feat: refatorando arquivos com funções úteis do langchain

_utils/bubble_integrations/obter_arquivo.py

@@ -5,7 +5,7 @@ import tempfile
 import requests
 
 from _utils.handle_files import return_document_list_with_llama_parser
-from _utils.splitters.splitter_util import SplitterUtils
+from _utils.langchain_utils.splitter_util import SplitterUtils
 
 splitter_utils = SplitterUtils()
 headers = {"Authorization": f"Bearer {os.environ.get("BUBBLE_TOKEN")}"}

_utils/gerar_relatorio_modelo_usuario/DocumentSummarizer_simples.py

@@ -1,5 +1,5 @@
 from typing import List, Dict, Tuple, Optional
-from _utils.splitters.Splitter_class import Splitter
+from _utils.langchain_utils.Splitter_class import Splitter
 from setup.easy_imports import (
     HuggingFaceEmbeddings,
     Chroma,

_utils/gerar_relatorio_modelo_usuario/GerarDocumento.py

@@ -2,8 +2,8 @@ import os
 from typing import List, Dict, Tuple, Optional, cast
 
 from pydantic import SecretStr
-from _utils.LLMs.LLM_class import LLM
-from _utils.vector_stores.Vector_store_class import VectorStore
+from _utils.langchain_utils.LLM_class import LLM
+from _utils.langchain_utils.Vector_store_class import VectorStore
 from setup.easy_imports import (
     Chroma,
     ChatOpenAI,
@@ -20,7 +20,7 @@ from _utils.models.gerar_relatorio import (
     RetrievalConfig,
 )
 from cohere import Client
-from _utils.splitters.Splitter_class import Splitter
+from _utils.langchain_utils.Splitter_class import Splitter
 
 
 class GerarDocumento:

_utils/gerar_relatorio_modelo_usuario/contextual_retriever.py

@@ -1,5 +1,5 @@
 import os
-from _utils.LLMs.LLM_class import LLM
+from _utils.langchain_utils.LLM_class import LLM
 from _utils.gerar_relatorio_modelo_usuario.utils import (
     get_response_from_auxiliar_contextual_prompt,
     validate_many_chunks_in_one_request,

_utils/gerar_relatorio_modelo_usuario/llm_calls.py

@@ -2,7 +2,7 @@ import os
 from typing import Literal
 
 from pydantic import SecretStr
-from _utils.LLMs.LLM_class import LLM
+from _utils.langchain_utils.LLM_class import LLM
 from setup.environment import default_model
 from langchain_core.messages import HumanMessage
 from langchain_openai import ChatOpenAI
@@ -58,4 +58,4 @@ def gpt_answer(
 async def agemini_answer(prompt, model: Literal["gemini-2.0-flash", "gemini-2.0-flash-lite"]="gemini-2.0-flash"):
     gemini = llm.google_gemini(model)
     resposta = await gemini.ainvoke([HumanMessage(content=prompt)])
-    return resposta.content
+    return resposta.content

_utils/gerar_relatorio_modelo_usuario/utils.py

@@ -3,8 +3,8 @@ from langchain_core.documents import Document
 from langchain_core.messages import HumanMessage
 
 from _utils.gerar_relatorio_modelo_usuario.llm_calls import agemini_answer
-from _utils.splitters.Splitter_class import Splitter
-from _utils.LLMs.LLM_class import LLM
+from _utils.langchain_utils.Splitter_class import Splitter
+from _utils.langchain_utils.LLM_class import LLM
 from _utils.gerar_relatorio_modelo_usuario.prompts import (
     create_prompt_auxiliar_do_contextual_prompt,
 )
@@ -89,51 +89,6 @@ def validate_many_chunks_in_one_request(
 
 
 # Esta função gera a resposta que será usada em cada um das requisições de cada chunk
-# async def get_response_from_auxiliar_contextual_prompt(
-#     full_text_as_array: List[str], full_text_as_string: str
-# ):
-#     print("full_text_as_string: ", full_text_as_string)
-
-#     print("\nCRIANDO PROMPT AUXILIAR DO CONTEXTUAL")
-#     # PROMPT PARA GERAR O RESUMO INICIAL DO PROCESSO
-#     prompt_auxiliar_summary = create_prompt_auxiliar_do_contextual_prompt(
-#         full_text_as_string
-#     )
-
-#     import tiktoken
-#     import re
-
-#     # full_text_as_string = re.sub(r"\s+", " ", full_text_as_string).strip()
-
-#     encoding = tiktoken.get_encoding("cl100k_base")
-#     # Count tokens
-#     num_tokens = len(encoding.encode(full_text_as_string))
-#     with open("output.txt", "w", encoding="utf-8") as file:
-#         file.write(full_text_as_string)
-#     print(f"CONTAGEM DE TOKENS - {num_tokens}")
-
-#     print("\n\nprompt_auxiliar_summary[0:500]: ", prompt_auxiliar_summary[0:500])
-
-#     # Claude comentado pois o limite de tokens estava sendo passado pela requisição e dava erro
-#     # response_auxiliar_summary = await aclaude_answer(
-#     #     self.claude_client, self.claude_context_model, prompt_auxiliar_summary
-#     # )
-
-#     llms = LLM()
-#     print("\nCOMEÇANDO REQUISIÇÃO AUXILIAR DO CONTEXTUAL")
-#     response_auxiliar_summary = await llms.google_gemini().ainvoke(
-#         [HumanMessage(content=prompt_auxiliar_summary)]
-#     )
-#     print("TERMINOU REQUISIÇÃO AUXILIAR DO CONTEXTUAL")
-
-#     print(
-#         "\n\nresponse_auxiliar_summary.content[0:500]: ",
-#         response_auxiliar_summary.content[0:500],
-#     )
-
-#     return response_auxiliar_summary.content
-
-
 async def get_response_from_auxiliar_contextual_prompt(full_text_as_array: List[str]):
     import tiktoken
 

