Added language options for summary

Author: downloadfreeram

Dockerfile

@@ -20,7 +20,7 @@ COPY . .
 
 EXPOSE 8501
 
-#RUN "pytest -v app/tests/test_http.py app/tests/test_file.py app/tests/test_file_ext.py" > logs.log
+# RUN "pytest -v app/tests/test_http.py app/tests/test_file.py app/tests/test_file_ext.py" > logs.log
 
 
 CMD ["streamlit", "run", "app/main.py", "--server.port=8501", "--server.address=0.0.0.0"]

app/main.py

@@ -1,8 +1,9 @@
 # Main giga kox file
 import streamlit as st
 import time
-from src import FileProcessor, Summarizer
-from config.settings import settings, quality, length
+from src.file_processor import FileProcessor
+from src.summarizer import Summarizer
+from config.settings import settings, quality, length, language
 
 def main():
     if 'tests_run' not in st.session_state:
@@ -33,6 +34,10 @@ def main():
             "intentionally lower quality, possibly missing key points or slightly incoherent"
         ],
     )
+    sumLang = st.radio(
+        "Summary language",
+        ["Polish", "English", "German", "French"],
+    )
 
 
     fileProcessor = FileProcessor()
@@ -50,7 +55,7 @@ def main():
 
     if st.session_state.submitted:
         summarizzler = Summarizer(settings)
-        result = summarizzler.summarize(fileProcessor.get_content(), length[sumLenght], quality[sumQuality])
+        result = summarizzler.summarize(fileProcessor.get_content(), length[sumLenght], quality[sumQuality], language[sumLang])
         with st.spinner("Summarizing your doc...."):
             time.sleep(1)
         st.success("Your summary is ready:")

app/src/summarizer.py

@@ -11,15 +11,15 @@ def __init__(self, config: Settings):
             api_key=self.api_key,
         )
 
-    def summarize(self, text: str, sumLen: str, sumQuality: str) -> str:
+    def summarize(self, text: str, sumLen: str, sumQuality: str, lang: str) -> str:
         try:
             completion = self.client.chat.completions.create(
                 model = self.model,
                 messages=[
                     {
                         "role": "user",
                         "content": f"""
-                        Summarize the following text in the same language.
+                        Summarize the following text in this language:{lang}.
                         Length of the summary:{sumLen}
                         Quality of the summary:{sumQuality}
                         Text to summarize:\n\n{text}"""

app/tests/test_file.py

@@ -1,13 +1,14 @@
 from log import Logger
 import pytest
 #from ../main import main
-from ..src import FileProcessor, Summarizer
+from src.file_processor import FileProcessor
 
 logger = Logger(__name__).get_logger()
 
 def test_file_name():
     file_procesor = FileProcessor()
-    file_procesor.file = "manu.txt"
+    file_procesor.file = "test.txt"
+    file_procesor.file = "test.docx"
     #file_procesor.upload_file()
     ret = file_procesor.getFile()
     assert ret != None

app/tests/test_file_ext.py

@@ -1,27 +1,16 @@
 from log import Logger
 import pytest
 import os
-from src import FileProcessor
+from src.file_processor import FileProcessor
 
 logger = Logger(__name__).get_logger()
 
-# def test_file_ext(file_name: str) -> bool:
-#     supported_extensions = {".txt"}
-    
-#     _, ext = os.path.splitext(file_name)
-
-#     if any(file_name.endswith(ext) for ext in supported_extensions):
-#         logger.info(f"Type {ext} is supported")
-#         return True
-#     else:
-#         logger.error(f"Type {ext} isn't supported")
-#         return False
-
 @pytest.mark.parametrize("file_name, expected", [
     (".jpeg", False),
     (".pdf", False),
     (".zip", False),
     (".txt", True),
+    (".docx", True),
     (".png", False),
 ])
 

config/settings.py

@@ -21,4 +21,11 @@ def __init__(self):
     'long':'high coherence, accurate, well-structured',
     'mid':'average quality, some simplifications allowed',
     'short':'intentionally lower quality, possibly missing key points or slightly incoherent',
+}
+
+language = {
+    'Polish':'only Polish words',
+    'English':'only English words',
+    'German':'only German words',
+    'French':'only French words',
 }