updating to use jsonb

Author: skainec-focusedlabs

importer/load_and_process.py

@@ -1,7 +1,7 @@
 import os
 from dotenv import load_dotenv
 from langchain_community.document_loaders import DirectoryLoader, UnstructuredPDFLoader
-from langchain_community.embeddings import OpenAIEmbeddings
+from langchain_openai import OpenAIEmbeddings
 from langchain_community.vectorstores import PGVector
 from langchain_experimental.text_splitter import SemanticChunker
 
@@ -15,7 +15,7 @@
     use_multithreading=True,
     show_progress=True,
     max_concurrency=50,
-    loader_cls=UnstructuredPDFLoader
+    loader_cls=UnstructuredPDFLoader,
 )
 
 docs = loader.load()
@@ -31,4 +31,5 @@
     collection_name=PG_COLLECTION_NAME,
     connection_string=os.getenv("POSTGRES_URL"),
     pre_delete_collection=True,
+    use_jsonb=True,
 )

langsmith_testing/dataset_evaluation.py

@@ -18,6 +18,11 @@ def predict(inputs: dict) -> dict:
 
 
 def format_evaluator_inputs(run: Run, example: Example):
+    # return StringEvaluatorInput(
+    #     input=example.inputs["inputs"],
+    #     prediction=run.outputs.values(),
+    #     reference=example.outputs["expected"],
+    # )
     return {
         "input": example.inputs["inputs"],
         "prediction": next(iter(run.outputs.values())),

modules/vector_store.py

@@ -10,4 +10,5 @@
     collection_name=PG_COLLECTION_NAME,
     connection_string=os.getenv("POSTGRES_URL"),
     embedding_function=OpenAIEmbeddings(model=EMBEDDING_MODEL),
+    use_jsonb=True,
 )