_utils/handle_files.py

@@ -4,7 +4,7 @@ from langchain_core.documents import Document as LangchainDocument
 from llama_index import Document
 from llama_parse import LlamaParse, ResultType
 
-from _utils.splitters.splitter_util import SplitterUtils
+from _utils.langchain_utils.splitter_util import SplitterUtils
 
 llama_parser_keys = [
     os.getenv("LLAMA_CLOUD_API_KEY_POPS"),

_utils/{splitters → langchain_utils}/Splitter_class.py

@@ -1,6 +1,6 @@
 from _utils.bubble_integrations.obter_arquivo import get_pdf_from_bubble
 from _utils.handle_files import return_document_list_with_llama_parser
-from _utils.splitters.splitter_util import (
+from _utils.langchain_utils.splitter_util import (
     SplitterUtils,
     combine_documents_without_losing_pagination,
 )

_utils/resumo_completo_cursor.py

@@ -1,7 +1,7 @@
 import os
 from langchain_core.messages import HumanMessage
 from typing import cast
-from _utils.LLMs.LLM_class import LLM
+from _utils.langchain_utils.LLM_class import LLM
 from _utils.bubble_integrations.enviar_resposta_final import enviar_resposta_final
 from _utils.custom_exception_handler import custom_exception_handler_wihout_api_handler
 from _utils.gerar_relatorio_modelo_usuario.prompts import (
@@ -24,7 +24,7 @@ from _utils.models.gerar_relatorio import (
 )
 import markdown
 
-from _utils.prompts.Prompt_class import Prompt
+from _utils.langchain_utils.Prompt_class import Prompt
 from _utils.utils import convert_markdown_to_HTML
 
 

gerar_documento/views.py

@@ -1,10 +1,10 @@
 from typing import Any, Dict, cast
 from langchain.prompts import PromptTemplate
-from _utils.LLMs.LLM_class import LLM
+from _utils.langchain_utils.LLM_class import LLM
 from _utils.gerar_relatorio_modelo_usuario.utils import (
     get_full_text_and_all_PDFs_chunks,
 )
-from _utils.prompts.Prompt_class import Prompt
+from _utils.langchain_utils.Prompt_class import Prompt
 from setup.easy_imports import (
     Response,
     AsyncAPIView,
@@ -24,7 +24,7 @@ from .serializer import (
     GerarEmentaSerializer,
 )
 import asyncio
-from _utils.splitters.Splitter_class import Splitter
+from _utils.langchain_utils.Splitter_class import Splitter
 
 
 class GerarDocumentoView(AsyncAPIView):

tests/LLMs/test_LLM_class.py

@@ -1,7 +1,7 @@
 import pytest
 import os
-from _utils.LLMs.LLM_class import LLM
-from _utils.splitters.Splitter_class import Splitter
+from _utils.langchain_utils.LLM_class import LLM
+from _utils.langchain_utils.Splitter_class import Splitter
 from _utils.models.gerar_relatorio import (
     DocumentChunk,
 )

tests/test_custom_exception_handler.py

@@ -1,6 +1,6 @@
 import pytest
 import os
-from _utils.splitters.Splitter_class import Splitter
+from _utils.langchain_utils.Splitter_class import Splitter
 from _utils.models.gerar_relatorio import (
     DocumentChunk,
 )
@@ -15,7 +15,7 @@ from django.core.handlers.wsgi import WSGIRequest
 #     @pytest.mark.asyncio
 #     def test_function_that_raises(self):
 #         client = APIClient()
-        
+
 #         invalid_data = {
 #             "name": "wrong property"
 #         }

tests/test_splitters.py

@@ -1,6 +1,6 @@
 import pytest
 import os
-from _utils.splitters.Splitter_class import Splitter
+from _utils.langchain_utils.Splitter_class import Splitter
 from _utils.models.gerar_relatorio import (
     DocumentChunk,
 )

tests/test_utils.py

@@ -1,6 +1,6 @@
 import pytest
 import os
-from _utils.splitters.Splitter_class import Splitter
+from _utils.langchain_utils.Splitter_class import Splitter
 from _utils.models.gerar_relatorio import (
     DocumentChunk,
 )