class FlickrQuery():
def __init__(self,flickrapikey,flickrSecret):
self.api_key = flickrapikey
self.fapi = FlickrAPI(flickrapikey, flickrSecret)
def searchbyid(self,eventid):
photolist = []
query = " prefix lode: <http://linkedevents.org/ontology/> \n\
prefix dc: <http://purl.org/dc/elements/1.1/> \n\
prefix ma: <http://www.w3.org/ns/ma-ont#> \n\
SELECT ?event ?eventTitle ?URI \n\
WHERE { \n\
?event dc:title ?eventTitle. \n\
?photo lode:illustrate ?event. \n\
?photo ma:locator ?URI. \n\
FILTER (?event = <http://data.linkedevents.org/event/eventURI>). \n\
} \n\
"
query = query.replace('eventURI',eventid)
searchbase = 'http://eventmedia.eurecom.fr/sparql'
params = urllib.urlencode({"format": "application/sparql-results+json", "query": query})
f = urllib.urlopen(searchbase + '?' + params)
results = simplejson.load(f)
try:
results = results['results']['bindings']
except:
return []
for result in results:
url = result['URI']['value']
photolist.append(url)
return photolist
def searchbytitle(self,title,time,eventid):
photolist = []
t1 = time
t2 = t1 + datetime.timedelta(days = 5)
starttime = t1 + datetime.timedelta(hours = - (t1.hour))
bReturn = 1
idx = 1
while (bReturn ==1):
try:
rsp = self.fapi.photos_search(api_key=self.api_key,
ispublic="1",
media="photos",
per_page="250",
page=str(idx),
min_taken_date = str(starttime),
max_taken_date = str(t2),
text = title.encode('utf-8'),
extras = 'date_upload, date_taken, owner_name, geo, tags, machine_tags, url_m'
)
idx = idx +1
self.fapi.testFailure(rsp)
total_images = rsp.photos[0]['total'];
null_test = int(total_images);
except:
null_test = 0
print sys.exc_info()[0]
print sys.exc_info()[1]
print ('Exception encountered while querying title for images\n')
print type(title),type(title.encode('utf-8'))
if null_test == 0:
break
if null_test >=250*(idx-1):
bReturn = 1
else:
bReturn = 0
tmpdir = os.path.join(gconfig.metadatadir,'querybytitle')
if not os.path.exists(tmpdir):
os.makedirs(tmpdir)
metadata = os.path.join(tmpdir,'%s_%d.xml' % (eventid,idx-1))
data = parseString(rsp.xml)
if not os.path.exists(metadata):
f = open(metadata,'w')
f.write(data.toprettyxml(encoding='UTF-8'))
f.close()
q = data.getElementsByTagName('photo')
for p in q:
url = p.getAttribute('url_m')
if url.find('.jpg')>0:
photolist.append(url)
return photolist
def searchbygeo(self,lat,lng,time,eventid):
photolist = []
photolist = []
t1 = time
t2 = t1 + datetime.timedelta(days = 3)
starttime = t1 + datetime.timedelta(hours = - (t1.hour))
bReturn = 1
idx = 1
while (bReturn ==1):
try:
rsp = self.fapi.photos_search(api_key=self.api_key,
ispublic="1",
media="photos",
per_page="250",
page=str(idx),
min_taken_date = str(starttime),
max_taken_date = str(t2),
lat = str(lat),
lon = str(lng),
radius = str('0.7'),
accuracy = '12',
extras = 'date_upload, date_taken, owner_name, geo, tags, machine_tags, url_m'
)
idx = idx +1
self.fapi.testFailure(rsp)
total_images = rsp.photos[0]['total'];
null_test = int(total_images);
except:
null_test = 0
print sys.exc_info()[0]
print sys.exc_info()[1]
print ('Exception encountered while querying for images\n')
if null_test == 0:
break
if null_test >=250*(idx-1):
bReturn = 1
else:
bReturn = 0
tmpdir = os.path.join(gconfig.metadatadir,'querybygeo')
if not os.path.exists(tmpdir):
