docs[minor]: Add missing use cases (#5353)

* docs[minor]: Add missing use cases * chatbots * Query analysis * sql use case * graph use case * fix broken image path * fix img path * one more missing import
langchain-ai · May 14, 2024 · 7f9859e · 7f9859e
1 parent 534009d
commit 7f9859e
Show file tree

Hide file tree

Showing 18 changed files with 7,310 additions and 0 deletions.
diff --git a/docs/core_docs/docs/how_to/chatbots_memory.ipynb b/docs/core_docs/docs/how_to/chatbots_memory.ipynb
diff --git a/docs/core_docs/docs/how_to/chatbots_retrieval.ipynb b/docs/core_docs/docs/how_to/chatbots_retrieval.ipynb
diff --git a/docs/core_docs/docs/how_to/chatbots_tools.ipynb b/docs/core_docs/docs/how_to/chatbots_tools.ipynb
diff --git a/docs/core_docs/docs/how_to/extraction_long_text.ipynb b/docs/core_docs/docs/how_to/extraction_long_text.ipynb
diff --git a/docs/core_docs/docs/how_to/extraction_parse.ipynb b/docs/core_docs/docs/how_to/extraction_parse.ipynb
@@ -0,0 +1,328 @@
+{
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "id": "ea37db49-d389-4291-be73-885d06c1fb7e",
+      "metadata": {},
+      "source": [
+        "# How to do extraction without using function calling\n",
+        "\n",
+        "LLMs that are able to follow prompt instructions well can be tasked with outputting information in a given format without using function calling.\n",
+        "\n",
+        "This approach relies on designing good prompts and then parsing the output of the LLMs to make them extract information well, though it lacks some of the guarantees provided by function calling or JSON mode.\n",
+        "\n",
+        "Here, we'll use Claude which is great at following instructions! See [here for more about Anthropic models](/docs/integrations/chat/anthropic).\n",
+        "\n",
+        "First, we'll install the integration package:\n",
+        "\n",
+        "```{=mdx}\n",
+        "import IntegrationInstallTooltip from \"@mdx_components/integration_install_tooltip.mdx\";\n",
+        "import Npm2Yarn from \"@theme/Npm2Yarn\";\n",
+        "\n",
+        "<IntegrationInstallTooltip></IntegrationInstallTooltip>\n",
+        "\n",
+        "<Npm2Yarn>\n",
+        "  @langchain/anthropic zod zod-to-json-schema\n",
+        "</Npm2Yarn>\n",
+        "```"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 1,
+      "id": "d71b32de-a6b4-45ed-83a9-ba1925f9470c",
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "import { ChatAnthropic } from \"@langchain/anthropic\";\n",
+        "\n",
+        "const model = new ChatAnthropic({\n",
+        "  model: \"claude-3-sonnet-20240229\",\n",
+        "  temperature: 0,\n",
+        "})"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "id": "3e412374-3beb-4bbf-966b-400c1f66a258",
+      "metadata": {},
+      "source": [
+        ":::{.callout-tip}\n",
+        "All the same considerations for extraction quality apply for parsing approach. Review the [guidelines](/docs/use_cases/extraction/guidelines) for extraction quality.\n",
+        "\n",
+        "This tutorial is meant to be simple, but generally should really include reference examples to squeeze out performance!\n",
+        ":::"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "id": "abc1a945-0f80-4953-add4-cd572b6f2a51",
+      "metadata": {},
+      "source": [
+        "## Using StructuredOutputParser\n",
+        "\n",
+        "The following example uses the built-in [`StructuredOutputParser`](/docs/modules/model_io/output_parsers/types/structured) to parse the output of a chat model. We use the built-in prompt formatting instructions contained in the parser."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 2,
+      "id": "497eb023-c043-443d-ac62-2d4ea85fe1b0",
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "import { z } from \"zod\";\n",
+        "import { StructuredOutputParser } from \"langchain/output_parsers\";\n",
+        "import { ChatPromptTemplate } from \"@langchain/core/prompts\";\n",
+        "\n",
+        "const personSchema = z.object({\n",
+        "  name: z.optional(z.string()).describe(\"The name of the person\"),\n",
+        "  hair_color: z.optional(z.string()).describe(\"The color of the person's hair, if known\"),\n",
