From aa23d911f5d657495eb1a6e965b22c9ce94ffbc5 Mon Sep 17 00:00:00 2001 From: Jai Date: Wed, 26 Apr 2023 14:50:09 -0700 Subject: [PATCH] Update langchain.md --- docs/src/examples/langchain.md | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/docs/src/examples/langchain.md b/docs/src/examples/langchain.md index 636a53ad..cf79e44e 100644 --- a/docs/src/examples/langchain.md +++ b/docs/src/examples/langchain.md @@ -2,6 +2,8 @@ ## simple Pandas 2.0 documentation Q&A answering bot using LangChain + + To demonstrate using Lance, we’re going to build a simple Q&A answering bot using LangChain — an open-source framework that allows you to build composable LLM-based applications easily. We’ll use chat-langchain, a simple Q&A answering bot app as an example. Note: in this fork of chat-langchain, we’re also using a forked version of LangChain integration where we’ve built a Lance integration. The first step is to generate embeddings. You could build a bot using your own data, like a wiki page or internal documentation. For this example, we’re going to use the Pandas API documentation. LangChain offers document loaders to read and pre-process many document types. Since the Pandas API is in HTML, reading the docs is straightforward: @@ -12,4 +14,4 @@ for p in Path("./pandas.documentation").rglob("*.html"): continue loader = UnstructuredHTMLLoader(p) raw_document = loader.load() - docs = docs + raw_document \ No newline at end of file + docs = docs + raw_document