os.makedirs(tmpdir)
metadata = os.path.join(tmpdir,'%s_%d.xml' % (eventid,idx-1))
data = parseString(rsp.xml)
if not os.path.exists(metadata):
f = open(metadata,'w')
f.write(data.toprettyxml(encoding='UTF-8'))
f.close()
q = data.getElementsByTagName('photo')
for p in q:
url = p.getAttribute('url_m')
if url.find('.jpg')>0:
photolist.append(url)
return photolist
def searchbygeoRadius(self,lat,lng,r,stime,etime, eventid):
photolist = []
bReturn = 1
idx = 1
while (bReturn ==1):
try:
rsp = self.fapi.photos_search(api_key=self.api_key,
ispublic="1",
media="photos",
per_page="250",
page=str(idx),
min_taken_date = str(stime),
max_taken_date = str(etime),
lat = str(lat),
lon = str(lng),
radius = str(r),
accuracy = '12',
extras = 'date_upload, date_taken, owner_name, geo, tags, machine_tags, url_m'
)
idx = idx +1
self.fapi.testFailure(rsp)
total_images = rsp.photos[0]['total'];
null_test = int(total_images);
except:
null_test = 0
print sys.exc_info()[0]
print sys.exc_info()[1]
print ('Exception encountered while querying for images\n')
if null_test == 0:
break
if null_test >=250*(idx-1):
bReturn = 1
else:
bReturn = 0
tmpdir = os.path.join(gconfig.metadatadir,'querybygeo')
if not os.path.exists(tmpdir):
os.makedirs(tmpdir)
metadata = os.path.join(tmpdir,'%s_%d.xml' % (eventid,idx-1))
data = parseString(rsp.xml)
if not os.path.exists(metadata):
f = open(metadata,'w')
f.write(data.toprettyxml(encoding='UTF-8'))
f.close()
q = data.getElementsByTagName('photo')
for p in q:
url = p.getAttribute('url_m')
if url.find('.jpg')>0:
photolist.append(url)
return photolist
def outputlist(self,list,id,fname):
fw = open(fname,'w')
for url in list:
fname = url.split('/')[-1]
fw.write('%s\%s\n' % (id,fname))
fw.close()
def geturlbyid(self,id,list):
photos = {}
results = []
for p in list:
t = p.split('/')[-1]
t = t.replace('.jpg','')
photos[t] = p
for idx in id:
if idx in photos:
results.append(photos[idx])
return results
def OutputList(self,listname,lst):
str = ''
str += '\n<table align="center" border="1" cellspacing="1" cellpadding="3" width=800><H2>query by %s</H2><tr>' % listname
N = len(lst)
num = 0
for i in range(0,N):
img_file = lst[i]
str += '\n<td align="center" valign=top width=30><IMG SRC="%s" width=160 border=1 /></td>' % img_file
num = num +1
if (num % 8)==0:
str +=('</tr>')
str +=('</table>')
return str
def OutputJson(self,lst):
N = len(lst)
tmp = []
num = 0
for i in range(0,N):
img_file = lst[i]
tmp.append({'photo':img_file})
mydict = {}
mydict['photos'] = tmp
mydict['number'] = str(len(tmp))
return mydict
def OutputHtml(self,id,idlist,titlelist,geolist,refinelist):
file = open(gconfig.outputdir + '/' + '%s.html' % id,'w')
head='<html><head><title> Media illustration </title></head>'
file.write('%s\n' % head)
head='<body BGCOLOR="#FFFFFF"><center><H1>enriching for Event %s</H1><HR HSIZE="50%%"/>' % id
file.write('%s\n'% head)
file.write(self.OutputList("machine tag",idlist) )
file.write(self.OutputList("Geo tag",geolist) )
file.write(self.OutputList("title + pruning + refine",refinelist) )
file.close()
def OutputXML(self,id,idlist,titlelist,geolist,refinelist):
fname = gconfig.outputdir + '/' + '%s.xml' % id
setid = set(idlist)
setgeo = set(geolist) - setid
settitle = set(titlelist) - setid
setrefinelist = set(refinelist) - setid
sets = [setid,setgeo,settitle,setrefinelist]
tmpinfo = ["query by ID","query by Geo - ID", "query by Title - ID"," Pruning and Refine"]
doc = Document()
query = doc.createElement("query")
query.setAttribute("id",id)
doc.appendChild(query)
results = doc.createElement("PhotoSets")
for tmpset,info in zip(sets,tmpinfo):
photoset = doc.createElement("photoset")
photoset.setAttribute('query',info)
photoset.setAttribute('photoNum',str(len(tmpset)))
for photo in tmpset:
ph = doc.createElement("photo")
ph.setAttribute('url', photo)
photoset.appendChild(ph)
results.appendChild(photoset)
query.appendChild(results)
f = open(fname, "w")
f.write(doc.toprettyxml(encoding='UTF-8'))
f.close()
upper_bound = mintime + timeskip * 20 #upper bound of the upper time limit
maxtime = .95 * lower_bound + .05 * upper_bound
print '\nBinary search on time range upper bound'
print 'Lower bound is ' + str(datetime.fromtimestamp(lower_bound))
print 'Upper bound is ' + str(datetime.fromtimestamp(upper_bound))
keep_going = 6 #search stops after a fixed number of iterations
while( keep_going > 0 and maxtime < endtime):
try:
rsp = fapi.photos_search(api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250",
page="1",
has_geo = "1", #bbox="-180, -90, 180, 90",
text=query_string,
accuracy="6", #6 is region level. most things seem 10 or better.
min_upload_date=str(mintime),
max_upload_date=str(maxtime))
##min_taken_date=str(datetime.fromtimestamp(mintime)),
##max_taken_date=str(datetime.fromtimestamp(maxtime)))
#we want to catch these failures somehow and keep going.
time.sleep(1)
fapi.testFailure(rsp)
total_images = rsp.photos[0]['total'];
null_test = int(total_images); #want to make sure this won't crash later on for some reason
null_test = float(total_images);
print '\nnumimgs: ' + total_images
print 'mintime: ' + str(mintime) + ' maxtime: ' + str(maxtime) + ' timeskip: ' + str(maxtime - mintime)
counter = -1
while( pagenum <= num_visit_pages ):
#for pagenum in range(1, num_visit_pages + 1): #page one is searched twice
print(' page number ' + str(pagenum))
try:
print("PAGE")
print(pagenum)
# WARNING THIS QUERY HAS TO MATCH THE SEARCH QUERY!!!!
rsp = fapi.photos_search(api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250",
page=str(pagenum),
has_geo = "0",
text=query_string,
#extras = "tags, original_format, license, geo, date_taken, date_upload, o_dims, views",
#accuracy="6", #6 is region level.
min_upload_date=str(1121832000),#mintime),
max_upload_date=str(1192165200))#maxtime))
#rsp = fapi.photos_search(api_key=flickrAPIKey,
# ispublic="1",
# media="photos",
# per_page="250",
# page='0', #str(pagenum),
# sort="interestingness-desc",
# has_geo = "0", #bbox="-180, -90, 180, 90",
# text=query_string,
# #accuracy="6", #6 is region level. most things seem 10 or better.
def run_flickr_query_general(query_args, max_photos = 1000, startDate = "1/1/2010", endDate = "31/12/2011"):
socket.setdefaulttimeout(30) #30 second time out on sockets before they throw
#an exception. I've been having trouble with urllib.urlopen hanging in the
#flickr API. This will show up as exceptions.IOError.
#the time out needs to be pretty long, it seems, because the flickr servers can be slow
#to respond to our big searches.