+        "  height_in_meters: z.optional(z.string()).describe(\"Height measured in meters\")\n",
+        "}).describe(\"Information about a person.\");\n",
+        "\n",
+        "const parser = StructuredOutputParser.fromZodSchema(personSchema);\n",
+        "\n",
+        "const prompt = ChatPromptTemplate.fromMessages([\n",
+        "  [\"system\", \"Answer the user query. Wrap the output in `json` tags\\n{format_instructions}\"],\n",
+        "  [\"human\", \"{query}\"],\n",
+        "]);\n",
+        "\n",
+        "const partialedPrompt = await prompt.partial({\n",
+        "  format_instructions: parser.getFormatInstructions(),\n",
+        "});"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "id": "c31aa2c8-05a9-4a12-80c5-ea1250dea0ae",
+      "metadata": {},
+      "source": [
+        "Let's take a look at what information is sent to the model"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 3,
+      "id": "20b99ffb-a114-49a9-a7be-154c525f8ada",
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "const query = \"Anna is 23 years old and she is 6 feet tall\";"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 4,
+      "id": "4f3a66ce-de19-4571-9e54-67504ae3fba7",
+      "metadata": {},
+      "outputs": [
+        {
+          "name": "stdout",
+          "output_type": "stream",
+          "text": [
+            "[\n",
+            "  SystemMessage {\n",
+            "    lc_serializable: true,\n",
+            "    lc_kwargs: {\n",
+            "      content: \"Answer the user query. Wrap the output in `json` tags\\n\" +\n",
+            "        \"You must format your output as a JSON value th\"... 1444 more characters,\n",
+            "      additional_kwargs: {}\n",
+            "    },\n",
+            "    lc_namespace: [ \"langchain_core\", \"messages\" ],\n",
+            "    content: \"Answer the user query. Wrap the output in `json` tags\\n\" +\n",
+            "      \"You must format your output as a JSON value th\"... 1444 more characters,\n",
+            "    name: undefined,\n",
+            "    additional_kwargs: {}\n",
+            "  },\n",
+            "  HumanMessage {\n",
+            "    lc_serializable: true,\n",
+            "    lc_kwargs: {\n",
+            "      content: \"Anna is 23 years old and she is 6 feet tall\",\n",
+            "      additional_kwargs: {}\n",
+            "    },\n",
+            "    lc_namespace: [ \"langchain_core\", \"messages\" ],\n",
+            "    content: \"Anna is 23 years old and she is 6 feet tall\",\n",
+            "    name: undefined,\n",
+            "    additional_kwargs: {}\n",
+            "  }\n",
+            "]\n"
+          ]
+        }
+      ],
+      "source": [
+        "const promptValue = await partialedPrompt.invoke({ query });\n",
+        "\n",
+        "console.log(promptValue.toChatMessages());"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 5,
+      "id": "3a46b5fd-9242-4b8c-a4e2-3f04fc19b3a4",
+      "metadata": {},
+      "outputs": [
+        {
+          "data": {
+            "text/plain": [
+              "{ name: \u001b[32m\"Anna\"\u001b[39m, hair_color: \u001b[32m\"\"\u001b[39m, height_in_meters: \u001b[32m\"1.83\"\u001b[39m }"
+            ]
+          },
+          "execution_count": 5,
+          "metadata": {},
+          "output_type": "execute_result"
+        }
+      ],
+      "source": [
+        "const chain = partialedPrompt.pipe(model).pipe(parser);\n",
+        "\n",
+        "await chain.invoke({ query });"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "id": "815b3b87-3bc6-4b56-835e-c6b6703cef5d",
+      "metadata": {},
+      "source": [
+        "## Custom Parsing\n",
+        "\n",
+        "You can also create a custom prompt and parser with `LangChain` and `LCEL`.\n",
+        "\n",
+        "You can use a raw function to parse the output from the model.\n",
+        "\n",
+        "In the below example, we'll pass the schema into the prompt as JSON schema. For convenience, we'll declare our schema with Zod, then use the [`zod-to-json-schema`](https://github.com/StefanTerdell/zod-to-json-schema) utility to convert it to JSON schema."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 6,
+      "id": "b1f11912-c1bb-4a2a-a482-79bf3996961f",
