remi/Perplexica
1
0
Fork 0
mirror of https://github.com/ItzCrazyKns/Perplexica.git synced 2025-08-11 10:18:41 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat(meta-search-agent): implement basic function calling with langgraph

Browse Source
This commit is contained in:
ItzCrazyKns
2025-08-04 21:47:55 +05:30
parent 653ce3cb19
commit 8811398040
3 changed files with 325 additions and 282 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
src/lib/prompts/webSearch.ts
View File

@ -65,12 +65,16 @@ export const webSearchResponsePrompt = `
You are Perplexica, an AI model skilled in web search and crafting detailed, engaging, and well-structured answers. You excel at summarizing web pages and extracting relevant information to create professional, blog-style responses.
Your task is to provide answers that are:
- **Informative and relevant**: Thoroughly address the user's query using the given context.
- **Informative and relevant**: Thoroughly address the user's query using the given search results.
- **Well-structured**: Include clear headings and subheadings, and use a professional tone to present information concisely and logically.
- **Engaging and detailed**: Write responses that read like a high-quality blog post, including extra details and relevant insights.
- **Cited and credible**: Use inline citations with [number] notation to refer to the context source(s) for each fact or detail included.
- **Cited and credible**: Use inline citations with [number] notation to refer to the search results source(s) for each fact or detail included.
- **Explanatory and Comprehensive**: Strive to explain the topic in depth, offering detailed analysis, insights, and clarifications wherever applicable.
You have access to the following tools which you have to use to answer the user:
1. **search_web**: Use this tool to search the web for information to answer the user's question.
2. **summarize**: Use this tool to summarize a link.
### Formatting Instructions
- **Structure**: Use a well-organized format with proper headings (e.g., "## Example heading 1" or "## Example heading 2"). Present information in paragraphs or concise bullet points where appropriate.
- **Tone and Style**: Maintain a neutral, journalistic tone with engaging narrative flow. Write as though you're crafting an in-depth article for a professional audience.
@ -80,13 +84,14 @@ export const webSearchResponsePrompt = `
- **Conclusion or Summary**: Include a concluding paragraph that synthesizes the provided information or suggests potential next steps, where appropriate.
### Citation Requirements
- Cite every single fact, statement, or sentence using [number] notation corresponding to the source from the provided \`context\`.
- Cite every single fact, statement, or sentence using [number] notation corresponding to the source from the search results
- Integrate citations naturally at the end of sentences or clauses as appropriate. For example, "The Eiffel Tower is one of the most visited landmarks in the world[1]."
- Ensure that **every sentence in your response includes at least one citation**, even when information is inferred or connected to general knowledge available in the provided context.
- Ensure that **every sentence in your response includes at least one citation**, even when information is inferred or connected to general knowledge available in the provided search results.
- Use multiple sources for a single detail if applicable, such as, "Paris is a cultural hub, attracting millions of visitors annually[1][2]."
- Always prioritize credibility and accuracy by linking all statements back to their respective context sources.
- Always prioritize credibility and accuracy by linking all statements back to their respective search results sources.
- Avoid citing unsupported assumptions or personal interpretations; if no source supports a statement, clearly indicate the limitation.
- Never return references to the citation or sources yourself, they're returned to the user internally.
### Special Instructions
- If the query involves technical, historical, or complex topics, provide detailed background and explanatory sections to ensure clarity.
- If the user provides vague input or if relevant information is missing, explain what additional details might help refine the search.
@ -99,12 +104,8 @@ export const webSearchResponsePrompt = `
### Example Output
- Begin with a brief introduction summarizing the event or query topic.
- Follow with detailed sections under clear headings, covering all aspects of the query if possible.
- Provide explanations or historical context as needed to enhance understanding.
- Provide explanations or historical search results as needed to enhance understanding.
- End with a conclusion or overall perspective if relevant.
<context>
{context}
</context>
Current date & time in ISO format (UTC timezone) is: {date}.
`;

4
src/lib/prompts/writingAssistant.ts
View File

@ -10,8 +10,4 @@ However you do not need to cite it using the same number. You can use different
### User instructions
These instructions are shared to you by the user and not by the system. You will have to follow them but give them less priority than the above instructions. If the user has provided specific instructions or preferences, incorporate them into your response while adhering to the overall guidelines.
{systemInstructions}
<context>
{context}
</context>
`;

580
src/lib/search/metaSearchAgent.ts
View File

@ -11,7 +11,12 @@ import {
RunnableMap,
RunnableSequence,
} from '@langchain/core/runnables';
import { BaseMessage } from '@langchain/core/messages';
import {
AIMessage,
BaseMessage,
isAIMessage,
ToolMessage,
} from '@langchain/core/messages';
import { StringOutputParser } from '@langchain/core/output_parsers';
import LineListOutputParser from '../outputParsers/listLineOutputParser';
import LineOutputParser from '../outputParsers/lineOutputParser';
@ -24,6 +29,15 @@ import computeSimilarity from '../utils/computeSimilarity';
import formatChatHistoryAsString from '../utils/formatHistory';
import eventEmitter from 'events';
import { StreamEvent } from '@langchain/core/tracers/log_stream';
import { DynamicStructuredTool, tool } from '@langchain/core/tools';
import { MessagesAnnotation, StateGraph } from '@langchain/langgraph';
import { z } from 'zod';
import {
IterableReadableStream,
IterableReadableStreamInterface,
} from '@langchain/core/utils/stream';
import EventEmitter from 'node:events';
import { BaseLanguageModel } from '@langchain/core/language_models/base';
export interface MetaSearchAgentType {
searchAndAnswer: (
@ -60,239 +74,6 @@ class MetaSearchAgent implements MetaSearchAgentType {
this.config = config;
}
private async createSearchRetrieverChain(llm: BaseChatModel) {
(llm as unknown as ChatOpenAI).temperature = 0;
return RunnableSequence.from([
PromptTemplate.fromTemplate(this.config.queryGeneratorPrompt),
llm,
this.strParser,
RunnableLambda.from(async (input: string) => {
const linksOutputParser = new LineListOutputParser({
key: 'links',
});
const questionOutputParser = new LineOutputParser({
key: 'question',
});
const links = await linksOutputParser.parse(input);
let question = this.config.summarizer
? await questionOutputParser.parse(input)
: input;
if (question === 'not_needed') {
return { query: '', docs: [] };
}
if (links.length > 0) {
if (question.length === 0) {
question = 'summarize';
}
let docs: Document[] = [];
const linkDocs = await getDocumentsFromLinks({ links });
const docGroups: Document[] = [];
linkDocs.map((doc) => {
const URLDocExists = docGroups.find(
(d) =>
d.metadata.url === doc.metadata.url &&
d.metadata.totalDocs < 10,
);
if (!URLDocExists) {
docGroups.push({
...doc,
metadata: {
...doc.metadata,
totalDocs: 1,
},
});
}
const docIndex = docGroups.findIndex(
(d) =>
d.metadata.url === doc.metadata.url &&
d.metadata.totalDocs < 10,
);
if (docIndex !== -1) {
docGroups[docIndex].pageContent =
docGroups[docIndex].pageContent + `\n\n` + doc.pageContent;
docGroups[docIndex].metadata.totalDocs += 1;
}
});
await Promise.all(
docGroups.map(async (doc) => {
const res = await llm.invoke(`
You are a web search summarizer, tasked with summarizing a piece of text retrieved from a web search. Your job is to summarize the
text into a detailed, 2-4 paragraph explanation that captures the main ideas and provides a comprehensive answer to the query.
If the query is \"summarize\", you should provide a detailed summary of the text. If the query is a specific question, you should answer it in the summary.
- **Journalistic tone**: The summary should sound professional and journalistic, not too casual or vague.
- **Thorough and detailed**: Ensure that every key point from the text is captured and that the summary directly answers the query.
- **Not too lengthy, but detailed**: The summary should be informative but not excessively long. Focus on providing detailed information in a concise format.
The text will be shared inside the \`text\` XML tag, and the query inside the \`query\` XML tag.
<example>
1. \`<text>
Docker is a set of platform-as-a-service products that use OS-level virtualization to deliver software in packages called containers.
It was first released in 2013 and is developed by Docker, Inc. Docker is designed to make it easier to create, deploy, and run applications
by using containers.
</text>
<query>
What is Docker and how does it work?
</query>
Response:
Docker is a revolutionary platform-as-a-service product developed by Docker, Inc., that uses container technology to make application
deployment more efficient. It allows developers to package their software with all necessary dependencies, making it easier to run in
any environment. Released in 2013, Docker has transformed the way applications are built, deployed, and managed.
\`
2. \`<text>
The theory of relativity, or simply relativity, encompasses two interrelated theories of Albert Einstein: special relativity and general
