In [12]:
import boto3
import requests
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
In [16]:
s3 = boto3.client("s3")
In [5]:
img_df = pd.read_csv("s3://dogfaces/reviews/pictures_log.csv")
img_df.shape
Out[5]:
In [9]:
test_link = img_df['pic_url'][1]
print test_link
In [11]:
img_data = requests.get(test_link).content
In [17]:
with open('photo.jpg', 'wb') as f:
f.write(img_data)
In [27]:
plt.imshow(plt.imread('n02086910_399.jpg'))
Out[27]:
In [24]:
s3.put_object(Bucket="dogfaces",Key='test.jpg', Body=img_data)
Out[24]:
In [ ]:
def download_review_pics(df):
N = df.shape[0]
for i in xrange(N):
if i%100 == 1:
print "processed till {} images".format(i+1)
pic_url = df.iloc[i]['pic_url']
pic_name = df.iloc[i]['pic_name']
r = requests.get(test_link)
if r.status_code == 200:
img_data = r.content
image_name = "reviews/review_pics/"+pic_name
s3.put_object(Bucket="dogfaces", Key=image_name, Body=img_data);
else:
print "images {} could not download".format(pic_name)
In [28]:
img_df.head()
Out[28]:
In [ ]: