moonq/q-tools
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Adding viewer capability for duplicate/similarity searches

Browse Source
This commit is contained in:
ville rantanen
2014-05-22 14:09:42 +03:00
parent c9bfa7b9e2
commit 39f92f3523
1 changed files with 62 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

89
image_list.py
View File

@@ -41,7 +41,10 @@ def setup_options():
parser.add_argument("--small",action="store_true",dest="searchsmall",default=False,
help="Return a list of small files, smaller than -m INT. This option will flip the 'Add new files' option. [%(default)s]")
parser.add_argument("--similar",type=str,dest="similarity",default=None,
help="Search list for similar images. Value 0-255 for similarity threshold. 0=high similarity. If value is a filename, search similar to that image.")
help="Search list for similar images. Value 0-255 for similarity threshold. 0=high similarity. "+
"If value is a filename, search similar to that image.")
parser.add_argument("--viewer",type=str,dest="viewer",default=None,
help="Program to view images, %%f refers to filename(s)")
parser.add_argument("-x",action="append",dest="exclude",default=[],
help="Exclude folder name from the lists. This option may be issued several times")
parser.add_argument('startpath', action="store",default='.', nargs='?')
@@ -308,40 +311,65 @@ def get_fingerprint(filename):
values+='0'
return str(int(values,2))
def find_fingerprint_similar(sqlfile,thr):
def find_fingerprint_similar(opts):
''' Find all similar images, nearest match more similar than thr '''
conn=sqlite3.connect(sqlfile)
thr=int(opts.similarity)
conn=sqlite3.connect(opts.sqlfile)
conn.text_factory=str
db1=conn.cursor()
db2=conn.cursor()
db1.execute("SELECT file,fingerprint,width,height FROM list ORDER BY file")
db1.execute("SELECT file,fingerprint,sharpness,width,height,R,G,B FROM list WHERE sharpness > 0 ORDER BY file")
hits=[]
hit_list=[]
for i,hit1 in enumerate(db1):
if hit1[0] in hits:
if hit1[0] in hit_list:
continue
this=[hit1[0],'',sys.maxint,int(hit1[1]),hit1[2]*hit1[3]]
db2.execute("SELECT file,fingerprint,width,height FROM list ORDER BY file")
cmp=hit1[0]
fp=int(hit1[1])
sp=hit1[2]
dims=hit1[3:5]
pixels=dims[0]*dims[1]
colors=hit1[5:8]
db2.execute("SELECT file,fingerprint,sharpness,width,height,R,G,B FROM list WHERE sharpness > 0 ORDER BY file")
this1=[ [cmp, 0,sp,int(hit1[3]),int(hit1[4]),0,pixels] ]
for hit2 in db2:
if hit2[0]==this[0]:
if hit2[0]==cmp:
continue
similarity=bin(this[3]^int(hit2[1])).count('1')
similarity=bin(fp^int(hit2[1])).count('1')
if similarity<thr:
if similarity<this[2]:
this[2]=similarity
this[1]=hit2[0]
if hit2[2]*hit2[3] < this[4]:
foo=this[1]
this[1]=this[0]
this[0]=foo
hits.append(hit2[0])
if this[2]<thr:
print('"%s" "%s":%i' % ( this[0], this[1], this[2] ))
this2=[hit2[0], similarity, hit2[2],
int(hit2[3]),int(hit2[4]),
abs( hit2[5] - colors[0] )+abs( hit2[6] - colors[1] )+abs( hit2[7] - colors[2] ),
int(hit2[3])*int(hit2[4])]
this1.append(this2)
hit_list.append(hit2[0])
this1.sort(key=lambda x: x[1])
if len(this1)>1:
hits.append(this1)
hit_list.append(cmp)
if i==0:
print("No measurements found")
sys.exit(1)
for src in hits:
file_len=str(max([len(x[0]) for x in src]))
print( ('{: <'+file_len+'} {: >4} {: >4} {: >4} {: ^5}x{: ^5}').format("File","Diff","CDel","Shrp","Wth","Hgt"))
for c in range(len(src)):
print( ('{: <'+file_len+'} {: >4} {: >4} {: >4} {: >5}x{: >5}').format(src[c][0],src[c][1],"%.2f"%src[c][5],
"%.1f" % src[c][2],src[c][3],src[c][4]))
if opts.viewer:
fnames=[x[0] for x in src]
subprocess.call(opts.viewer.replace('%f', " ".join(fnames)), shell=True)
return
def find_fingerprint_nearest(sqlfile,cmp):
''' Find nearest match '''
cmp=os.path.abspath(cmp)
conn=sqlite3.connect(sqlfile)
def find_fingerprint_nearest(opts):
''' Find nearest match to given file '''
cmp=os.path.abspath(opts.similarity)
conn=sqlite3.connect(opts.sqlfile)
conn.text_factory=str
db1=conn.cursor()
if is_listed(db1, cmp):
@@ -377,6 +405,10 @@ def find_fingerprint_nearest(sqlfile,cmp):
print( ('{: <'+file_len+'} {: >4} {: >4} {: >4} {: ^5}x{: ^5}').format("File","Diff","CDel","Shrp","Wth","Hgt"))
print( ('{: <'+file_len+'} {: >4} {: >4} {: >4} {: >5}x{: >5}').format(cmp,"","","%.1f" % sp,dims[0],dims[1]))
print( ('{: <'+file_len+'} {: >4} {: >4} {: >4} {: >5}x{: >5}').format(this[0], this[1],"%.2f"%this[5], "%.1f" % this[2],this[3], this[4]))
if opts.viewer:
subprocess.call(opts.viewer.replace('%f', " ".join((cmp,this[0]))), stderr=subprocess.STDOUT, shell=True)
def append_sharpness(sqlfile):
conn=sqlite3.connect(sqlfile)
@@ -511,7 +543,7 @@ def print_structure(files):
i+=1
return
def print_dup_structure(files):
def print_dup_structure(files,opts):
i=1
for hash in files:
#print(hash[0])
@@ -519,6 +551,9 @@ def print_dup_structure(files):
for f in hash[1]:
fnames.append(' "'+f[0]+'"')
print "%(i)d:%(n)d:%(f)s " % {'i':i, 'n':len(fnames), 'f':",".join(fnames)}
if opts.viewer:
fnames=[x[0] for x in hash[1]]
subprocess.call(opts.viewer.replace('%f', " ".join(fnames)), shell=True)
i+=1
return
@@ -548,12 +583,12 @@ def main():
find_color_nearest(options.sqlfile,options.nearestcolor)
if options.similarity!=None:
if os.path.exists(options.similarity):
find_fingerprint_nearest(options.sqlfile,options.similarity)
find_fingerprint_nearest(options)
else:
find_fingerprint_similar(options.sqlfile,int(options.similarity))
find_fingerprint_similar(options)
if options.duplicate:
files=find_duplicates(options.sqlfile,options.startpath)
print_dup_structure(files)
print_dup_structure(files,options)
if options.searchsmall:
files=find_smalls(options.minsize,options.sqlfile)
if options.deleteFiles: