From 922193475ad9f59fb7041a53cf30c242f4711f35 Mon Sep 17 00:00:00 2001
From: Bagatur <22008038+baskaryan@users.noreply.github.com>
Date: Tue, 24 Oct 2023 11:52:47 -0400
Subject: [PATCH] Docs: Add LCEL to chains/foundational/transform (#12212)

---
 .../chains/foundational/transformation.ipynb  | 101 ++++++++++++++----
 1 file changed, 78 insertions(+), 23 deletions(-)

diff --git a/docs/docs/modules/chains/foundational/transformation.ipynb b/docs/docs/modules/chains/foundational/transformation.ipynb
index 9114ab7359..f7caa2b9f9 100644
--- a/docs/docs/modules/chains/foundational/transformation.ipynb
+++ b/docs/docs/modules/chains/foundational/transformation.ipynb
@@ -7,21 +7,27 @@
    "source": [
     "# Transformation\n",
     "\n",
-    "This notebook showcases using a generic transformation chain.\n",
+    "Often we want to transform inputs as they are passed from one component to another.\n",
     "\n",
-    "As an example, we will create a dummy transformation that takes in a super long text, filters the text to only the first 3 paragraphs, and then passes that into an `LLMChain` to summarize those."
+    "As an example, we will create a dummy transformation that takes in a super long text, filters the text to only the first 3 paragraphs, and then passes that into a chain to summarize those."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
-   "id": "bbbb4330",
+   "execution_count": 2,
+   "id": "d257f50d-c53d-41b7-be8a-df23fbd7c017",
    "metadata": {},
    "outputs": [],
    "source": [
-    "from langchain.chains import TransformChain, LLMChain, SimpleSequentialChain\n",
-    "from langchain.llms import OpenAI\n",
-    "from langchain.prompts import PromptTemplate"
+    "from langchain.prompts import PromptTemplate\n",
+    "\n",
+    "prompt = PromptTemplate.from_template(\n",
+    "    \"\"\"Summarize this text:\n",
+    "\n",
+    "{output_text}\n",
+    "\n",
+    "Summary:\"\"\"\n",
+    ")"
    ]
   },
   {
@@ -35,9 +41,67 @@
     "    state_of_the_union = f.read()"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "4c938536-e3fb-45eb-a1b3-cb82be410e32",
+   "metadata": {},
+   "source": [
+    "## Using LCEL\n",
+    "\n",
+    "With LCEL this is trivial, since we can add functions in any `RunnableSequence`."
+   ]
+  },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 17,
+   "id": "1e53e851-b1bd-424f-a144-5f2e8b413dcf",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'The speaker acknowledges the presence of important figures in the government and addresses the audience as fellow Americans. They highlight the impact of COVID-19 on keeping people apart in the previous year but express joy in being able to come together again. The speaker emphasizes the unity of Democrats, Republicans, and Independents as Americans.'"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from langchain.chat_models import ChatOpenAI\n",
+    "from langchain.schema import StrOutputParser\n",
+    "\n",
+    "runnable = {\"output_text\": lambda text: \"\\n\\n\".join(text.split(\"\\n\\n\")[:3])} | prompt | ChatOpenAI() | StrOutputParser()\n",
+    "runnable.invoke(state_of_the_union)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a9b9bd07-155f-4777-9215-509d39ecfe3f",
+   "metadata": {},
+   "source": [
+    "## [Legacy] TransformationChain\n",
+    "\n",
+    "::note:: This is a legacy class, using LCEL as shown above is preffered.\n",
+    "\n",
+    "This notebook showcases using a generic transformation chain."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "bbbb4330",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from langchain.chains import TransformChain, LLMChain, SimpleSequentialChain\n",
+    "from langchain.llms import OpenAI\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
    "id": "98739592",
    "metadata": {},
    "outputs": [],
@@ -47,7 +111,6 @@
     "    shortened_text = \"\\n\\n\".join(text.split(\"\\n\\n\")[:3])\n",
     "    return {\"output_text\": shortened_text}\n",
     "\n",
-    "\n",
     "transform_chain = TransformChain(\n",
     "    input_variables=[\"text\"], output_variables=[\"output_text\"], transform=transform_func\n",
     ")"
@@ -55,7 +118,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 14,
    "id": "e9397934",
    "metadata": {},
    "outputs": [],
@@ -71,7 +134,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 15,
    "id": "06f51f17",
    "metadata": {},
    "outputs": [],
@@ -81,17 +144,17 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 16,
    "id": "f7caa1ee",
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "' The speaker addresses the nation, noting that while last year they were kept apart due to COVID-19, this year they are together again. They are reminded that regardless of their political affiliations, they are all Americans.'"
+       "' In an address to the nation, the speaker acknowledges the hardships of the past year due to the COVID-19 pandemic, but emphasizes that regardless of political affiliation, all Americans can come together.'"
       ]
      },
-     "execution_count": 7,
+     "execution_count": 16,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -99,14 +162,6 @@
    "source": [
     "sequential_chain.run(state_of_the_union)"
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "e3ca6409",
-   "metadata": {},
-   "outputs": [],
-   "source": []
   }
  ],
  "metadata": {
@@ -125,7 +180,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.3"
+   "version": "3.9.1"
   }
  },
  "nbformat": 4,