chore: Update langchain_community.vectorstores imports and add Chroma vectorstore

.gitattributes

@@ -34,6 +34,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 langchain_faiss/* filter=lfs diff=lfs merge=lfs -text
+langchain_chroma/* filter=lfs diff=lfs merge=lfs -text
 *.pdf filter=lfs diff=lfs merge=lfs -text
 *.faiss filter=lfs diff=lfs merge=lfs -text
 *.msg filter=lfs diff=lfs merge=lfs -text

app.py

@@ -16,7 +16,7 @@ from langchain_community.document_loaders.parsers.language.language_parser impor
     LanguageParser,
 )
 from langchain_community.embeddings import HuggingFaceBgeEmbeddings
-from langchain_community.vectorstores import FAISS
+from langchain_community.vectorstores import FAISS, Chroma
 from langchain_core.callbacks.manager import CallbackManager
 from langchain_core.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
 from langchain_core.output_parsers import StrOutputParser
@@ -129,18 +129,37 @@ cached_embeddings = CacheBackedEmbeddings.from_bytes_store(
 
 # Create and save FAISS index
 FAISS_DB_INDEX = "./langchain_faiss"
-# db = FAISS.from_documents(combined_documents, cached_embeddings)
-# db.save_local(folder_path=FAISS_DB_INDEX)
-db = FAISS.load_local(
+# faiss_db = FAISS.from_documents(
+#     documents=combined_documents,
+#     embedding=cached_embeddings,
+# )
+# faiss_db.save_local(folder_path=FAISS_DB_INDEX)
+
+# Create and save Chroma index
+CHROMA_DB_INDEX = "./langchain_chroma"
+# chroma_db = Chroma.from_documents(
+#     documents=combined_documents,
+#     embedding=cached_embeddings,
+#     persist_directory=CHROMA_DB_INDEX,
+# )
+
+# load vectorstore
+faiss_db = FAISS.load_local(
     FAISS_DB_INDEX, cached_embeddings, allow_dangerous_deserialization=True
 )
+chroma_db = Chroma(
+    embedding_function=cached_embeddings,
+    persist_directory=CHROMA_DB_INDEX,
+)
 
 # Create retrievers
-faiss_retriever = db.as_retriever(search_type="mmr", search_kwargs={"k": 10})
+faiss_retriever = faiss_db.as_retriever(search_type="mmr", search_kwargs={"k": 10})
+chroma_retriever = chroma_db.as_retriever(search_type="mmr", search_kwargs={"k": 10})
 bm25_retriever = BM25Retriever.from_documents(combined_documents)
 bm25_retriever.k = 10
 ensemble_retriever = EnsembleRetriever(
-    retrievers=[bm25_retriever, faiss_retriever], weights=[0.5, 0.5], search_type="mmr"
+    retrievers=[bm25_retriever, faiss_retriever, chroma_retriever],
+    weights=[0.4, 0.3, 0.3],
 )
 
 # Create prompt template
@@ -172,11 +191,14 @@ class StreamCallback(BaseCallbackHandler):
         print(token, end="", flush=True)
 
 
+streaming = os.getenv("STREAMING", "true") == "true"
+print("STREAMING", streaming)
+
 # Initialize LLMs with configuration
 llm = ChatOpenAI(
     model="gpt-4o",
     temperature=0,
-    streaming=True,
+    streaming=streaming,
     callbacks=[StreamCallback()],
 ).configurable_alternatives(
     ConfigurableField(id="llm"),
@@ -220,11 +242,11 @@ rag_chain = (
 )
 
 
-model_key = os.getenv("LLM_MODEL", "gpt4")
-print("model", model_key)
+model_key = os.getenv("MODEL_KEY", "gemini")
+print("MODEL_KEY", model_key)
 
 
-def respond(
+def respond_stream(
     message,
     history: list[tuple[str, str]],
 ):
@@ -234,11 +256,18 @@ def respond(
         yield response
 
 
+def respond(
+    message,
+    history: list[tuple[str, str]],
+):
+    return rag_chain.with_config(configurable={"llm": model_key}).invoke(message)
+
+
 """
 For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
 """
 demo = gr.ChatInterface(
-    respond,
+    respond_stream if streaming else respond,
     title="랭체인에 대해서 물어보세요!",
     description="안녕하세요!\n저는 랭체인에 대한 인공지능 QA봇입니다. 랭체인에 대해 깊은 지식을 가지고 있어요. 랭체인 개발에 관한 도움이 필요하시면 언제든지 질문해주세요!",
 )

langchain_chroma/chroma.sqlite3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:825e12c771f36b9c3e1f7aa3410a662227600a7c2c361960a0369b29b633871f
+size 189218816

langchain_chroma/dcb5b5f5-3dac-4239-b657-276cb0bca164/data_level0.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9f756c81ce3c4911b5c264c5332a0f1dcd2b3bc98dd84dbdb94d01bd9cf927b
+size 63540000

langchain_chroma/dcb5b5f5-3dac-4239-b657-276cb0bca164/header.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9fcf6e0965f08173cd9d1d2b68e7d8c9a6971bd865faa9a8cfd8ba994cccc64b
+size 100

langchain_chroma/dcb5b5f5-3dac-4239-b657-276cb0bca164/index_metadata.pickle

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff4868df92f9bdadacd1abe27aef8cdb3704a0b16304fdd33ae539ce6c39d371
+size 868175

langchain_chroma/dcb5b5f5-3dac-4239-b657-276cb0bca164/length.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57bba9bb25e471e0266df7225b0e7b12ae290d7237cbf2cf16cb3f9503c4a75d
+size 60000

langchain_chroma/dcb5b5f5-3dac-4239-b657-276cb0bca164/link_lists.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b31f9bfa91a0635928b7b7f2b64977111f7f8fb6207ed42b3d0443518680d5c
+size 132080

langchain_faiss/index.faiss

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f50b9cdc2968dd1fe5875e7e1f8ed2689a3e938d505a0e2f06b5257083339bd2
-size 2621485
+oid sha256:7794ae040ff5830ba667d78cd4a1648e52e8bd860175f68a74ab85ea56ab3bb2
+size 61632557

langchain_faiss/index.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ba84841f2d61493243e47d654cff88c8864c5fa6119469b7569677e4f82f3c5f
-size 862597
+oid sha256:2a9c82717db2899e9cdd3196f4806a0807167fe9e7892e3a2c00b5b5c6cb6c9b
+size 19914288