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "import { z } from \"zod\";\n",
+        "import { zodToJsonSchema } from \"zod-to-json-schema\";\n",
+        "\n",
+        "const personSchema = z.object({\n",
+        "  name: z.optional(z.string()).describe(\"The name of the person\"),\n",
+        "  hair_color: z.optional(z.string()).describe(\"The color of the person's hair, if known\"),\n",
+        "  height_in_meters: z.optional(z.string()).describe(\"Height measured in meters\")\n",
+        "}).describe(\"Information about a person.\");\n",
+        "\n",
+        "const peopleSchema = z.object({\n",
+        "  people: z.array(personSchema),\n",
+        "});\n",
+        "\n",
+        "const SYSTEM_PROMPT_TEMPLATE = [\n",
+        "  \"Answer the user's query. You must return your answer as JSON that matches the given schema:\",\n",
+        "  \"```json\\n{schema}\\n```.\",\n",
+        "  \"Make sure to wrap the answer in ```json and ``` tags. Conform to the given schema exactly.\",\n",
+        "].join(\"\\n\");\n",
+        "\n",
+        "const prompt = ChatPromptTemplate.fromMessages([\n",
+        "  [\"system\", SYSTEM_PROMPT_TEMPLATE],\n",
+        "  [\"human\", \"{query}\"],\n",
+        "]);\n",
+        "\n",
+        "const extractJsonFromOutput = (message) => {\n",
+        "  const text = message.content;\n",
+        "\n",
+        "  // Define the regular expression pattern to match JSON blocks\n",
+        "  const pattern = /```json\\s*((.|\\n)*?)\\s*```/gs;\n",
+        "\n",
+        "  // Find all non-overlapping matches of the pattern in the string\n",
+        "  const matches = pattern.exec(text);\n",
+        "\n",
+        "  if (matches && matches[1]) {\n",
+        "    try {\n",
+        "      return JSON.parse(matches[1].trim());\n",
+        "    } catch (error) {\n",
+        "      throw new Error(`Failed to parse: ${matches[1]}`);\n",
+        "    }\n",
+        "  } else {\n",
+        "    throw new Error(`No JSON found in: ${message}`);\n",
+        "  }\n",
+        "}"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 7,
+      "id": "cda52ef5-a354-47a7-9c25-45153c2389e2",
+      "metadata": {},
+      "outputs": [
+        {
+          "data": {
+            "text/plain": [
+              "\u001b[32m\"System: Answer the user's query. You must return your answer as JSON that matches the given schema:\\n\"\u001b[39m... 170 more characters"
+            ]
+          },
+          "execution_count": 7,
+          "metadata": {},
+          "output_type": "execute_result"
+        }
+      ],
+      "source": [
+        "const query = \"Anna is 23 years old and she is 6 feet tall\";\n",
+        "\n",
+        "const promptValue = await prompt.invoke({\n",
+        "  schema: zodToJsonSchema(peopleSchema),\n",
+        "  query\n",
+        "});\n",
+        "\n",
+        "promptValue.toString();"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 8,
+      "id": "993dc61a-229d-4795-a746-0d17df86b5c0",
+      "metadata": {},
+      "outputs": [
+        {
+          "data": {
+            "text/plain": [
+              "{ name: \u001b[32m\"Anna\"\u001b[39m, age: \u001b[33m23\u001b[39m, height: { feet: \u001b[33m6\u001b[39m, inches: \u001b[33m0\u001b[39m } }"
+            ]
+          },
+          "execution_count": 8,
+          "metadata": {},
+          "output_type": "execute_result"
+        }
+      ],
+      "source": [
+        "const chain = prompt.pipe(model).pipe(extractJsonFromOutput);\n",
+        "\n",
+        "await chain.invoke({\n",
+        "  schema: zodToJsonSchema(peopleSchema),\n",
+        "  query,\n",
+        "});"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "id": "fc6cd89f",
+      "metadata": {},
+      "outputs": [],
+      "source": []
+    }
+  ],
+  "metadata": {
+    "kernelspec": {
+      "display_name": "Deno",
+      "language": "typescript",
+      "name": "deno"
+    },
+    "language_info": {
+      "file_extension": ".ts",
+      "mimetype": "text/x.typescript",
+      "name": "typescript",
+      "nb_converter": "script",
+      "pygments_lexer": "typescript",
+      "version": "5.3.3"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 5
+}