修订 11

15 工程化/1 松耦合/5 插件/plugins-src/frequencies1.py

@@ -1,6 +1,6 @@
 import operator
 
-def top25(word_list):
+def top_word(word_list):
     word_freqs = {}
     for w in word_list:
         if w in word_freqs:

15 工程化/1 松耦合/5 插件/plugins-src/frequencies2.py

@@ -1,6 +1,6 @@
-import operator, collections
+import  collections
 
-def top25(word_list):
+def top_word(word_list):
     counts = collections.Counter(w for w in word_list)
     return counts.most_common(10)
 

15 工程化/1 松耦合/5 插件/plugins-src/words1.py

@@ -1,14 +0,0 @@
-import sys, re, string
-from cppy.cp_util import *
-
-def extract_words(path_to_file):
-    with open(path_to_file,encoding='utf-8') as f:
-        str_data = f.read()
-    pattern = re.compile('[\W_]+')
-    word_list = pattern.sub(' ', str_data).lower().split()
-
-    
-    stop_words = get_stopwords()
-
-    return [w for w in word_list if not w in stop_words]
-

15 工程化/1 松耦合/5 插件/plugins-src/words2.py

@@ -1,8 +0,0 @@
-import sys, re, string
-from cppy.cp_util import *
-
-def extract_words(path_to_file):
-    words = re.findall('[a-z]{2,}', open(path_to_file,encoding='utf-8').read().lower())
-    stopwords = get_stopwords()
-    return [w for w in words if w not in stopwords]
-

15 工程化/1 松耦合/5 插件/tf-20.py

@@ -1,19 +1,19 @@
 import configparser, importlib.machinery
 from cppy.cp_util import *
 
-
 def load_plugins():
-    config = configparser.ConfigParser()
+    config = configparser.ConfigParser()    
     script_dir = os.path.dirname(os.path.abspath(__file__))  
-    os.chdir(script_dir)
+    os.chdir(script_dir)   
-    config.read("config.ini")
+    config.read("config.ini")    
-    words_plugin = config.get("Plugins", "words")
     frequencies_plugin = config.get("Plugins", "frequencies")
-    global tfwords, tffreqs
+    
-    tfwords = importlib.machinery.SourcelessFileLoader('tfwords', words_plugin).load_module()
+    global get_frequencies
-    tffreqs = importlib.machinery.SourcelessFileLoader('tffreqs', frequencies_plugin).load_module()
+    get_frequencies = importlib.machinery.SourcelessFileLoader('tffreqs', frequencies_plugin).load_module()
+
 
 load_plugins()
-word_freqs = tffreqs.top25(tfwords.extract_words( testfilepath ))
 
+wordlist = extract_file_words( testfilepath ) 
+word_freqs = get_frequencies.top_word( wordlist )
 print_word_freqs(word_freqs)