#=============================================================================
# Copyright 2010-2014 Fallgatter Technologies, Inc. All Rights Reserved.
#
# This software is the confidential and proprietary information of Fallgatter
# Technologies, Inc. ("Confidential Information"). You shall not
# disclose such Confidential Information and shall use it only in
# accordance with the terms of the license agreement you entered into
# with Fall-tech.
#
# FALL-TECH MAKES NO REPRESENTATIONS OR WARRANTIES ABOUT THE SUITABILITY OF THE
# SOFTWARE, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
# IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
# PURPOSE, OR NON-INFRINGEMENT. FALL-TECH SHALL NOT BE LIABLE FOR ANY DAMAGES
# SUFFERED BY LICENSEE AS A RESULT OF USING, MODIFYING OR DISTRIBUTING
# THIS SOFTWARE OR ITS DERIVATIVES.
#=============================================================================
# This script parses through a given folder,
# and find the most recent NxM images and creates a web page showing
# them all off in an HTML file
#
# If there are more than NxM images in the folder, we'll delete the oldest
# ones.
#
# Place the script and the contents of the img folder in the same folder.
# Feel free to place all of this outside the public_html, just will have to
# set the -diOut setting here to correctly reference the public_html
#
# @author KS
# @date 02.08.12
# @updated 03.04.13
# @version 0.141
#
# imports
import argparse, os, glob, datetime, time, shutil, re
import Image, ImageFont, ImageDraw, PngImagePlugin
#
# Parameters for text generation of the web-page. These should be
# a little easier to change as we need them.
#
PAGE_TYPE_LIVE = 0;
PAGE_TYPE_SHOWCASE = 1;
PAGE_TYPE_TRIPTYCH = 2;
#
# define header text for each of the above (in the same order)
#
# title of the page
PAGE_TITLES = [ "MASC Showcase: Snowflakes in Freefall",
"MASC Showcase: Snowflakes in Freefall",
"MASC Showcase: Triplet Images of Snowflakes in Freefall"
]
# what to show before the table of images (no need to use to set width of document any more)
PAGE_PREAMABLE = [
# live
"""
This is a live showcase of snowflake images captured in freefall at the School of Marine and Atmospheric Sciences at Stony Brook University using the MASC (Multi Angle Snowflake Camera), which was developed at the University of Utah (Tim Garrett). This is part of a National Science Foundation project between Stony Brook (Dr. Brian Colle) and North Carolina State University (Dr. Sandra Yuter and Dr. Matthew Miller). The data is used to understand the microphysical evolution within East Coast winter storms. For more information about the MASC, please visit the SnowflakeStereography and Fallspeed home page.
Images are taken at f/5.6 with an exposure of 1/25,000th of a second using
a 1.2MP industrial camera with a 16 mm lens. The image width is 33 mm (1 1/3 inches).
Images are taken at f/5.6 with an exposure of 1/25,000th of a second using
a 1.2MP industrial camera with a 16 mm lens. The image width is 33 mm (1 1/3 inches).
Images are taken at f/5.6 with an exposure of 1/25,000th of a second using
a 1.2MP industrial camera with a 16 mm lens. The image width is 33 mm (1 1/3 inches).
Click on any snowflake to see it in full resolution. Note that all times are in EST.
"""
]
# the footer of the page (after "Updated on:")
PAGE_FOOTER = """
Copyright, Tim Garrett,
Konstantin Shkurko,
Cale Fallgatter,
Daniel \"Howie\" Howlett, University of Utah, """ + datetime.datetime.now().strftime( "%Y" ) + ""
#
# try to parse out the date/time from file name
# we assume that the filename has the following
# organization: "%Y.%m.%d_%H.%M.%S_",
# where %<> corresponds to the Python time
#
# @param inFile - input file name (string)
# @return datetime, if name parsing succeeded, otherwise - False
#
def parseDateTimeFromFileName( inFile ):
# try splitting our filename
try:
tmp = os.path.basename( inFile )
tmp = os.path.splitext( tmp )[0]
tmp = tmp.split( '_' )
except:
return False
# try parsing 1st item as date
try:
#print( "parsed time: " + tmp[0] + ", " + tmp[1] )
d = time.strptime( tmp[0]+" "+tmp[1], "%Y.%m.%d %H.%M.%S" )
return time.mktime( d )
except:
