添加inc和srcHEADmaster

1 files changed, 44 insertions, 0 deletions

diff --git a/src/dataset_build/based_sfh.py b/src/dataset_build/based_sfh.py
new file mode 100644
index 0000000..b3281ce
--- /dev/null
+++ b/src/dataset_build/based_sfh.py
@@ -0,0 +1,44 @@
+import re
+import ConfigParser
+import bisect
+import random
+
+term = {'not_null':(lambda x : len(x)!=0)}
+	
+config = ConfigParser.RawConfigParser()
+config.read("based_sfh.conf")
+raw_file_address = config.get("file","raw_file_address")
+ripe_file_address = config.get("file","ripe_file_address")
+
+class sfh_fingerprint(object):
+	
+	def __init__(self,sfh):
+		self.sfh = sfh
+
+	@staticmethod
+	def get_hashed_len(sfh):
+		p = r"\[+\d+?:+\d+?\]"
+		pattern = re.compile(p)
+		hashed_len_set = pattern.findall(sfh)
+		if (term['not_null'](hashed_len_set)):
+			hashed_len = 0
+			for x in xrange(0,len(hashed_len_set)):
+				hashed_len_num = re.split(r"\[|\]|:",hashed_len_set[x])
+				hashed_len = hashed_len + int(hashed_len_num[2]) - int(hashed_len_num[1])
+			return hashed_len/len(hashed_len_set)
+		else :
+			return -1
+
+i=0
+sfh_set = list()
+with open(raw_file_address,'r') as infile:
+	with open(ripe_file_address,'w') as outfile: 
+		for line in infile:
+			i+=1
+			if(i%100000==0):
+				print i
+			result = re.split(r';',line)
+			if(term['not_null'](result[3]) and term['not_null'](result[19])):
+				hashed_len = sfh_fingerprint.get_hashed_len(result[19])
+				if(hashed_len/int(result[3])>0.8):
+					outfile.write(result[19]+'\n')
\ No newline at end of file