v2

run.py

@@ -3,12 +3,12 @@ import pandas as pd
 import csv
 import os
 import kenlm
-from nltk import word_tokenize
-from collections import defaultdict, Counter
+from collections import Counter, defaultdict
+from math import log10
 
 #%%
 def clean(text):
-    text = str(text).lower().strip().replace("’", "'").replace('\\n', " ").replace("'t", " not").replace("'s", " is").replace("'ll", " will").replace("'m", " am").replace("'ve", " have").replace(",", "").replace("-", "").replace(".", "").replace("'", "".replace("”", ""))
+    text = str(text).lower().strip().replace("’", "'").replace('\\n', " ").replace("'t", " not").replace("'s", " is").replace("'ll", " will").replace("'m", " am").replace("'ve", " have").replace(",", "").replace("-", "").replace(".", "").replace("'", "".replace("”", "").replace(">", ""))
     return text
 
 train_in = pd.read_csv("train/in.tsv.xz", sep='\t', header=None, encoding="UTF-8", on_bad_lines="skip", quoting=csv.QUOTE_NONE, nrows=300000)[[6, 7]]
@@ -23,14 +23,13 @@ if not os.path.isfile('train_file.txt'):
             f.write(text + "\n")
 
 #%%
-#!../kenlm/build/bin/lmplz -o 4 < train_file.txt > model.arpa
+#get_ipython().system('../kenlm/build/bin/lmplz -o 4 < train_file.txt > model.arpa --skip_symbols')
 model = kenlm.Model("model.arpa")
-
 #%%
 import nltk
+from nltk import word_tokenize
 nltk.download('punkt')
 
-#%%
 most_common = defaultdict(lambda: 0)
 for text in data:
     words = word_tokenize(text)
@@ -42,24 +41,26 @@ most_common = Counter(most_common).most_common(8000)
 
 #%%
 def predict(path, result_path):
-    data = pd.read_csv(path, sep='\t', header=None, encoding="UTF-8", on_bad_lines="skip", quoting=csv.QUOTE_NONE)[7]
+    data = pd.read_csv(path, sep='\t', header=None, encoding="UTF-8", on_bad_lines="skip", quoting=csv.QUOTE_NONE)
     with open(result_path, "w+", encoding="UTF-8") as f:
-        for text in data:
+        for i, row in data.iterrows():
             result = {}
-            for word in most_common:
-                prob = model.score(text + f" {word[0]}")
-                result[word[0]] = prob
-            predictions = dict(Counter(result).most_common(6))
-            result = ""
-            for word, prob in predictions.items():
-                result += f"{word}:{prob} "
-            result = result.rstrip()
-            if len(result) == 0:
+            before = word_tokenize(clean(str(row[6])))[-3:]
+            if(len(before) < 2):
                 result = "a:0.2 the:0.2 to:0.2 of:0.1 and:0.1 of:0.1 :0.1"
+            else:
+                for w in most_common:
+                    word = w[0]
+                    prob = model.score(" ".join(before + [word]))
+                    result[word] = prob
+                predictions = dict(Counter(result).most_common(12))
+                result = ""
+                for word, prob in predictions.items():
+                    result += f"{word}:{prob} "
+            result  += f':{log10(0.99)}'
             f.write(result + "\n")
             print(result)
 
 
 predict("dev-0/in.tsv.xz", "dev-0/out.tsv")
-predict("test-A/in.tsv.xz", "test-A/out.tsv")
-# %%
+predict("test-A/in.tsv.xz", "test-A/out.tsv")