simple img show added

Author: Catoverflow

requirements.txt

@@ -1,3 +1,4 @@
 nltk
-gensim
-zstd
+zstd
+pillow
+requests

semantic_search.py

@@ -4,6 +4,7 @@
 from bool_search import Bool_Search
 from utils.data_process import Data
 from utils.tf_idf import TF_IDF
+from utils.img import show_image
 import logging
 
 
@@ -106,5 +107,11 @@ def load():
         else:
             for docid in range(len(res)):
                 if(res[docid][0] > 0):
-                    print('{}:\t{}'.format(res[docid][0],
-                                           metadata[res[docid][1]]['title']))
+                    imgurl = metadata[res[docid][1]]['img']
+                    title = metadata[res[docid][1]]['title']
+                    print('{}:\t{}'.format(res[docid][0],title),end='\t')
+                    if len(imgurl) > 0:
+                        print('Found image')
+                        show_image(imgurl,title)
+                    else:
+                        print('Image not found')

utils/data_process.py

@@ -52,9 +52,10 @@ def load(self, path="./data", maxfile=-1):
                     self.metadata.append({
                         "id": rawdata["uuid"],
                         # "time": rawdata["published"],
-                        "title": rawdata["title"]
+                        "title": rawdata["title"],
                         # "author": rawdata["author"],
                         # "url": rawdata["url"]
+                        "img": rawdata["thread"]["main_image"]
                     })
                     if file_id == maxfile:
                         logging.info(f'Maxfile reached, {file_id} file loaded')
@@ -154,4 +155,4 @@ def process(self):
                 self.strip_stop_words(
                     self.pre_process(self.headerdata[docid])))
         self.parse_syn()
-        self.gen_dict()
+        self.gen_dict()

utils/img.py

@@ -0,0 +1,8 @@
+from requests import get
+from PIL import Image
+from io import BytesIO
+
+def show_image(url,title):
+    raw = get(url)
+    img = Image.open(BytesIO(raw.content))
+    img.show(title=title)