Merge branch 'master' of https://mit.cs.uchicago.edu/mawh/project_kitty

Searcresults.py

@@ -77,25 +77,31 @@ class SearchResults:
         '''
 
         self.__query = query
+        print('query',query)
         self.__filters = filters
+        print('filters',filters)
         self.__required_words = self.get_required_words(query)
         if self.__required_words != None:
             self.__filters['required_words'] = self.__required_words
         self.__forbidden_words = self.get_forbidden_words(query)
         if self.__forbidden_words != None:
             self.__filters['forbidden_words'] = self.__forbidden_words
-        
         self.__soup = self.get_search_soup(query)
         self.__links = [a['href'] for a in self.__soup.find_all(\
                         'a', class_='result_link')]
         self.__results = self.__soup.find_all('a', class_ = 'result_link')
+        print('num reuslts',len(self.__results))
+        print(self.__results[0])
         self.__unfiltered_outer_articles = [OuterArticle(r) for r in self.__results if \
                'Category' not in r.find('div', class_ = 'result_title').text]
+        print('unfiltered outer aticles',len(self.__unfiltered_outer_articles))
         self.__categories = [r for r in self.__results if\
              'Category' in r.find('div', class_ = 'result_title')]
         self.__outer_articles = [r for r in self.__unfiltered_outer_articles\
              if self.passes_outer_filters(r)]
+        print('num outer articles',len(self.__outer_articles))
         self.__unfiltered_articles = [Article(r.url) for r in self.__outer_articles]
+        print('making articles')
         self.articles = [a for a in self.__unfiltered_articles if\
             self.passes_all_filters(a)]
 
@@ -108,6 +114,7 @@ class SearchResults:
 
     
     def passes_outer_filters(self, outer_article):
+        print('checking')
         if 'date_updated' in self.__filters:
             num, interval = outer_article.date_updated.split()[:2]
             num = int(num)
@@ -160,8 +167,11 @@ class SearchResults:
             if self.__filters['child_safe']:
                 self.__forbidden_words += inappropiate_words 
         if 'pct_helpful' in self.__filters:
-            if article.pct_helpful < self.__filters['pct_helpful']:
-                return False
+            try:
+                if article.pct_helpful < self.__filters['pct_helpful']:
+                    return False
+            except TypeError:
+                pass
         if 'num_voters' in self.__filters:
             if article.num_voters < self.__filters['num_voters']:
                 return False
@@ -176,6 +186,7 @@ class SearchResults:
 
     def __repr__(self, limit_results = None):
         info = 'Results for ' + self.__query + ':\n'
+        print(len(self.articles))
         if limit_results != None:
             results = self.articles[:limit_results]
         else:
@@ -219,9 +230,10 @@ class Article:
         self.languages = self.find_languages(self.__soup)
         self.view = self.__repr__()
         self.words = self.scrape_words(self.url, self.__soup)
+        '''
         self.salient = self.find_salient_ngrams(self.words,\
                        1, False, 0.5)
-
+        '''
 
     def get_page_soup(self, url):
         response = requests.get(url)
@@ -403,7 +415,7 @@ class Article:
                     links.append(filtered_link)
 
         return links
-    
+    '''
 
     def find_salient_ngrams(self, wh_page, n, case_sensitive, threshold):
 
@@ -476,7 +488,7 @@ class Article:
         return idf
 
     def find_salient(self, docs, threshold):
-        '''
+        """
         Compute the salient words for each document.  A word is salient if
         its tf-idf score is strictly above a given threshold.
 
@@ -485,7 +497,7 @@ class Article:
         threshold: float
 
         Returns: list of sets of salient words
-        '''
+        """
 
         salient_lst = []
         for tok_lst in docs:
@@ -496,8 +508,8 @@ class Article:
                     new_set.add(a_token)
             salient_lst.append(new_set)        
         return salient_lst
-
-
+    '''
+    
     def __repr__(self):
         print('repr')
         info = self.title

WikiHowHow.py

@@ -2,73 +2,77 @@
 import sys
 import PySimpleGUI as sg
 import SearchResults
-
-
-
-
-
+ 
+ 
+ 
+ 
+ 
 def callback_function1():
-
-   sg.popup('In Callback Function 1')
-   print('In the callback function 1')
-
+ 
+  sg.popup('In Callback Function 1')
+  print('In the callback function 1')
+ 
 def callback_function2():
-
-   sg.popup('In Callback Function 2')
-   print('In the callback function 2')
-
-
-layout = [[sg.Text('Wikihow Tool')],
-         [sg.Text('Search Here'), sg.Input(key = 'Query'), sg.Button('Search')],
-         [sg.Text('Filters', font = ('Helvetica', 15, 'underline'))],
-         [sg.Text('ChildSafety'), sg.CBox('On', key = 'child_safe'), sg.Spin([i for i in range(53)], initial_value=0, key = "date_updated"), sg.Text('Max time (months) since last update')], 
-         [sg.Text('Expert Co-authored'), sg.CBox('', key = 'e_verif'), sg.Text('Quality Tested'), sg.CBox('', key = 'q_verif')],
-         [sg.Text('Minimum % viewers who found it helpful'), sg.Slider(range=(0,100), default_value=50, size=(20,15), orientation='horizontal', font=('Helvetica', 12), key = 'pct_helpful')],
-         [sg.Spin([i for i in range(100)], initial_value=0, key = "num_voters"), sg.Text('Number of people who voted on helpfulness')],
-         [sg.Text('Categories'), sg.CBox('House & Garden', key = 'House'), sg.CBox('Computers & Internet', key = 'CP & I'), sg.CBox('Art and Literature', key = 'Arts'), sg.CBox('Social and Organizational Skills', key = 'Social keys')],
-         [sg.Spin([i for i in range(100)], initial_value=0, key = "num_sources"), sg.Text('Number of sources cited')],
-         [sg.Spin([i for i in range(100)], initial_value=0, key = "num_views"), sg.Text('Views')],
-         [sg.Combo(['English', 'Español', 'Português', 'Italiano', 'Français', 'русский язык', 'Deutsch', '简体中文', 'Nederlands', 'Čeština', 'Bahasa Indonesia', '日本語', 'हिंदी', 'ภาษาไทย',' العَرَبِيةُ', 'Tiếng Việt', '한국말', 'Türkçe',' فارسی']), sg.Text('Language of the Article')],
-         [sg.StatusBar( text=f'| Filters On: {None}', size=(30,1), pad=(0,0), text_color='black', background_color='white', relief=sg.RELIEF_FLAT, justification='left', visible=True, key='status_bar' )]]
-
-
-
+ 
+  sg.popup('In Callback Function 2')
+  print('In the callback function 2')
+ 
+ 
 def interact(filters = {}):
-    window = sg.Window('WikiHow Search and Filters', layout)
-    while True:             
-        event, values = window.read()
-        if event == sg.WIN_CLOSED:
-            break    
-        elif event == 'Search':
-            filters = values
-            wh_query = values['Query']
-            print(event, values)
-            display_results(filters)
-    window.close()
-
-
-def display_results(filters):
-    query = filters['Query']
-    results = SearchResults.SearchResults(query, filters)
-    print(results)
-    #code to add window here
-
-
-'''
-sg.theme('DarkAmber')   # Add a little color to your windows
-# All the stuff inside your window. This is the PSG magic code compactor...
-layout = [  [sg.Text('Some text on Row 1')],
-            [sg.Text('Enter something on Row 2'), sg.InputText()],
-            [sg.OK(), sg.Cancel()]]
-
-# Create the Window
-window = sg.Window('Window Title', layout)
-# Event Loop to process "events"
-while True:             
-    event, values = window.read()
-    if event in (sg.WIN_CLOSED, 'Cancel'):
-        break
-    print('you entered', values[0])
+ 
+   layout = [[sg.Text('Wikihow Tool')],
+        [sg.Text('Search Here'), sg.Input(key = 'Query'), sg.Button('Search')],
+        [sg.Text('Filters', font = ('Helvetica', 15, 'underline'))],
+        [sg.Text('ChildSafety'), sg.CBox('On', key = 'child_safe'), sg.Spin([i for i in range(53)], initial_value=0, key = "date_updated"), sg.Text('Max time (months) since last update')],
+        [sg.Text('Expert Co-authored'), sg.CBox('', key = 'e_verif'), sg.Text('Quality Tested'), sg.CBox('', key = 'q_verif')],
+        [sg.Text('Minimum % viewers who found it helpful'), sg.Slider(range=(0,100), default_value=50, size=(20,15), orientation='horizontal', font=('Helvetica', 12), key = 'pct_helpful')],
+        [sg.Spin([i for i in range(100)], initial_value=0, key = "num_voters"), sg.Text('Number of people who voted on helpfulness')],
+        [sg.Text('Categories'), sg.CBox('House & Garden', key = 'House'), sg.CBox('Computers & Internet', key = 'CP & I'), sg.CBox('Art and Literature', key = 'Arts'), sg.CBox('Social and Organizational Skills', key = 'Social keys')],
+        [sg.Spin([i for i in range(100)], initial_value=0, key = "num_sources"), sg.Text('Number of sources cited')],
+        [sg.Spin([i for i in range(100)], initial_value=0, key = "num_views"), sg.Text('Views')],
+        [sg.Combo(['English', 'Español', 'Português', 'Italiano', 'Français', 'русский язык', 'Deutsch', '简体中文', 'Nederlands', 'Čeština', 'Bahasa Indonesia', '日本語', 'हिंदी', 'ภาษาไทย',' العَرَبِيةُ', 'Tiếng Việt', '한국말', 'Türkçe',' فارسی']), sg.Text('Language of the Article')],
+        [sg.StatusBar( text=f'| Filters On: {None}', size=(30,1), pad=(0,0), text_color='black', background_color='white', relief=sg.RELIEF_FLAT, justification='left', visible=True, key='status_bar' )]]
+ 
+   #layout_x = layout
+   print('im interacting')
+   window = sg.Window('WikiHow Search and Filters', layout)
+   results_pg_active = False
+ 
+   while True:
+       print(results_pg_active)            
+       event, values = window.read()
+       if event == sg.WIN_CLOSED:
+           break   
+       elif not results_pg_active and event == 'Search':
+           results_pg_active = True
+           filters = values
+           wh_query = values['Query']
+           print(event, values)
+           display_results(filters, results_pg_active)
+           results_pg_active = False
+   window.close()
+ 
+ 
+def display_results(filters, results_pg_active):
+   print(results_pg_active)
+   query = filters['Query']
+   input_filters = dict(filters)
+   if filters['date_updated'] == 0:
+       del input_filters['date_updated']
+   results = SearchResults.SearchResults(query, input_filters)
+   print(results)
+   layout2 = [[sg.Text('Results')],
+               [sg.Text(str(filters))]]
+   win2 = sg.Window('Window 2', layout2)
+ 
+   if results_pg_active:
+       ev2, vals2 = win2.read(timeout=100)
+       if ev2 == sg.WIN_CLOSED or ev2 == 'Exit':
+           results_pg_active  = False
+           win2.close()
+   #code to add window here
+ 
+ 
+ 
+ 
 
-window.close()
-'''
\ No newline at end of file