img = videoInput.read()
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = cv2.equalizeHist(gray)
face_position_list, rects = cv_image.face_detect(gray, frame_position, (85, 85))
#face_position_list, rects = faceDetection(gray, frame_position)
if 0xFF & cv2.waitKey(5) == 27:
cv2.destroyAllWindows()
sys.exit(1)
if len(face_position_list) == 1:
print 'detected'
image_name = keyword + str(frame_position) + '.jpg'
cv_image.output_image(rects, img, OUTPUT_PATH + '/img/' + image_name)
for face_position in face_position_list:
face_count += 1
print face_count
frame[face_count] = { 'keyword' : keyword,
'face_position': face_position.tolist(),
'ID' : face_count,
'frame_position': frame_position,
'face_id': face_count}
frame_position += FRAME_INTERVAL
#close video
videoInput.release()
json_io.write_json(OUTPUT_PATH + 'frame.json', frame)
def find_relation(keyword_list_file, search_result_file, time_interval):
time_to_keyword = csv_io.read_csv(search_result_file)
keyword_list = csv_io.read_csv(keyword_list_file)
leading_keyword = keyword_list[0]
frame_to_keyword = {}
for row in time_to_keyword:
start_frame, end_frame = time_format.to_frame(row)
while start_frame in frame_to_keyword:
start_frame = start_frame + 0.001
while end_frame in frame_to_keyword:
end_frame = end_frame + 0.001
frame_to_keyword[start_frame] = row[1]
# Transfrom to timeline format
frame_list = frame_to_keyword.keys()
frame_list.sort()
relations = {}
for i in range(1, len(keyword_list)):
relations.update( {keyword_list[i] : count_ralation(keyword_list[i], frame_list, frame_to_keyword, time_interval)} )
count = 0
proper_relation = {}
for name, relation in relations.iteritems():
total = sum(relation.values())
proper_relation[name] = {}
print name,
for person in relation:
if proper_test(total, leading_keyword, person, relation):
proper_relation[name][person] = relation[person]
print person , relation[person],
count += 1
print
print str(time_interval/(24*60)) + ',' + str(count)
json_io.write_json('output/relations.json', proper_relation)
def video_processing(movie_file, role_list_file, search_result_file,
role_input_way):
# load frame-keyword files
keyword_search_result = csv_io.read_csv(search_result_file)
role_list = csv_io.read_csv(role_list_file)
# load video
videoInput = cv2.VideoCapture(movie_file)
frame = {}
keyword_id = 0
frame_number = 0
for row in keyword_search_result:
start_frame, end_frame, keyword = float(row[0]), float(row[1]), row[2]
frame_position = round(start_frame) + 24 * BACKWORD_EXPAND_TIME
finish_frame = round(end_frame) + 24 * FORWORD_EXPAND_TIME
keyword_id += 1
keyword_time = keyword + '_t' + str(keyword_id)
while frame_position <= finish_frame:
face_position_list, rects, img = frame_caputre(
videoInput, frame_position)
if len(face_position_list) >= 1:
print "detect face..."
image_name = OUTPUT_PATH + 'img/' + keyword_time + str(
frame_number)
#if role_input_way == 0:
cv_image.output_image(rects, img, image_name)
count = 0
for face_position in face_position_list:
if role_input_way == 'auto':
break
role_name = role_identify(
image_name + '-' + str(count) + '.jpg', role_list)
else:
role_name = role_input(role_list)
count += 1
if role_name == -1:
continue
else:
if keyword_time not in frame:
print keyword_time, role_name
frame[keyword_time] = {}
frame[keyword_time][role_name] = {
'keyword': keyword,
'face_position': face_position.tolist(),
'frame_position': frame_position,
'keyword_id': keyword_id,
'weight': 1,
'speaker': True
}
elif role_name in frame[keyword_time]:
frame[keyword_time][role_name]['weight'] += 1
else:
frame[keyword_time][role_name] = {
'keyword': keyword,
'face_position': face_position.tolist(),
'frame_position': frame_position,
'keyword_id': keyword_id,
'weight': 1,
'speaker': False
}
frame_number += 1
frame_position += FRAME_INTERVAL
#close video
videoInput.release()
json_io.write_json(OUTPUT_PATH + 'keywordt_roles.json', frame)
def keyword_search(name_file, relationship_file, subtitle_file):
# Read files
name_list = csv_io.read_csv(name_file)
relation_list = csv_io.read_csv(relationship_file)
subtitle = read_subtitle_file(subtitle_file)
# Create regular expression pattern for reuse
name_patterns = {}
for name in name_list:
name_patterns[name] = '[\s]*' + name.lower() + "[^'\w]"
relation_patterns = {}
for relation in relation_list:
relation_patterns[relation] = '[\s]*' + relation.lower() + "[^'\w]"
# Find keyword
time_to_keyword = []
subtitle_interval = []
keyword_number = 0
keyword_list = [""]
keyword_count = {}
for line in subtitle:
if line.strip():
subtitle_interval.append(line)
if len(subtitle_interval) < 2:
continue
if len(subtitle_interval) == 2:
subtitle_time = line[:-2]
continue
for name in name_patterns:
if keyword_number < MAX_KEYWORDS_IN_ONE_INTERVAL and re.search(name_patterns[name], line.lower()):
time_to_keyword.append([subtitle_time, name])
keyword_number += 1
if name not in keyword_list:
keyword_list.append(name)
keyword_count[name] = 1
else:
keyword_count[name] += 1
for relation in relation_patterns:
if keyword_number < MAX_KEYWORDS_IN_ONE_INTERVAL and re.search(relation_patterns[relation], line.lower()):
time_to_keyword.append([subtitle_time, relation])
keyword_number += 1
if relation not in keyword_list:
keyword_list.append(relation)
keyword_count[relation] = 1
else:
keyword_count[relation] += 1
else:
if keyword_number == MAX_KEYWORDS_IN_ONE_INTERVAL:
for i in range(MAX_KEYWORDS_IN_ONE_INTERVAL):
time_to_keyword.pop()
subtitle_interval=[]
keyword_number=0
count = Counter(values[1] for values in time_to_keyword)
total_count = sum(keyword_count.values())
filter_list = []
for name, freq in count.iteritems():
if float(freq)/total_count >= 0.012:
print name
else:
filter_list.append(name)
for name in filter_list:
keyword_list.remove(name)
time_to_keyword = list( (values[0], values[1]) for values in time_to_keyword if values[1] != name)
# Find the max keyword count as leading keyword
keyword_list[0] = max(keyword_count, key=keyword_count.get)
csv_io.write_csv(OUTPUT_ROOT_PATH + 'search_result.csv', time_to_keyword)
csv_io.write_csv(OUTPUT_ROOT_PATH + 'keyword_list.csv', [keyword_list])
def video_processing(movie_file, role_list_file, search_result_file, role_input_way):
# load frame-keyword files
keyword_search_result = csv_io.read_csv(search_result_file)
role_list = csv_io.read_csv(role_list_file)
# load video
videoInput = cv2.VideoCapture(movie_file)
frame = {}
keyword_id = 0
frame_number = 0
for row in keyword_search_result:
start_frame, end_frame, keyword = float(row[0]), float(row[1]), row[2]
frame_position = round(start_frame) + 24 * BACKWORD_EXPAND_TIME
finish_frame = round(end_frame) + 24 * FORWORD_EXPAND_TIME
keyword_id += 1
keyword_time = keyword + '_t' + str(keyword_id)
while frame_position <= finish_frame:
face_position_list, rects, img = frame_caputre(videoInput, frame_position)
if len(face_position_list) >= 1:
print "detect face..."
image_name = OUTPUT_PATH + 'img/' + keyword_time + str(frame_number)
#if role_input_way == 0:
cv_image.output_image(rects, img, image_name)
count = 0
for face_position in face_position_list:
if role_input_way == 'auto':
break
role_name = role_identify( image_name + '-' + str(count) + '.jpg', role_list)
else:
role_name = role_input(role_list)
count += 1
if role_name == -1:
continue
else:
if keyword_time not in frame:
print keyword_time, role_name
frame[keyword_time] = {}
frame[keyword_time][role_name] = {'keyword' : keyword,
'face_position' : face_position.tolist(),
'frame_position' : frame_position,
'keyword_id' : keyword_id,
'weight' : 1,
'speaker': True}
elif role_name in frame[keyword_time]:
frame[keyword_time][role_name]['weight'] += 1
else:
frame[keyword_time][role_name] = {'keyword' : keyword,
'face_position' : face_position.tolist(),
'frame_position' : frame_position,
'keyword_id' : keyword_id,
'weight' : 1,
'speaker': False}
frame_number += 1
frame_position += FRAME_INTERVAL
#close video
videoInput.release()
json_io.write_json(OUTPUT_PATH + 'keywordt_roles.json', frame)
for term, count in term_tf.iteritems():
count = count / length
json_io.write_json(output_path+doc+'.json', term_tf)
def to_db(mydb, term_id, document_list, doc_hash, input_dir):
for doc in document_list:
terms_tf = json_io.read_json(input_dir+doc)
for term, tf in terms_tf.iteritems():
term = term.replace("'", "")
if len(term) > 255:
term = term[:254]
sql = "INSERT INTO doc_lookups (doc_id,title,tf,term_id) VALUES (" \
+ "'" + str(doc_hash[doc[:-5]]) + "','" + doc[:-5] + "','" + str(tf) + "','" + str(term_id[term]) + "');"
mydb.exe_sql(sql)
if __name__=='__main__':
if len(sys.argv) > 1:
doc_hash = csv_io.read_csv('output/doc_hash.json')
input_dir = sys.argv[1]
output_dir = 'output/zh_tf/'
else:
doc_hash = csv_io.read_csv('output/doc_hash.json')
input_dir = 'output/en_tokens/'
output_dir = 'output/en_tf/'
document_list = get_docs_list(input_dir)
for doc in document_list:
terms = csv_io.read_csv(input_dir+doc)
tf(terms, output_dir)
def __init__(self):
self.relative_path = os.path.join("my_class/")
self.stopword_list = csv_io.read_csv(self.relative_path + 'stopword.csv') + [u',']
self.stemmer = PorterStemmer()
def reconstruct_role(recongition_merge_file, keword_list_file):
keyword_to_frame = json_io.read_json(recongition_merge_file)
keword_list = csv_io.read_csv(keword_list_file)
leading_keyword = keword_list[0]
for keyword, frame_list in keyword_to_frame.iteritems():
for frame in frame_list:
for face in frame_list[frame]:
name = keyword + str(face['frame_position']) + '.jpg'
face['img'] = cv2.imread(OUTPUT_PATH + '/img/' + name)
detector, matcher = cv_face.init_feature('orb')
# Find other characters
face_list = {}
character_list = {}
for keyword, frame_list in keyword_to_frame.iteritems():
print keyword
for frame in frame_list:
for face in frame_list[frame]:
if face and face['face_id'] not in face_list:
face_list[face['face_id']] = []
if face:
face_list[face['face_id']].append(face)
rank = sorted(face_list, key=lambda k: len(face_list[k]), reverse=True)
character_list[keyword] = [face_list[rank[0]]]
i=0
#for face in face_list[rank[5]]:
# i+=1
for j in rank:
face = face_list[j][0]
cv2.imwrite(OUTPUT_PATH + '/result2/' + keyword + str(i) + '.jpg', face['img'])
i += 1
if len(rank) > 1 and '-' in keyword:
character_list[keyword].append(face_list[rank[1]])
''' for i in range(1, len(rank)):
