[IMP] translate to russian

app.py

@@ -2,6 +2,8 @@ import warnings
 
 from langchain_core._api import LangChainDeprecationWarning
 
+import translator
+
 warnings.filterwarnings("ignore", category=DeprecationWarning)
 warnings.filterwarnings("ignore", category=UserWarning)
 warnings.filterwarnings("ignore", category=LangChainDeprecationWarning)
@@ -23,7 +25,7 @@ prev_files = None
 retriever = None
 
 
-def handle_files_and_query(query, files, chunk_overlap=50, token_per_chunk=256, bm_25_answers=200):
+def handle_files_and_query(query, files, chunk_overlap=50, token_per_chunk=256, bm_25_answers=200, translate_to_ru=False):
     results = ""
     global prev_files, retriever
     files = [f.name for f in files]
@@ -51,10 +53,16 @@ def handle_files_and_query(query, files, chunk_overlap=50, token_per_chunk=256,
         reranked_results = FAISS.from_documents(search_results, embeddings,
                                                 distance_strategy=DistanceStrategy.COSINE).similarity_search(query,
                                                                                                              k=25)
-        results = "\n".join([
-            f"Source: {re.search(pattern, result.metadata['file_path']).group(0)}\nPage: {result.metadata['page']}\nContent:\n{result.page_content}\n"
-            for result in reranked_results
-        ])
+        if translate_to_ru:
+            results = "\n".join([
+                f"Source: {re.search(pattern, result.metadata['file_path']).group(0)}\nPage: {result.metadata['page']}\nContent:\n{translator.translate(result.page_content, 'russian')}\n"
+                for result in reranked_results
+            ])
+        else:
+            results = "\n".join([
+                f"Source: {re.search(pattern, result.metadata['file_path']).group(0)}\nPage: {result.metadata['page']}\nContent:\n{result.page_content}\n"
+                for result in reranked_results
+            ])
     return results
 
 
@@ -66,7 +74,8 @@ interface = gr.Interface(
         gr.Slider(minimum=1, maximum=100, value=50, label="Chunk Overlap"),
         gr.Slider(minimum=64, maximum=512, value=256, label="Tokens Per Chunk (чем больше - тем бОльшие куски книги "
                                                             "сможем находить)"),
-        gr.Slider(minimum=1, maximum=1000, value=200, label="BM25 Answers (чем больше - тем больше будем учитывать неявные смысловые сравнения слов)")
+        gr.Slider(minimum=1, maximum=1000, value=200, label="BM25 Answers (чем больше - тем больше будем учитывать неявные смысловые сравнения слов)"),
+        gr.Checkbox(label="Translate to Russian", value=False),
     ],
     outputs="text",
     title="Similarity Search for eksmo books"

translator.py

@@ -0,0 +1,14 @@
+import openai
+
+
+def translate(text, target_lang):
+    """Translate text to target language using OpenAI's GPT-3 API."""
+    client = openai.Client()
+    response = client.chat.completions.create(
+        messages=[{"role": "system", "content": f"You are AI-translator and you should translate text to {target_lang}"},
+                  {'role': 'user', 'content': f'Please translate this text to {target_lang}: {text}. '
+                                              f'Answer with tranlsatrion and no additional information.'},
+                  ],
+        model="gpt-3.5-turbo",
+    )
+    return response.choices[0].message.content