###########################################################################
# Modify this section to reflect your data and specific search
###########################################################################
# flickr auth information:
# change these to your flickr api keys and secret
# make a new FlickrAPI instance
fapi = FlickrAPI(flickrAPIKey, flickrSecret)
#print '\n\nquery arguments\n'
#print query_args
total_images_queried = 0;
# number of seconds to skip per query
timeskip = 8 * 604800 #one week
starttime = convertDate(startDate)
#mintime = convertDate(startDate)
endtime = convertDate(endDate)
# maxtime = startime+20*timeskip
maxtime = endtime
mintime = endtime-10*timeskip
timeskip = min(timeskip, endtime-mintime)
print 'Start time: ' + str(datetime.fromtimestamp(starttime))
print 'End time: ' + str(datetime.fromtimestamp(endtime))
#this is the desired number of photos in each block
desired_photos = min(250, max_photos)
total_image_num = 0
results = {}
print 'here'
print starttime
print mintime
# while (maxtime < endtime):
while (starttime < mintime):
#new approach - adjust maxtime until we get the desired number of images
#within a block. We'll need to keep upper bounds and lower
#lower bound is well defined (mintime), but upper bound is not. We can't
#search all the way from endtime.
# lower_bound = mintime + 900 #lower bound OF the upper time limit. must be at least 15 minutes or zero results
# upper_bound = mintime + timeskip * 20 #upper bound of the upper time limit
# upper_bound = min(upper_bound, endtime)
# maxtime = .95 * lower_bound + .05 * upper_bound
lower_bound = mintime - 20 * timeskip #lower bound OF the upper time limit. must be at least 15 minutes or zero results
upper_bound = maxtime #upper bound of the upper time limit
lower_bound = max(lower_bound, starttime)
mintime = 0.05 * lower_bound + 0.95 * upper_bound
# print '\nBinary search on time range upper bound'
# print 'Lower bound is ' + str(datetime.fromtimestamp(lower_bound))
# print 'Upper bound is ' + str(datetime.fromtimestamp(upper_bound))
if total_image_num > max_photos:
print 'Number of photos %d > %d limit.' % (total_image_num, max_photos)
break
print 'here'
keep_going = 6 #search stops after a fixed number of iterations
while( keep_going > 0 and starttime < mintime):
#while( keep_going > 0 and maxtime < endtime):
try:
# print 'Calling api'
rsp = fapi.photos_search(api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250",
page="1",
#has_geo = "1", #bbox="-180, -90, 180, 90",
#text=query_string,
#accuracy="6", #6 is region level. most things seem 10 or better.
min_upload_date=str(mintime),
max_upload_date=str(maxtime),
**query_args)
##min_taken_date=str(datetime.fromtimestamp(mintime)),
##max_taken_date=str(datetime.fromtimestamp(maxtime)))
#we want to catch these failures somehow and keep going.
os_time.sleep(1)
fapi.testFailure(rsp)
print rsp
total_images = rsp.photos[0]['total'];
if total_images == '':
total_images = '0'
# print total_images
# print rsp.photos[0]
null_test = int(total_images); #want to make sure this won't crash later on for some reason
null_test = float(total_images);
# print 'numimgs: ' + total_images
# print 'mintime: ' + str(mintime) + ' maxtime: ' + str(maxtime) + ' timeskip: ' + str(maxtime - mintime)
if( int(total_images) > desired_photos ):
# print 'too many photos in block, increasing mintime'
lower_bound = mintime
mintime = (upper_bound + mintime) / 2 #midpoint between current value and lower bound.
