fix elastic rag template in playground

templates/chat-bot-feedback/chat_bot_feedback/chain.py

@@ -165,7 +165,6 @@ def format_chat_history(chain_input: dict) -> dict:
 # with the new `tool.langserve.export_attr`
 chain = (
     (format_chat_history | _prompt | _model | StrOutputParser())
-    .with_types(input_type=ChainInput)
     # This is to add the evaluators as "listeners"
     # and to customize the name of the chain.
     # Any chain that accepts a compatible input type works here.
@@ -180,3 +179,5 @@ def format_chat_history(chain_input: dict) -> dict:
         ],
     )
 )
+
+chain = chain.with_types(input_type=ChainInput)

templates/rag-elasticsearch/README.md

@@ -1,7 +1,7 @@
 
 # rag-elasticsearch
 
-This template performs RAG using ElasticSearch.
+This template performs RAG using [ElasticSearch](https://python.langchain.com/docs/integrations/vectorstores/elasticsearch).
 
 It relies on sentence transformer `MiniLM-L6-v2` for embedding passages and questions.
 
@@ -19,7 +19,12 @@ export ELASTIC_PASSWORD = <ClOUD_PASSWORD>
 For local development with Docker, use:
 
 ```bash
-export ES_URL = "http://localhost:9200"
+export ES_URL="http://localhost:9200"
+```
+
+And run an Elasticsearch instance in Docker with
+```bash
+docker run -p 9200:9200 -e "discovery.type=single-node" -e "xpack.security.enabled=false" -e "xpack.security.http.ssl.enabled=false" docker.elastic.co/elasticsearch/elasticsearch:8.9.0
 ```
 
 ## Usage
@@ -83,7 +88,7 @@ runnable = RemoteRunnable("http://localhost:8000/rag-elasticsearch")
 For loading the fictional workplace documents, run the following command from the root of this repository:
 
 ```bash
-python ./data/load_documents.py
+python ingest.py
 ```
 
 However, you can choose from a large number of document loaders [here](https://python.langchain.com/docs/integrations/document_loaders).  

templates/rag-elasticsearch/rag_elasticsearch/chain.py

@@ -1,12 +1,13 @@
 from operator import itemgetter
-from typing import List, Tuple
+from typing import List, Optional, Tuple
 
 from langchain.chat_models import ChatOpenAI
 from langchain.embeddings import HuggingFaceEmbeddings
-from langchain.schema import format_document
+from langchain.schema import BaseMessage, format_document
 from langchain.schema.output_parser import StrOutputParser
 from langchain.schema.runnable import RunnableMap, RunnablePassthrough
 from langchain.vectorstores.elasticsearch import ElasticsearchStore
+from pydantic import BaseModel, Field
 
 from .connection import es_connection_details
 from .prompts import CONDENSE_QUESTION_PROMPT, DOCUMENT_PROMPT, LLM_CONTEXT_PROMPT
@@ -41,6 +42,13 @@ def _format_chat_history(chat_history: List[Tuple]) -> str:
     return buffer
 
 
+class ChainInput(BaseModel):
+    chat_history: Optional[List[BaseMessage]] = Field(
+        description="Previous chat messages."
+    )
+    question: str = Field(..., description="The question to answer.")
+
+
 _inputs = RunnableMap(
     standalone_question=RunnablePassthrough.assign(
         chat_history=lambda x: _format_chat_history(x["chat_history"])
@@ -56,3 +64,5 @@ def _format_chat_history(chat_history: List[Tuple]) -> str:
 }
 
 chain = _inputs | _context | LLM_CONTEXT_PROMPT | llm | StrOutputParser()
+
+chain = chain.with_types(input_type=ChainInput)