Reran pre-commit hooks

app/home.py

@@ -8,9 +8,7 @@
 
 from config.logging import setup_logging
 from config.settings import ENV
-from hackathon.streamlit.utils import (
-    check_password,
-)
+from hackathon.streamlit.utils import check_password
 
 get_logger = setup_logging()
 logger = get_logger(__name__)

hackathon/llm/chain_config.py

@@ -5,9 +5,7 @@
 from langchain.schema.output_parser import StrOutputParser
 from langchain_core.output_parsers.transform import BaseTransformOutputParser
 
-from hackathon.llm.prompts.core import (
-    PromptTemplate,
-)
+from hackathon.llm.prompts.core import PromptTemplate
 
 
 @dataclass

hackathon/llm/llm.py

@@ -36,9 +36,7 @@ def __init__(
         self, llm_model_path: str, stop_sequences: List[str] = ["ANSWER:"]
     ) -> None:
         self.llm_model_path = llm_model_path
-        self.callback_manager = CallbackManager(
-            [StreamingStdOutCallbackHandler()]
-        )
+        self.callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
         self.stop_sequences = stop_sequences
         self.initialise_llm()
 

hackathon/llm/llm_chains.py

@@ -27,10 +27,7 @@ def __init__(self, config: ChainConfig, llm: LLM):
             llm (LLM): to attach to the LLMChain
         """
         self.chain: RunnableSequence = (
-            config.var_input
-            | config.prompt
-            | llm.get_llm()
-            | config.out_parser()
+            config.var_input | config.prompt | llm.get_llm() | config.out_parser()
         )
 
     def invoke_query(self, query: Dict):

hackathon/llm/prompts/core.py

@@ -19,12 +19,7 @@
 
 
 _advice_prompt = (
-    B_INST
-    + B_SYS
-    + _advice_system_prompt
-    + E_SYS
-    + _advice_instruction
-    + E_INST
+    B_INST + B_SYS + _advice_system_prompt + E_SYS + _advice_instruction + E_INST
 )
 
 ADVICE_PROMPT = PromptTemplate.from_template(_advice_prompt)

hackathon/loader/processors.py

@@ -52,14 +52,10 @@ def _dataframe_process(self, df) -> List[Document]:
         else:
             df["metadata"] = {}
 
-        df["metadata"] = df["metadata"].apply(
-            lambda x: None if pd.isna(x) else x
-        )
+        df["metadata"] = df["metadata"].apply(lambda x: None if pd.isna(x) else x)
 
         docs = df.apply(
-            lambda row: Document(
-                page_content=row["content"], metadata=row["metadata"]
-            ),
+            lambda row: Document(page_content=row["content"], metadata=row["metadata"]),
             axis=1,
         ).tolist()
 
@@ -127,14 +123,8 @@ def get_processor(
         """
         extension = file_path.split(".")[-1].lower()
         if extension == "csv":
-            return CSVProcessor(
-                source_column, metadata_columns, content_columns
-            )
+            return CSVProcessor(source_column, metadata_columns, content_columns)
         elif extension == "parquet":
-            return ParquetProcessor(
-                source_column, metadata_columns, content_columns
-            )
+            return ParquetProcessor(source_column, metadata_columns, content_columns)
         else:
-            ValueError(
-                f"File type {extension} does not have a supported processor"
-            )
+            ValueError(f"File type {extension} does not have a supported processor")

hackathon/vectorstore/opensearch.py

@@ -22,9 +22,7 @@ def __init__(self, index_name, endpoint_name=None, region=None) -> None:
         self.client = None
         self.endpoint_name = endpoint_name
         self.index_name = index_name
-        self.opensearch_endpoint = self._get_opensearch_endpoint(
-            endpoint_name, region
-        )
+        self.opensearch_endpoint = self._get_opensearch_endpoint(endpoint_name, region)
         self.client = self.get_client()
 
     def get_client(self):
@@ -70,9 +68,7 @@ def _get_opensearch_endpoint(self, endpoint_name, region=None):
             self.http_auth = ("admin", "admin")
             return OPENSEARCH_URL
         client = boto3.client("es", region_name=region)
-        response = client.describe_elasticsearch_domain(
-            DomainName=endpoint_name
-        )
+        response = client.describe_elasticsearch_domain(DomainName=endpoint_name)
         return response["DomainStatus"]["Endpoints"]["vpc"]
 
     def _put_bulk_in_opensearch(self, docs):
@@ -111,9 +107,7 @@ def _create_index(self, index_name=None):
         if not index_name:
             index_name = self.index_name
         settings = {
-            "settings": {
-                "index": {"knn": True, "knn.space_type": "cosinesimil"}
-            }
+            "settings": {"index": {"knn": True, "knn.space_type": "cosinesimil"}}
         }
         response = self.client.indices.create(index=index_name, body=settings)
         return bool(response["acknowledged"])

hackathon/vectorstore/vectorstore.py

@@ -64,9 +64,7 @@ def get_documents(self, limit: Optional[int] = None):
         return self.vectorstore.get(limit=limit)
 
     def search_with_score(self, query: str):
-        return self.vectorstore.similarity_search_with_relevance_scores(
-            query, k=100
-        )
+        return self.vectorstore.similarity_search_with_relevance_scores(query, k=100)
 
     def similarity_search_with_filter(
         self, query: str, filter: Optional[Dict[str, str]]
@@ -159,9 +157,7 @@ def get_documents(self, limit: Optional[int] = None):
         return self.vectorstore.get(limit=limit)
 
     def retrieve_data_with_relevance_scores(self, query):
-        return self.vectorstore.similarity_search_with_relevance_scores(
-            query, k=10
-        )
+        return self.vectorstore.similarity_search_with_relevance_scores(query, k=10)
 
 
 class VectorStoreClient(ABC):
@@ -200,9 +196,7 @@ def __init__(self, embedding_function: Embeddings, name: str = "") -> None:
     def store_data(self, documents):
         logger.info("storing docs %s in Chroma", len(documents))
         response = self.client.add_documents(documents)
-        logger.info(
-            "Put docs %s in Chroma with response %s", len(documents), response
-        )
+        logger.info("Put docs %s in Chroma with response %s", len(documents), response)
         self.client.persist()
         return response
 
@@ -277,9 +271,7 @@ def store_data(self, documents: List[Document]):
         Args:
             Documents (list[Document]): list of documents to store into opensearch.
         """
-        logger.info(
-            f"Starting Opensearch ingestion of {len(documents)} documents"
-        )
+        logger.info(f"Starting Opensearch ingestion of {len(documents)} documents")
         for batch in self._split_into_batches(documents):
             try:
                 response = self.vectorstore.add_documents(batch)