if cv_face.list_match(MIN_MATCH, character_list[keyword][0], face_list[rank[i]], detector, matcher):
continue
else:
character_list[keyword].append(face_list[rank[i]])
break
if len(character_list[keyword]) == 1:
character_list[keyword].append(face_list[rank[1]])'''
face_list = {}
print
role_list = {}
# Use leading role image to check
lead_role_list = character_list[leading_keyword]
for keyword, characters in character_list.iteritems():
if keyword == leading_keyword or len(characters) < 2:
continue
if leading_keyword in keyword:
print keyword, '---'
match_count1 = 0
match_count2 = 0
for face in character_list[leading_keyword][0]:
match_count1 += cv_face.get_match_rate(face['img'], characters[0][0]['img'])
cv2.imwrite(OUTPUT_PATH + '/result/' + '000' + keyword + '.jpg', characters[0][0]['img'])
for face in character_list[leading_keyword][0]:
match_count2 += cv_face.get_match_rate(face['img'], characters[1][0]['img'])
cv2.imwrite(OUTPUT_PATH + '/result/' + '001' + keyword + '.jpg', characters[1][0]['img'])
if match_count1 > match_count2:
print 'characters1', match_count1, match_count2
del characters[0]
else:
print 'characters2', match_count1, match_count2
del characters[1]
role_list[keyword.split('-')[0]] = characters[0]
for keyword, characters in character_list.iteritems():
if leading_keyword in keyword or len(characters) < 2:
continue
important_person = keyword.split('-')[1]
if important_person in role_list:
print keyword, important_person, '---'
match_count1 = 0
match_count2 = 0
for face in role_list[important_person]:
match_count1 += cv_face.get_match_rate(face['img'], characters[0][0]['img'])
#cv2.imwrite(OUTPUT_PATH + '/result/' + '000' + keyword + '.jpg', characters[0][0]['img'])
for face in role_list[important_person]:
match_count2 += cv_face.get_match_rate(face['img'], characters[1][0]['img'])
#cv2.imwrite(OUTPUT_PATH + '/result/' + '001' + keyword + '.jpg', characters[1][0]['img'])
if match_count1 > match_count2:
print 'characters1', match_count1, match_count2
del characters[0]
else:
print 'characters2', match_count1, match_count2
del characters[1]
else:
del characters[1]
# Output
for keyword, characters in character_list.iteritems():
for character in characters:
if '-' in keyword:
keyword = keyword.split('-')[0]
cv2.imwrite(OUTPUT_PATH + '/result/' + keyword + '.jpg', character[0]['img'])
def __init__(self):
self.relative_path = os.path.join("my_class/")
self.stopword_list = csv_io.read_csv(self.relative_path +
'stopword.csv') + [u',']
self.stemmer = PorterStemmer()
json_io.write_json(output_path + doc + '.json', term_tf)
def to_db(mydb, term_id, document_list, doc_hash, input_dir):
for doc in document_list:
terms_tf = json_io.read_json(input_dir + doc)
for term, tf in terms_tf.iteritems():
term = term.replace("'", "")
if len(term) > 255:
term = term[:254]
sql = "INSERT INTO doc_lookups (doc_id,title,tf,term_id) VALUES (" \
+ "'" + str(doc_hash[doc[:-5]]) + "','" + doc[:-5] + "','" + str(tf) + "','" + str(term_id[term]) + "');"
mydb.exe_sql(sql)
if __name__ == '__main__':
if len(sys.argv) > 1:
doc_hash = csv_io.read_csv('output/doc_hash.json')
input_dir = sys.argv[1]
output_dir = 'output/zh_tf/'
else:
doc_hash = csv_io.read_csv('output/doc_hash.json')
input_dir = 'output/en_tokens/'
output_dir = 'output/en_tf/'
document_list = get_docs_list(input_dir)
for doc in document_list:
terms = csv_io.read_csv(input_dir + doc)
tf(terms, output_dir)
