format the code

src/sherpa_ai/connectors/base.py

@@ -1,5 +1,6 @@
 from abc import ABC, abstractmethod
 from typing import List
+
 from langchain.docstore.document import Document
 
 

src/sherpa_ai/memory/shared_memory_with_vectordb.py

@@ -3,6 +3,7 @@
 from langchain.embeddings.openai import OpenAIEmbeddings
 
 from sherpa_ai.actions.planning import Plan
+
 # from sherpa_ai.agents import AgentPool
 from sherpa_ai.connectors.base import BaseVectorDB
 from sherpa_ai.events import Event, EventType
@@ -24,9 +25,8 @@ def __init__(
         self,
         objective: str,
         session_id: str,
-        agent_pool:  None,
-        vectorStorage:BaseVectorDB = None,
-
+        agent_pool: None,
+        vectorStorage: BaseVectorDB = None,
     ):
         self.objective = objective
         self.agent_pool = agent_pool
@@ -43,7 +43,9 @@ def observe(self, belief: Belief):
 
         # based on the current task search similarity on the context and add it as an
         # event type user_input which is going to be used as a context on the prompt
-        contexts = self.vectorStorage.similarity_search(task.content, session_id=self.session_id)
+        contexts = self.vectorStorage.similarity_search(
+            task.content, session_id=self.session_id
+        )
         # Loop through the similarity search results, add the chunks as user_input events which will be added as a context in the belief class.
         for context in contexts:
             super().add(

src/sherpa_ai/utils.py

@@ -10,7 +10,7 @@
 from langchain.docstore.document import Document
 from langchain.document_loaders import UnstructuredMarkdownLoader, UnstructuredPDFLoader
 from langchain.llms import OpenAI
-from langchain.text_splitter import TokenTextSplitter
+from langchain.text_splitter import CharacterTextSplitter, TokenTextSplitter
 from loguru import logger
 from nltk.metrics import edit_distance, jaccard_distance
 from pypdf import PdfReader
@@ -19,7 +19,7 @@
 import sherpa_ai.config as cfg
 from sherpa_ai.database.user_usage_tracker import UserUsageTracker
 from sherpa_ai.models.sherpa_base_model import SherpaOpenAI
-from langchain.text_splitter import CharacterTextSplitter
+
 
 HTTP_GET_TIMEOUT = 2.5
 
@@ -121,7 +121,6 @@ def count_string_tokens(string: str, model_name: str) -> int:
 
 
 def chunk_and_summarize(text_data: str, question: str, link: str, llm):
-
     instruction = (
         "include any information that can be used to answer the "
         f"question '{question}' the given literal text is a data "
@@ -153,7 +152,6 @@ def chunk_and_summarize_file(
     llm,
     title: str = None,
 ):
-
     title = f",title {title} " if title is not None else ""
 
     instruction = (

src/tests/integration_tests/test_qa_agent_with_vector_shared_memory.py

@@ -3,6 +3,7 @@
 
 import pytest
 from langchain.chat_models import ChatOpenAI
+from langchain.docstore.document import Document
 from loguru import logger
 
 from sherpa_ai.agents import QAAgent
@@ -12,7 +13,6 @@
 from sherpa_ai.memory.shared_memory_with_vectordb import SharedMemoryWithVectorDB
 from sherpa_ai.test_utils.llms import get_llm
 from sherpa_ai.utils import file_text_splitter
-from langchain.docstore.document import Document
 
 
 data = """Avocados are a fruit, not a vegetable. They're technically considered a single-seeded berry, believe it or not.
@@ -65,6 +65,7 @@ def fake_embedding(input, default_dimension=1536):
             results.append([0] * default_dimension)
     return results
 
+
 @pytest.fixture
 def mock_chroma_vector_store(external_api):
     if external_api:
@@ -162,6 +163,29 @@ def test_shared_memory_with_vector(get_llm, mock_chroma_vector_store):  # noqa F
 
     results = shared_memory.get_by_type(EventType.result)
     logger.debug(results[0].content)
-    expected = ["Avocados", "Eiffel Tower", "Trypophobia", "Allodoxaphobia", "Australia", "Mellifluous", "Spice Girls", "Emma Bunton", "Human teeth", "guinea pig", "Ancient Romans", "shrimp", "Amy Poehler", "shower", "Baby rabbits", "unicorn", "aeroplane", "Venus", "Nutmeg", "French Burgundy"]
-    assert any(item in results[0].content for item in expected), "Result does not contain any expected items"
+    expected = [
+        "Avocados",
+        "Eiffel Tower",
+        "Trypophobia",
+        "Allodoxaphobia",
+        "Australia",
+        "Mellifluous",
+        "Spice Girls",
+        "Emma Bunton",
+        "Human teeth",
+        "guinea pig",
+        "Ancient Romans",
+        "shrimp",
+        "Amy Poehler",
+        "shower",
+        "Baby rabbits",
+        "unicorn",
+        "aeroplane",
+        "Venus",
+        "Nutmeg",
+        "French Burgundy",
+    ]
+    assert any(
+        item in results[0].content for item in expected
+    ), "Result does not contain any expected items"
     assert len(results) == 1

src/tests/unit_tests/scrape/test_prompt_reconstructor.py

@@ -5,6 +5,7 @@
 from sherpa_ai.scrape.prompt_reconstructor import PromptReconstructor
 from sherpa_ai.test_utils.llms import get_llm
 
+
 # Assuming that 'your_module' contains the 'PromptReconstructor' class
 
 
@@ -33,12 +34,12 @@ def test_reconstruct_prompt_with_link_inside_succeeds(
     mock_get_link_from_slack_client_conversation,
     mock_scrape_with_url,
     mock_chunk_and_summarize,
-    get_llm
+    get_llm,
 ):
     question = "Here's a <https://google.com>"
     slack_message = ""
     llm = get_llm(__file__, test_reconstruct_prompt_with_link_inside_succeeds.__name__)
-    reconstructor = PromptReconstructor(question, slack_message , llm)
+    reconstructor = PromptReconstructor(question, slack_message, llm)
     with patch(
         "sherpa_ai.scrape.prompt_reconstructor.chunk_and_summarize",
         return_value=mock_chunk_and_summarize,