# print 'too many photos in block, reducing maxtime'
# upper_bound = maxtime
# maxtime = (lower_bound + mintime) / 2 #midpoint between current value and lower bound.
if( int(total_images) < desired_photos):
# print 'too few photos in block, reducing mintime'
upper_bound = mintime
mintime = (lower_bound + mintime) / 2
# print 'too few photos in block, increasing maxtime'
# lower_bound = maxtime
# maxtime = (upper_bound + maxtime) / 2
# print 'Lower bound is ' + str(datetime.fromtimestamp(lower_bound))
# print 'Upper bound is ' + str(datetime.fromtimestamp(upper_bound))
if( int(total_images) > 0): #only if we're not in a degenerate case
keep_going = keep_going - 1
else:
upper_bound = upper_bound + timeskip;
except KeyboardInterrupt:
print('Keyboard exception while querying for images, exiting\n')
raise
# except:
# print sys.exc_info()[0]
#print type(inst) # the exception instance
#print inst.args # arguments stored in .args
#print inst # __str__ allows args to printed directly
# print ('Exception encountered while querying for images\n')
#end of while binary search
# print 'finished binary search'
print 'mintime: ' + str(datetime.fromtimestamp(mintime)) + ' maxtime: ' + str(datetime.fromtimestamp(maxtime)) + ' numimgs: ' + total_images
i = getattr(rsp,'photos',None)
if i:
s = 'numimgs: ' + total_images
print s
current_image_num = 1;
num = int(rsp.photos[0]['pages'])
s = 'total pages: ' + str(num)
print s
#only visit 16 pages max, to try and avoid the dreaded duplicate bug
#16 pages = 4000 images, should be duplicate safe. Most interesting pictures will be taken.
num_visit_pages = min(16,num)
s = 'visiting only ' + str(num_visit_pages) + ' pages ( up to ' + str(num_visit_pages * 250) + ' images)'
print s
total_images_queried = total_images_queried + min((num_visit_pages * 250), int(total_images))
#print 'stopping before page ' + str(int(math.ceil(num/3) + 1)) + '\n'
pagenum = 1;
while( pagenum <= num_visit_pages ):
#for pagenum in range(1, num_visit_pages + 1): #page one is searched twice
print ' page number ' + str(pagenum)
try:
rsp = fapi.photos_search(api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250",
page=str(pagenum),
sort="interestingness-desc",
#has_geo = "1", #bbox="-180, -90, 180, 90",
#text=query_string,
#accuracy="6", #6 is region level. most things seem 10 or better.
extras = "tags, original_format, license, geo, date_taken, date_upload, o_dims, views",
min_upload_date=str(mintime),
max_upload_date=str(maxtime),
**query_args)
##min_taken_date=str(datetime.fromtimestamp(mintime)),
##max_taken_date=str(datetime.fromtimestamp(maxtime)))
os_time.sleep(1)
fapi.testFailure(rsp)
except KeyboardInterrupt:
print('Keyboard exception while querying for images, exiting\n')
raise
# except:
# print sys.exc_info()[0]
# #print type(inst) # the exception instance
# #print inst.args # arguments stored in .args
# #print inst # __str__ allows args to printed directly
# print ('Exception encountered while querying for images\n')
else:
# and print them
k = getattr(rsp,'photos',None)
if k:
m = getattr(rsp.photos[0],'photo',None)
if m:
for b in rsp.photos[0].photo:
if b!=None:
photo_id = b['id']
photo_data = { }
photo_data['id'] = b['id']
photo_data['secret'] = b['secret']
photo_data['server'] = b['server']
photo_data['farm'] = b['farm']
photo_data['owner'] = b['owner']
photo_data['title'] = b['title']
photo_data['originalsecret'] = b['originalsecret']
photo_data['originalformat'] = b['originalformat']
photo_data['o_height'] = b['o_height']
photo_data['o_width'] = b['o_width']
photo_data['datetaken'] = b['datetaken'].encode("ascii","replace")
photo_data['dateupload'] = b['dateupload'].encode("ascii","replace")
photo_data['tags'] = b['tags'].encode("ascii","replace")
photo_data['license'] = b['license'].encode("ascii","replace")
photo_data['latitude'] = b['latitude'].encode("ascii","replace")
photo_data['longitude'] = b['longitude'].encode("ascii","replace")
photo_data['accuracy'] = b['accuracy'].encode("ascii","replace")
photo_data['views'] = b['views']
photo_data['interestingness'] = (current_image_num, total_images)
results[photo_id] = photo_data
current_image_num = current_image_num + 1;
total_image_num = total_image_num + 1;
pagenum = pagenum + 1; #this is in the else exception block. It won't increment for a failure.
