From 1ca4058a3fa2907999ca1337925f544af8365422 Mon Sep 17 00:00:00 2001
From: Klaudia <s407228>
Date: Sun, 25 Apr 2021 00:51:40 +0200
Subject: [PATCH] Done

---
 Skrypt.py      | 64 +++++++++++++++++++++++++++++++++++++
 dev-0/out.tsv  | 87 ++++++++++++++++++++++++++++++++++++++++++++++++++
 test-A/out.tsv | 87 ++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 238 insertions(+)
 create mode 100644 Skrypt.py
 create mode 100644 dev-0/out.tsv
 create mode 100644 test-A/out.tsv

diff --git a/Skrypt.py b/Skrypt.py
new file mode 100644
index 0000000..4ab9d30
--- /dev/null
+++ b/Skrypt.py
@@ -0,0 +1,64 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+
+from sklearn.feature_extraction.text import TfidfVectorizer
+from sklearn.cluster import KMeans
+import string
+from stop_words import get_stop_words
+
+stop_words = get_stop_words('polish')
+
+
+
+inp1 = open('dev-0/in.tsv', 'r', encoding="utf-8")
+out1 = open("dev-0/out.tsv", "w")
+
+linia1 = inp1.readlines()
+inp1.close()
+
+prep=[]
+for x in linia1:
+    temp = ""
+    for y in x.split():
+        y = y.strip().replace(",", "")
+        if y not in stop_words:
+             temp = temp + " " + y
+    prep.append(temp)
+    
+vectorizer1 = TfidfVectorizer()
+vectorizer1 = vectorizer1.fit_transform(prep)
+
+predict1 = KMeans(n_clusters=25, max_iter=1000).fit_predict(vectorizer1)
+print(predict1)
+
+for x in predict1:
+    out1.write(str(x) + '\n')
+out1.close()
+
+
+
+inp2 = open('test-A/in.tsv', 'r', encoding="utf-8")
+out2 = open("test-A/out.tsv", "w")
+
+linia2 = inp2.readlines()
+inp2.close()
+
+prep2=[]
+for x2 in linia1:
+    temp2 = ""
+    for y2 in x2.split():
+        y2 = y2.strip().replace(",", "")
+        if y2 not in stop_words:
+             temp2 = temp2 + " " + y2
+    prep2.append(temp2)
+    
+vectorizer2 = TfidfVectorizer()
+vectorizer2 = vectorizer2.fit_transform(prep)
+
+predict2 = KMeans(n_clusters=25, max_iter=1000).fit_predict(vectorizer2)
+print(predict2)
+
+for y in predict2:
+    out2.write(str(y) + '\n')
+out2.close()
diff --git a/dev-0/out.tsv b/dev-0/out.tsv
new file mode 100644
index 0000000..6bd13c5
--- /dev/null
+++ b/dev-0/out.tsv
@@ -0,0 +1,87 @@
+7
+0
+1
+6
+22
+20
+3
+4
+8
+5
+4
+20
+0
+7
+21
+11
+7
+22
+12
+17
+4
+24
+9
+2
+18
+15
+10
+9
+24
+0
+3
+3
+3
+7
+21
+23
+7
+7
+10
+6
+8
+14
+3
+6
+14
+7
+23
+0
+16
+3
+18
+1
+3
+15
+13
+2
+0
+0
+3
+19
+3
+7
+3
+0
+20
+19
+3
+14
+7
+23
+22
+12
+14
+3
+0
+4
+18
+13
+6
+7
+5
+4
+5
+17
+3
+15
+0
diff --git a/test-A/out.tsv b/test-A/out.tsv
new file mode 100644
index 0000000..be5cc6f
--- /dev/null
+++ b/test-A/out.tsv
@@ -0,0 +1,87 @@
+9
+0
+2
+16
+10
+18
+13
+23
+18
+3
+6
+18
+0
+9
+12
+12
+21
+10
+14
+0
+11
+24
+20
+22
+23
+4
+15
+20
+20
+0
+7
+7
+7
+8
+12
+1
+8
+9
+15
+16
+18
+17
+13
+19
+17
+1
+1
+0
+18
+7
+23
+2
+7
+4
+5
+22
+0
+0
+7
+24
+1
+21
+13
+0
+18
+24
+7
+17
+8
+1
+10
+14
+17
+7
+0
+0
+23
+5
+19
+0
+3
+6
+3
+10
+7
+9
+0