return False
#
# Parses file name of our snowflake, splits it based on '_'
# and returns the camera id of the image.
# For now assume %d_%T_flake_%f_cam_%c.png format (camId=5):
# +%f flake id
# +%c camera id
# +%d date (yyyy.mm.dd) %Y - year, %M - month, %D - day
# +%T time (hh.mm.ss) %h - hour, %m - minute, %s - second
#
# @param inFile - input file name to parse
# @param camId - id of the piece separated by '_' which corresponds to
# the camera id (first is 0). If not passed in, return
# full split of filename
# @return camera id of given image filename
#
def parseFileName( inFile, camId ):
fileName = os.path.splitext( os.path.split( inFile )[1] )[0]
#splitStr = fileName.split( '_' )
#dt = splitStr[0]
#tm = splitStr[1]
#fk = splitStr[3]
#cm = splitStr[5]
#return [cm, fk, dt, tm, fk]
if( camId==None ):
return fileName.split( '_' )
else:
return fileName.split( '_' )[camId]
#
# Gets the PNG images in the directory sorted by modified date
#
# The date is extraced in the following order (until one succeeds)
# 1. Based on the "Creation Time" within image meta-data
# 2. Using parseDateTimeFromFileName function above
# 3. Using the os file system, file creation time (Windows) or
# last metadata modification time (*nix)
#
# @param dir - directory which to check (default ".")
# @return list of PNG filenames sorted on timestamp in descending order
#
def getImagesInDir( dir = "." ):
if( not os.path.exists( dir ) ):
print "ERROR: output path (" + dir + ")doesn't exist!"
# init file info
outInfo = [];
# open the directory
files = glob.glob( os.path.join( dir, '*.png' ) );
for file in files:
# get image tags to resave later (and parse out date)
img = Image.open( file )
imgTags = img.info
if( imgTags.has_key('Creation Time') ):
imgCTime = time.strptime( imgTags['Creation Time'], "%d %b %Y %H:%M:%S +0000" )
outInfo.append( [file, time.mktime( imgCTime )] )
#print "found: {0}, {1}".format( file, fileTimestamp )
else:
tmp = parseDateTimeFromFileName( file )
if( tmp != False ):
outInfo.append( [file, tmp] )
#print "parse: {0}, {1}".format( file, tmp )
else:
fInfo = os.stat( file )
outInfo.append( [file, fInfo.st_ctime] )
#print "NOT: {0}, {1}".format( file, fInfo.st_ctime )
del img
# sort in descending order based on created timestamp (first NxM images)
outInfo.sort( key=lambda tmp: tmp[1], reverse=True )
return outInfo
#
# Searches for images in a directory which fit a given naming convention
# while replacing camera ids...
# The basic idea is to match in the following order: exact, up to 1s seconds,
# up to 10s of seconds. All of the matches are accumulated into a list
# that is returned.
#
# @param inDir - input directory to check
# @param inBaseName - basename of the image to check with
# @param inCamIds - camera ID to check. Can be None, , list
# + None: matches to the camera ID in the basename
# + : matches to given camera ID as this int
# + list: matches to given list of ids as int
# @param inCamIDIndex - index of camera ID in file name, counted between _, starting at 0
# @return list of file basenames that fit
#
def findImageFileInFolder( inDir, inBaseName, inCamIds, inCamIDIndex ):
# get basename and such so we can mess with it
thisFilesSplit = parseFileName( inBaseName, None )
parsedTime = thisFilesSplit[1].split('.')
if( inCamIds==None ):
inCamIds = [thisFilesSplit[inCamIDIndex][0]]
# check if list... if not, listify
if( type(inCamIds)!=type(list()) ):
inCamIds = [inCamIds]
# run through all camera IDs
foundNames = []
for c in inCamIds:
# check full name first
tmpFilesSplit = list(thisFilesSplit)
tmpFilesSplit[inCamIDIndex] = str(c)
tmpName = "_".join(tmpFilesSplit) + ".png"
if( os.path.isfile( os.path.join( inDir, tmpName ) ) ):
foundNames.append(tmpName)
del tmpFilesSplit
continue
# check to 1s of seconds
else:
tmpTime = list(parsedTime)
tmpTime[2] = tmpTime[2][0] + '([0-9]+)(\.*)([0-9]*)'
if( len(tmpTime)>3 ):
tmpFilesSplit[1] = '.'.join(tmpTime[0:-1])
else:
tmpFilesSplit[1] = '.'.join(tmpTime)
tmpName = "_".join(tmpFilesSplit) + ".png"
namesBefore = list( [os.path.basename(f1) for f1 in os.listdir(inDir) if re.search(tmpName, f1)] )
if( len(namesBefore)>0 ):
foundNames.append(namesBefore[0])
del tmpTime, tmpFilesSplit, namesBefore
continue
# check to 10s of seconds
else:
tmpTime[2] = '([0-9]+)(\.*)([0-9]*)'
if( len(tmpTime)>3 ):
tmpFilesSplit[1] = '.'.join(tmpTime[0:-1])
else:
tmpFilesSplit[1] = '.'.join(tmpTime)
tmpName = "_".join(tmpFilesSplit) + ".png"
namesBefore = list( [os.path.basename(f1) for f1 in os.listdir(inDir) if re.search(tmpName, f1)] )
if( len(namesBefore)>0 ):
foundNames.append(namesBefore[0])
del tmpTime, tmpFilesSplit, namesBefore
continue
del tmpTime
# done
del tmpFilesSplit
del thisFilesSplit, parsedTime
return foundNames
#
# Generates the output file based on all passed in parameters. See the README file
# which describes in detail meaning of each parameter in greater detail
#
# @param inDir - input directory with images (this is where original image files are located,
# and then moved to output directory. Original images are NOT modified)
# @param outDir - output directory with images (this is where HTML file will be placed)
# If inDir is NOT passed in, this directory will house the original images
# which will be modified in place
# @param outFile - name of the HTML file
# @param N - N in NxM matrix of thumbnails
# @param M - M in NxM matrix of thumbnails
# @param iw - image width of the thumbnail (generated from original image)
# @param ih - image height of the thumbnail (generated from original image)
# @param rTime - refresh time (milliseconds) for the web page. If 0, then page will not have
# any refreshing and the introductory text is different. Also no
# image timestamps will be shown underneath thumbnails.
# Otherwise, page will refresh in rTime. Time stamps will be placed
# underneath thumbnails.
# NOTE: use rTime=0 for gallery, rTime>0 for live feed.
# @param tripFlg - flag whether to generate a triptych web page. It will trump
# time flag, etc.
# @param nolineFlg - flag whether to remove text and lines underneath the images
# @param camIdx - index of the camera ID in the file name (as passed to parseFileName)
#
def genOutputHTML( inDir, outDir, outFile, N, M, iw, ih, rTime, tripFlg, nolineFlg, camIdx ):
widthPerImgCell = 30
if( nolineFlg ):
widthPerImgCell = 0
reloadTimeInMSec = rTime;
bodyWidth = (widthPerImgCell + iw)*M + 15
# determine which type of page we're generating based on some inputs
pageTypeToGenerate = PAGE_TYPE_LIVE;
# triptych
if( tripFlg ):
pageTypeToGenerate = PAGE_TYPE_TRIPTYCH
reloadTimeInMSec = 0
M = 6
N = 1000000
bodyWidth = (widthPerImgCell + iw)*6 + widthPerImgCell + (int)(iw/2) + 15
# LIVE
elif( reloadTimeInMSec!=0 ):
pageTypeToGenerate = PAGE_TYPE_LIVE
else:
# triptych
if( tripFlg ):
pageTypeToGenerate = PAGE_TYPE_TRIPTYCH
reloadTimeInMSec = 0
M = 6
N = 1000000
bodyWidth = (widthPerImgCell + iw)*6 + widthPerImgCell + (int)(iw/2) + 15
# showcase
else:
pageTypeToGenerate = PAGE_TYPE_SHOWCASE
# generate the header for the file
outStr = """
"""
# add page title
outStr += ("%(s0)s" % { 's0':PAGE_TITLES[pageTypeToGenerate] } )
# body onload
if( reloadTimeInMSec!=0 ):
outStr += ("" % {'s0':str(reloadTimeInMSec), 's1':str(bodyWidth)})
else:
outStr += ("" % {'s0':str(bodyWidth)})
# add page header and pre-amble text
outStr += ("
%(s0)s
%(s1)s" % { 's0':PAGE_TITLES[pageTypeToGenerate], 's1':PAGE_PREAMABLE[pageTypeToGenerate] })
# add automagic reloading flag
if( reloadTimeInMSec!=0 ):
outStr += """Auto Refresh (in """ + str(reloadTimeInMSec/1000) + """ sec): Off |
On
""";
# compare the input and output directories, only if inDir is set!
# if different, copy only images that aren't in outDirectory
if( inDir != "" ):
fullInPath = os.path.abspath( inDir )
fullOutPath = os.path.abspath( outDir )
if( not os.path.exists( fullInPath ) ):
print "ERROR: input path (" + fullInPath + ")doesn't exist!"
if( not os.path.exists( fullOutPath ) ):
print "ERROR: output path (" + fullOutPath + ")doesn't exist!"
if( fullInPath != fullOutPath ):
filesIn = glob.glob( os.path.join( fullInPath, '*.png' ) )
filesOut = glob.glob( os.path.join( fullOutPath, '*.png' ) )
filesIn1 = [os.path.basename(f) for f in filesIn]
filesOut1 = [os.path.basename(f) for f in filesOut]
del filesIn, filesOut
filesToMove = list( set(filesIn1) - set(filesOut1) )
for f in filesToMove:
if( pageTypeToGenerate!=PAGE_TYPE_TRIPTYCH ):
# try copy for now, but might not work quite too well...
bname = os.path.basename( f )
#print "- copying file " + f + " to " + os.path.join( fullOutPath, bname )
shutil.copy2( f, os.path.join( fullOutPath, bname ) )
# if we are in triptych mode, we must rescale images to largest one during "copy"
else:
# check if this filename already moved
bname = os.path.basename( f )
# find all the files that fit
stuffFound = findImageFileInFolder( fullOutPath, bname, None, camIdx )
if( len(stuffFound)>0 ):
del stuffFound
continue
# check if input actually has a tripplet
stuffFound = findImageFileInFolder( fullInPath, bname, [0,1,2], camIdx )
if( len(stuffFound)<3 ):
print "ERROR: we can't find a triplet using file: " + bname
print "Instead we found: "
print stuffFound
del stuffFound
continue
# parse out the file name, then get snowflake's sizes!
# NOTE: we'll normalize the time to the 1st found snowflake
# as in. Let's say we have 3 flake images:
# - 2013.02.06_14.14.20.123456_flake_3_cam_0.png
# - 2013.02.06_14.14.20.777777_flake_3_cam_1.png
# - 2013.02.06_14.14.21.111111_flake_3_cam_2.png
# the resulting (copied) 3 images will be
# - 2013.02.06_14.14.20.123456_flake_3_cam_0.png
# - 2013.02.06_14.14.20.123456_flake_3_cam_1.png
# - 2013.02.06_14.14.20.123456_flake_3_cam_2.png
thisFilesSplit = parseFileName( f, None )
# get ith camera
#namesBefore = [None,None,None]
namesBefore = findImageFileInFolder( fullInPath, f, [0,1,2], camIdx )
namesAfter = [None,None,None]
img = [None,None,None]
ws = [None,None,None]
hs = [None,None,None]
hMax = -1
wMax = -1
for c in range(0,3):
thisFilesSplit[camIdx] = str(c)
tmpName = "_".join(thisFilesSplit) + ".png"
namesBefore[c] = os.path.join( fullInPath, namesBefore[c] )
namesAfter[c] = os.path.join( fullOutPath, tmpName )
#print "looking at: " + namesBefore[c] + ", going to : " + namesAfter[c]
img[c] = Image.open( namesBefore[c] )
w, h = img[c].size
ws[c] = w
hs[c] = h
if( h > hMax ):
hMax = h
wMax = w
# get max height and rescale images proportionally if need be,
# saving them to output
for c in range(0,3):
# max image -> just copy
if( ws[c]==hMax ):
bname = os.path.basename( namesAfter[c] )
shutil.copy2( namesBefore[c], os.path.join( fullOutPath, bname ) )
continue
# other image -> resize
tmpw = wMax #ws[c]
tmph = hMax #hs[c]
#if( ws[c] > hs[c] ):
# tmph = int( tmpw * float( hs[c] ) / ws[c] )
#else:
tmpw = int( tmph * float( ws[c] ) / hs[c] )
#print "--- original size (" + str(ws[c]) + ", " + str(hs[c]) + "), saving to (" + str(tmpw) +", " + str(tmph) +"), max size: (" + str(wMax) + ", " + str(hMax) + ")"
imgTags = img[c].info
#img2 = img.convert("RGB")
img2 = img[c].resize( (tmpw, tmph), Image.ANTIALIAS )
reservedTags = ('interlace', 'gamma', 'dpi', 'transparency', 'aspect', 'icc_profile' )
meta = PngImagePlugin.PngInfo()
mc = 0
for k,v in imgTags.iteritems():
if k in reservedTags:
meta.add( k, v )
continue
meta.add_text( k, v, 0 )
#print "found meta {0}".format( k )
mc = mc+1
saveName = os.path.join( fullOutPath, os.path.basename( namesAfter[c] ) )
if( mc>0 ):
img2.save( saveName, "PNG", pnginfo=meta )
else:
img2.save( saveName, "PNG" )
del img2, meta, imgTags, reservedTags
#for f in namesBefore:
# print f
#print "after:"
#for f in namesAfter:
# print f
#print "----------------------------"
# clean up!
del namesBefore, namesAfter, img, ws, hs
# get images in folder + clean up the oldest ones
numImgs = N*M
imgs = getImagesInDir( outDir )
if( len(imgs)>numImgs ):
for i in range(numImgs, len(imgs) ):
os.remove( imgs[i][0] )
os.remove( os.path.splitext(imgs[i][0])[0] + "_s.jpg" )
#print "- removing file " + imgs[i][0]
imgs = imgs[0:numImgs]
# define font for overlay (just in case font I got is malicious, use it as last resort)
fontSizeB = 16
fontSizeS = 8
try:
overFontB = ImageFont.truetype( "Verdana.ttf", fontSizeB )
overFontS = ImageFont.truetype( "Verdana.ttf", fontSizeS )
except:
overFontB = ImageFont.truetype( "img/Verdana.ttf", fontSizeB )
overFontS = ImageFont.truetype( "img/Verdana.ttf", fontSizeS )
copyMsg = "Tim Garrett, University of Utah, "
reservedTags = ('interlace', 'gamma', 'dpi', 'transparency', 'aspect', 'icc_profile' )
# get two images for overlay - larger versions
#overNSFL = Image.open( 'img/ssl_NSF_desat.png' )
##overNSFL = Image.open( 'img/ssl_NSF.png' )
#overALTAL = Image.open( 'img/ssl_ALTA.png' )
#overNSFLw, overNSFLh = overNSFL.size
#overALTALw, overALTALh = overALTAL.size
#overCombWL = overNSFLw + overALTALw + 5
# get two images for overlay - smaller versions
overNSF = Image.open( 'img/ss_NSF_desat.png' )
#overNSF = Image.open( 'img/ss_NSF.png' )
overALTA = Image.open( 'img/ss_ALTA.png' )
overNSFw, overNSFh = overNSF.size
overALTAw, overALTAh = overALTA.size
overCombW = overNSFw + overALTAw + 5
# insert the table with images
c = 1 # image counter
lImgs = len(imgs) # total number of images
locTriplet = 0 # triplet counter (used for triptych page only!)
# do we need to include space around images and lines/writing underneath them
if( nolineFlg ):
outStr += "
\n\t
\n"
else:
outStr += "
\n\t
\n"
for f in imgs:
#print "- file: " + f[0]
# get the correct image dimensions
tmpw = iw
tmph = ih
relp = f[0]
thumbFileInfo = os.path.splitext( f[0] )
trelp= thumbFileInfo[0] + "_s.jpg";
img = Image.open( f[0] )
w, h = img.size
if( w>iw or h>ih ):
if( w > h ):
tmph = int( tmpw * float( h ) / w )
else:
tmpw = int( tmph * float( w ) / h )
else:
tmpw = w
tmph = h
# do we need to generate a thumbnail (and create overlay)?
# if need to do thumbnail -> newly added image
timestp = datetime.datetime.fromtimestamp( f[1] )
if( os.path.isfile( trelp )==False ):
imgTags = img.info
img2 = img.convert("RGB")
img3 = img2.copy()
# add copyright text + sponsor logos
yearStr = timestp.strftime( '%Y' )
draw = ImageDraw.Draw( img2 )
tmpStr = copyMsg + yearStr
txtBW, txtBH = overFontB.getsize( tmpStr )
# - largest logo + large font
#if( w-overCombWL-10-txtBW > 0 ):
# img2.paste( overNSFL, (w-overNSFLw,h-overNSFLh) )
# img2.paste( overALTAL, (w-overCombWL,h-overALTALh) )
# txtDim = (w-txtBW-10-overCombWL, (int)(h-1.5*fontSizeB))
# draw.text( txtDim, tmpStr, (255,255,255), font=overFontB )
# - small logo + large font
#elif( w-overCombW-10-txtBW > 0 ):
if( w-overCombW-10-txtBW > 0 ):
img2.paste( overNSF, (w-overNSFw,h-overNSFh) )
img2.paste( overALTA, (w-overCombW,h-overALTAh) )
txtDim = (w-txtBW-10-overCombW, (int)(h-1.5*fontSizeB))
draw.text( txtDim, tmpStr, (255,255,255), font=overFontB )
# label in top-left corner camera id
if( pageTypeToGenerate==PAGE_TYPE_TRIPTYCH ):
imgCamId = parseFileName( relp, camIdx )
imgCamIdStr = str(imgCamId)
txtBW2, txtBH2 = overFontB.getsize( imgCamIdStr )
txtDim2 = (txtBW2+5, fontSizeB)
draw.text( txtDim2, imgCamIdStr, (255,255,255), font=overFontB )
# - no logo + large font
elif( w-txtBW-5 > 0 ):
txtDim = (w-txtBW-5, (int)(h-1.5*fontSizeB))
draw.text( txtDim, tmpStr, (255,255,255), font=overFontB )
# label in top-left corner camera id
if( pageTypeToGenerate==PAGE_TYPE_TRIPTYCH ):
imgCamId = parseFileName( relp, camIdx )
imgCamIdStr = str(imgCamId)
txtBW2, txtBH2 = overFontB.getsize( imgCamIdStr )
txtDim2 = (txtBW2, fontSizeB)
draw.text( txtDim2, imgCamIdStr, (255,255,255), font=overFontB )
# - no logo + small font
else:
txtSW, txtSH = overFontS.getsize( tmpStr )
draw.text( (w-txtSW-5,h-1.5*fontSizeS), tmpStr, (255,255,255), font=overFontS )
# label in top-left corner camera id
if( pageTypeToGenerate==PAGE_TYPE_TRIPTYCH ):
imgCamId = parseFileName( relp, camIdx )
imgCamIdStr = str(imgCamId)
txtSW2, txtSH2 = overFontS.getsize( imgCamIdStr )
txtDim2 = (txtSW2, fontSizeS)
draw.text( txtDim2, imgCamIdStr, (255,255,255), font=overFontS )
meta = PngImagePlugin.PngInfo()
mc = 0
for k,v in imgTags.iteritems():
if k in reservedTags:
meta.add( k, v )
continue
meta.add_text( k, v, 0 )
#print "found meta {0}".format( k )
mc = mc+1
if( mc>0 ):
img2.save( thumbFileInfo[0] + ".png", "PNG", pnginfo=meta )
else:
img2.save( thumbFileInfo[0] + ".png", "PNG" )
# save full res + thumbnail (no overlay)
img3.thumbnail( (tmpw,tmph), Image.ANTIALIAS )
#draw2 = ImageDraw.Draw( img3 )
#txtSW, txtSH = overFontS.getsize( tmpStr )
#draw2.text( (tmpw-txtSW-5,tmph-1.5*fontSizeS), tmpStr, (255,255,255), font=overFontS )
img3.save( trelp, "JPEG" )
#del draw, img2, draw2, img3, meta, imgTags
del draw, img2, img3, meta, imgTags
# note, we'll keep the time stamp only for the live feed generation
# LIVE
if( reloadTimeInMSec!=0 ):
if( nolineFlg ):
outStr += ("\t\t
![\"NSF\"](\"img/ss_NSF.png\" "\\"NSF\\"/")
![\"Alta\"](\"img/ss_ALTA.png\" "\\"Alta\\"/")