#this block is indented such that it will only run if there are no exceptions
#in the original query. That means if there are exceptions, mintime won't be incremented
#and it will try again
# timeskip = maxtime - mintime #used for initializing next binary search
# mintime = maxtime
timeskip = maxtime - mintime #used for initializing next binary search
maxtime = mintime
return results
counter = -1
while (pagenum <= num_visit_pages):
#for pagenum in range(1, num_visit_pages + 1): #page one is searched twice
print ' page number ' + str(pagenum)
try:
print("PAGE")
print(pagenum)
# WARNING THIS QUERY HAS TO MATCH THE SEARCH QUERY!!!!
rsp = fapi.photos_search(
api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250",
page=str(pagenum),
has_geo="0",
text=query_string,
#extras = "tags, original_format, license, geo, date_taken, date_upload, o_dims, views",
#accuracy="6", #6 is region level.
min_upload_date=str(1121832000), #mintime),
max_upload_date=str(1192165200)) #maxtime))
#rsp = fapi.photos_search(api_key=flickrAPIKey,
# ispublic="1",
# media="photos",
# per_page="250",
# page='0', #str(pagenum),
# sort="interestingness-desc",
# has_geo = "0", #bbox="-180, -90, 180, 90",
# text=query_string,
# #accuracy="6", #6 is region level. most things seem 10 or better.
def run_flickr_query(query_string, max_photos = 1000, startDate = "1/1/2010", endDate = "31/12/2011"):
socket.setdefaulttimeout(30) #30 second time out on sockets before they throw
#an exception. I've been having trouble with urllib.urlopen hanging in the
#flickr API. This will show up as exceptions.IOError.
#the time out needs to be pretty long, it seems, because the flickr servers can be slow
#to respond to our big searches.
###########################################################################
# Modify this section to reflect your data and specific search
###########################################################################
# flickr auth information:
# change these to your flickr api keys and secret
# make a new FlickrAPI instance
fapi = FlickrAPI(flickrAPIKey, flickrSecret)
print '\n\nquery_string is ' + query_string
total_images_queried = 0;
# number of seconds to skip per query
#timeskip = 62899200 #two years
timeskip = 604800 #one week
#timeskip = 172800 #two days
#timeskip = 86400 #one day
#timeskip = 3600 #one hour
#timeskip = 2257 #for resuming previous query
#mintime = 1121832000 #from im2gps
#mintime = 1167407788 # resume crash england
#mintime = 1177828976 #resume crash japan
#mintime = 1187753798 #resume crash greece
#mintime = 1171416400 #resume crash WashingtonDC
mintime = 1287878400 # 10/24/2010
maxtime = mintime+timeskip
#endtime = 1192165200 #10/12/2007, at the end of im2gps queries
endtime = 1351100325 # 10/24/2012
#this is the desired number of photos in each block
desired_photos = 250
print "hola"
print datetime.fromtimestamp(mintime)
print datetime.fromtimestamp(endtime)
total_image_num = 0
while (maxtime < endtime):
#new approach - adjust maxtime until we get the desired number of images
#within a block. We'll need to keep upper bounds and lower
#lower bound is well defined (mintime), but upper bound is not. We can't
#search all the way from endtime.
lower_bound = mintime + 900 #lower bound OF the upper time limit. must be at least 15 minutes or zero results
upper_bound = mintime + timeskip * 20 #upper bound of the upper time limit
maxtime = .95 * lower_bound + .05 * upper_bound
print '\nBinary search on time range upper bound'
print 'Lower bound is ' + str(datetime.fromtimestamp(lower_bound))
print 'Upper bound is ' + str(datetime.fromtimestamp(upper_bound))
if total_image_num > max_photos:
print 'Number of photos %d > %d limit.' % (total_image_num, max_photos)
break
keep_going = 6 #search stops after a fixed number of iterations
while( keep_going > 0 and maxtime < endtime):
try:
rsp = fapi.photos_search(api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250",
page="1",
#has_geo = "1", #bbox="-180, -90, 180, 90",
text=query_string,
#accuracy="6", #6 is region level. most things seem 10 or better.
min_upload_date=str(mintime),
max_upload_date=str(maxtime))
##min_taken_date=str(datetime.fromtimestamp(mintime)),
##max_taken_date=str(datetime.fromtimestamp(maxtime)))
#we want to catch these failures somehow and keep going.
time.sleep(1)
fapi.testFailure(rsp)
total_images = rsp.photos[0]['total'];
null_test = int(total_images); #want to make sure this won't crash later on for some reason
null_test = float(total_images);
print '\nnumimgs: ' + total_images
print 'mintime: ' + str(mintime) + ' maxtime: ' + str(maxtime) + ' timeskip: ' + str(maxtime - mintime)
if( int(total_images) > desired_photos ):
print 'too many photos in block, reducing maxtime'
upper_bound = maxtime
maxtime = (lower_bound + maxtime) / 2 #midpoint between current value and lower bound.
if( int(total_images) < desired_photos):
print 'too few photos in block, increasing maxtime'
lower_bound = maxtime
maxtime = (upper_bound + maxtime) / 2
print 'Lower bound is ' + str(datetime.fromtimestamp(lower_bound))
print 'Upper bound is ' + str(datetime.fromtimestamp(upper_bound))
if( int(total_images) > 0): #only if we're not in a degenerate case
keep_going = keep_going - 1
else:
upper_bound = upper_bound + timeskip;
except KeyboardInterrupt:
print('Keyboard exception while querying for images, exiting\n')
raise
except:
print sys.exc_info()[0]
#print type(inst) # the exception instance
#print inst.args # arguments stored in .args
#print inst # __str__ allows args to printed directly
print ('Exception encountered while querying for images\n')
#end of while binary search
print 'finished binary search'
s = '\nmintime: ' + str(mintime) + ' maxtime: ' + str(maxtime)
print s
out_file.write(s + '\n')
i = getattr(rsp,'photos',None)
if i:
s = 'numimgs: ' + total_images
print s
out_file.write(s + '\n')
current_image_num = 1;
num = int(rsp.photos[0]['pages'])
s = 'total pages: ' + str(num)
print s
out_file.write(s + '\n')
#only visit 16 pages max, to try and avoid the dreaded duplicate bug
#16 pages = 4000 images, should be duplicate safe. Most interesting pictures will be taken.
num_visit_pages = min(16,num)
s = 'visiting only ' + str(num_visit_pages) + ' pages ( up to ' + str(num_visit_pages * 250) + ' images)'
print s
out_file.write(s + '\n')
total_images_queried = total_images_queried + min((num_visit_pages * 250), int(total_images))
#print 'stopping before page ' + str(int(math.ceil(num/3) + 1)) + '\n'
pagenum = 1;
while( pagenum <= num_visit_pages ):
#for pagenum in range(1, num_visit_pages + 1): #page one is searched twice
print ' page number ' + str(pagenum)
try:
rsp = fapi.photos_search(api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250",
page=str(pagenum),
sort="interestingness-desc",
#has_geo = "1", #bbox="-180, -90, 180, 90",
text=query_string,
#accuracy="6", #6 is region level. most things seem 10 or better.
extras = "tags, original_format, license, geo, date_taken, date_upload, o_dims, views",
min_upload_date=str(mintime),
max_upload_date=str(maxtime))
##min_taken_date=str(datetime.fromtimestamp(mintime)),
##max_taken_date=str(datetime.fromtimestamp(maxtime)))
time.sleep(1)
fapi.testFailure(rsp)
except KeyboardInterrupt:
print('Keyboard exception while querying for images, exiting\n')
raise
except:
print sys.exc_info()[0]
#print type(inst) # the exception instance
#print inst.args # arguments stored in .args
#print inst # __str__ allows args to printed directly
print ('Exception encountered while querying for images\n')
else:
# and print them
k = getattr(rsp,'photos',None)
if k:
m = getattr(rsp.photos[0],'photo',None)
if m:
for b in rsp.photos[0].photo:
if b!=None:
out_file.write('photo: ' + b['id'] + ' ' + b['secret'] + ' ' + b['server'] + ' ' + b['farm'] + '\n')
out_file.write('owner: ' + b['owner'] + '\n')
out_file.write('title: ' + b['title'].encode("ascii","replace") + '\n')
out_file.write('originalsecret: ' + b['originalsecret'] + '\n')
out_file.write('originalformat: ' + b['originalformat'] + '\n')
out_file.write('o_height: ' + b['o_height'] + '\n')
out_file.write('o_width: ' + b['o_width'] + '\n')
out_file.write('datetaken: ' + b['datetaken'].encode("ascii","replace") + '\n')
out_file.write('dateupload: ' + b['dateupload'].encode("ascii","replace") + '\n')
out_file.write('tags: ' + b['tags'].encode("ascii","replace") + '\n')
out_file.write('license: ' + b['license'].encode("ascii","replace") + '\n')
out_file.write('latitude: ' + b['latitude'].encode("ascii","replace") + '\n')
out_file.write('longitude: ' + b['longitude'].encode("ascii","replace") + '\n')
out_file.write('accuracy: ' + b['accuracy'].encode("ascii","replace") + '\n')
out_file.write('views: ' + b['views'] + '\n')
out_file.write('interestingness: ' + str(current_image_num) + ' out of ' + str(total_images) + '\n');
out_file.write('\n')
current_image_num = current_image_num + 1;
total_image_num = total_image_num + 1;
pagenum = pagenum + 1; #this is in the else exception block. It won't increment for a failure.
#this block is indented such that it will only run if there are no exceptions
#in the original query. That means if there are exceptions, mintime won't be incremented
#and it will try again
timeskip = maxtime - mintime #used for initializing next binary search
mintime = maxtime
out_file.write('Total images queried: ' + str(total_images_queried) + '\n')
out_file.close
out_file = open(output_filename,'w')
fapi = FlickrAPI(flickrAPIKey, flickrSecret)
i = 0
page_num = range(1,5)
for i in page_num:
page_nu = str(i)
try:
rsp = fapi.photos_search(api_key = flickrAPIKey,
ispublic = "1",
media = "photos",
per_page = "250", #seems like it is max
page = page_nu,
text = query_string)
time.sleep(1)
fapi.testFailure(rsp)
total_images = rsp.photos[0]['total']
for b in rsp.photos[0].photo:
if b!=None:
out_file.write('photo: ' + b['id'] + ' ' + b['secret'] + ' ' + b['server'] + '\n')
out_file.write('owner: ' + b['owner'] + '\n')
out_file.write('title: ' + b['title'].encode("ascii","replace") + '\n')
out_file.write('tags: ' + b['tags'].encode("ascii","replace") + '\n')
out_file.write('\n')
fapi = FlickrAPI(flickrAPIKey, flickrSecret)
i = 0
page_num = range(1, 5)
for i in page_num:
page_nu = str(i)
try:
rsp = fapi.photos_search(
api_key=flickrAPIKey,
ispublic="1",
media="photos",
per_page="250", #seems like it is max
page=page_nu,
text=query_string)
time.sleep(1)
fapi.testFailure(rsp)
total_images = rsp.photos[0]['total']
for b in rsp.photos[0].photo:
if b != None:
out_file.write('photo: ' + b['id'] + ' ' + b['secret'] +
' ' + b['server'] + '\n')
out_file.write('owner: ' + b['owner'] + '\n')
out_file.write('title: ' +
b['title'].encode("ascii", "replace") +
'\n')