Initial commit

.gitignore

@@ -0,0 +1,6 @@
+__pycache__
+.vscode
+venv
+.env
+/data/*
+*.session

grab_samples.py

@@ -0,0 +1,34 @@
+import os
+from dotenv import load_dotenv
+from telethon.sync import TelegramClient
+
+load_dotenv()
+
+api_id = os.getenv('API_ID')
+api_hash = os.getenv('API_HASH')
+chat = 'tg1337const'
+
+def cleanup(msg):
+  unwanted_strings = [
+    '[1337const | подписаться](https://t.me/tg1337const)',
+    '[1337const | Подписаться](https://t.me/tg1337const)'
+  ]
+
+  try:
+    for unwanted in unwanted_strings:
+      msg = msg.replace(unwanted, '')
+
+    return msg
+  except:
+    return msg
+
+
+with TelegramClient('session_name', api_id, api_hash) as client:
+  f = open("data/data.txt", "w")
+  for message in client.iter_messages(chat):
+    try:
+      f.write(cleanup(message.text))
+    except:
+      pass
+
+  f.close()

main.py

@@ -0,0 +1,10 @@
+import markovify
+
+f = open("data/data.txt", "r")
+text = f.read()
+
+text_model = markovify.Text(text)
+
+for i in range(3):
+  # print(text_model.make_short_sentence(280))
+  print(text_model.make_sentence())

requirements.txt

@@ -0,0 +1,3 @@
+markovify
+telethon
+python-dotenv