From 15bc16101177ee1b7354a3018dd6221f35684197 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dominik=20Dembi=C5=84ski?= <domdem2@st.amu.edu.pl>
Date: Tue, 26 Apr 2022 23:21:35 +0200
Subject: [PATCH] Zaktualizuj 'run.py'

---
 run.py | 66 ++++++++++++++++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 62 insertions(+), 4 deletions(-)

diff --git a/run.py b/run.py
index 9eff044..0f09e05 100644
--- a/run.py
+++ b/run.py
@@ -1,7 +1,65 @@
-for dataset in 'dev-0', 'dev-1', 'test-A':
-    with open(f'{dataset}/in.tsv') as f_in, open(f'{dataset}/out.tsv','w') as f_out:
-        for line in f_in:
-            f_out.write('0.45\n')
+men = ['samochód', 'auto', 'audi', 'bmw', 'mercedes', 'windows', 'silnika', 'mb', 'gb', 'mecz', 'legia', 'lech', 'real', 'komputer', 'telefon', 'apple', 'iphone', 'zwiastuny', 'hd',
+       'windows', 'system', 'serwer', 'serwery', 'siłownia', 'biceps', 'triceps']
+woman = ['zrobiłam', 'byłam', 'pojechałam','ciąży', 'miesiączki', 'makijaż', 'perfumy', 'kompleks', 'ciąża', 'ciazy', 'antykoncepcyjne', 'ginekologa',
+         'tabletki', 'sukienka', 'spódniczka', 'mąż', 'miesiączkę', 'samotność', 'kolczyki']
+def database(men, woman):
+    men_base = []
+    woman_base = []
+    for i in men:
+        men_base.append(i[:10].lower())
+    for i in woman:
+        woman_base.append(i[:10].lower())
+    return men_base, woman_base
+def predict(men_base, woman_base, data):
+    res = []
+    i = 0
+    for slowo in data:
+        men_count = 0
+        woman_count = 0
+        term = slowo.split()
+
+        for w in term:
+            if w[:10].lower() in men_base:
+                men_count = men_count + 1
+            elif w[:10].lower() in woman_base:
+                woman_count = woman_count + 1
+        if men_count >= woman_count:
+            result.append('1')
+        elif men_count < woman_count:
+            result.append('0')
+    return result
+
+
+def out_file(result, name):
+    with open(name, 'w') as file:
+        for program in result:
+            file.write(program + "\n")
+
+
+men_dict, woman_dict = database(men, woman)
+data = []
+with open('dev-0/in.tsv') as data:
+    for idx, line in enumerate(data.readlines()):
+        data.append(line.replace('\n', ''))
+
+result = predict(men_dict, woman_dict, data)
+out_file(result, 'dev-0/out.tsv')
+
+data2 = []
+with open('dev-1/in.tsv') as data:
+    for idx, line in enumerate(data.readlines()):
+        data2.append(line.replace('\n', ''))
+
+result = predict(men_dict, woman_dict, data2)
+out_file(result, 'dev-1/out.tsv')
+
+data3 = []
+with open('test-A/in.tsv') as data:
+    for idx, line in enumerate(data.readlines()):
+        data3.append(line.replace('\n', ''))
+
+result = predict(men_dict, woman_dict, data3)
+out_file(result, 'test-A/out.tsv')