440058 plusalpha

run.py

@@ -0,0 +1,102 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+# In[1]:
+
+
+from collections import defaultdict, Counter
+import csv
+import regex as re
+import pandas as pd
+from nltk import trigrams, word_tokenize
+
+
+# In[26]:
+
+def prepare_data():
+    x_train = pd.read_csv('train/in.tsv.xz', sep='\t',  header=None,
+                          quoting=csv.QUOTE_NONE, nrows=70000, on_bad_lines='skip')
+    y_train = pd.read_csv('train/expected.tsv', sep='\t', header=None,
+                          quoting=csv.QUOTE_NONE, nrows=70000, on_bad_lines='skip')
+    x_train = x_train[[6, 7]]
+    x_train = pd.concat([x_train, y_train], axis=1)
+    x_train['l'] = x_train[6] + x_train[0] + x_train[7]
+
+    return x_train, y_train
+
+
+x_train, y_train = prepare_data()
+
+
+# In[39]:
+
+
+def train(x_train):
+    model = defaultdict(lambda: defaultdict(lambda: 0))
+    setOf = set()
+    alpha = 0.02
+    count = x_train.iterrows()
+    for i, (_, row) in enumerate(count):
+        text = re.sub(r'\p{P}', '', str(row['l']).lower().replace(
+            '-\\n', '').replace('\\n', ' '))
+        words = word_tokenize(text)
+        for word_1, word_2, word_3 in trigrams(words, pad_right=True, pad_left=True):
+            if word_1 and word_2 and word_3:
+                model[(word_1, word_3)][word_2] += 1
+                setOf.add(word_1)
+                setOf.add(word_2)
+                setOf.add(word_3)
+
+    for words in model:
+        num_n_grams = float(sum(model[words].values()))
+        for word in model[words]:
+            model[words][word] = (model[words][word] + alpha) / \
+                (num_n_grams + alpha * len(setOf))
+
+    return model
+
+
+# In[41]:
+
+
+def predict(before, after):
+    result = ''
+    p = 0.0
+    pred = dict(Counter(dict(model[before, after])).most_common(7))
+    for key, value in pred.items():
+        p += value
+        result += f'{key}:{value} '
+    if p == 0.0:
+        result = 'to:0.02 the:0.02 be:0.02 and:0.01 or:0.01 and:0.01 a:0.01 :0.9'
+        return result
+    result += f':{max(1 - p, 0.01)}'
+
+    return result
+
+
+# In[42]:
+def preprocess(text):
+    text = text.lower().replace('-\\n', '').replace('\\n', ' ')
+    return re.sub(r'\p{P}', '', text)
+
+
+def gap_predict(file):
+    X_test = pd.read_csv(f'{file}/in.tsv.xz', sep='\t', header=None,
+                         quoting=csv.QUOTE_NONE,  error_bad_lines=False)
+    with open(f'{file}/out.tsv', 'w', encoding='utf-8') as result_file:
+        for _, row in X_test.iterrows():
+            before, after = word_tokenize(preprocess(
+                str(row[6]))), word_tokenize(preprocess(str(row[7])))
+            if len(before) < 3 or len(after) < 3:
+                result = 'to:0.02 the:0.02 be:0.02 and:0.01 or:0.01 and:0.01 a:0.01 :0.9'
+            else:
+                result = predict(before[-1], after[0])
+            result_file.write(result + '\n')
+
+
+# In[43]:
+
+
+model = train(x_train)
+gap_predict('dev-0')
+gap_predict('test-A')