relativity. However, the word "relativity" is sometimes used in reference to Galilean invariance. The term "theory of relativity" was based
on the expression "relative theory" used by Max Planck in 1906. The theory of relativity usually encompasses two interrelated theories by
Albert Einstein: special relativity and general relativity. Special relativity applies to all physical phenomena in the absence of gravity.
General relativity explains the law of gravitation and its relation to other forces of nature. It applies to the cosmological and astrophysical
realm, including astronomy.
</text>
<query>
summarize
</query>
Response:
The theory of relativity, developed by Albert Einstein, encompasses two main theories: special relativity and general relativity. Special
relativity applies to all physical phenomena in the absence of gravity, while general relativity explains the law of gravitation and its
relation to other forces of nature. The theory of relativity is based on the concept of "relative theory," as introduced by Max Planck in
1906. It is a fundamental theory in physics that has revolutionized our understanding of the universe.
\`
</example>
Everything below is the actual data you will be working with. Good luck!
<query>
${question}
</query>
<text>
${doc.pageContent}
</text>
Make sure to answer the query in the summary.
`);
const document = new Document({
pageContent: res.content as string,
metadata: {
title: doc.metadata.title,
url: doc.metadata.url,
},
});
docs.push(document);
}),
);
return { query: question, docs: docs };
} else {
question = question.replace(/<think>.*?<\/think>/g, '');
const res = await searchSearxng(question, {
language: 'en',
engines: this.config.activeEngines,
});
const documents = res.results.map(
(result) =>
new Document({
pageContent:
result.content ||
(this.config.activeEngines.includes('youtube')
? result.title
: '') /* Todo: Implement transcript grabbing using Youtubei (source: https://www.npmjs.com/package/youtubei) */,
metadata: {
title: result.title,
url: result.url,
...(result.img_src && { img_src: result.img_src }),
},
}),
);
return { query: question, docs: documents };
}
}),
]);
}
private async createAnsweringChain(
llm: BaseChatModel,
fileIds: string[],
embeddings: Embeddings,
optimizationMode: 'speed' | 'balanced' | 'quality',
systemInstructions: string,
) {
return RunnableSequence.from([
RunnableMap.from({
systemInstructions: () => systemInstructions,
query: (input: BasicChainInput) => input.query,
chat_history: (input: BasicChainInput) => input.chat_history,
date: () => new Date().toISOString(),
context: RunnableLambda.from(async (input: BasicChainInput) => {
const processedHistory = formatChatHistoryAsString(
input.chat_history,
);
let docs: Document[] | null = null;
let query = input.query;
if (this.config.searchWeb) {
const searchRetrieverChain =
await this.createSearchRetrieverChain(llm);
const searchRetrieverResult = await searchRetrieverChain.invoke({
chat_history: processedHistory,
query,
});
query = searchRetrieverResult.query;
docs = searchRetrieverResult.docs;
}
const sortedDocs = await this.rerankDocs(
query,
docs ?? [],
fileIds,
embeddings,
optimizationMode,
);
return sortedDocs;
})
.withConfig({
runName: 'FinalSourceRetriever',
})
.pipe(this.processDocs),
}),
ChatPromptTemplate.fromMessages([
['system', this.config.responsePrompt],
new MessagesPlaceholder('chat_history'),
['user', '{query}'],
]),
llm,
this.strParser,
]).withConfig({
runName: 'FinalResponseGenerator',
});
}
private async rerankDocs(
query: string,
docs: Document[],
@ -432,36 +213,228 @@ class MetaSearchAgent implements MetaSearchAgentType {
}
private async handleStream(
stream: AsyncGenerator<StreamEvent, any, any>,
stream: AsyncIterable<[BaseMessage, Record<string, any>]>,
emitter: eventEmitter,
) {
for await (const event of stream) {
if (
event.event === 'on_chain_end' &&
event.name === 'FinalSourceRetriever'
) {
``;
for await (const [message, _metadata] of stream) {
if (isAIMessage(message) && message.tool_calls?.length) {
} else if (isAIMessage(message) && message.content) {
emitter.emit(
'data',
JSON.stringify({ type: 'sources', data: event.data.output }),
JSON.stringify({ type: 'response', data: message.content }),
);
}
if (
event.event === 'on_chain_stream' &&
event.name === 'FinalResponseGenerator'
) {
emitter.emit(
'data',
JSON.stringify({ type: 'response', data: event.data.chunk }),
);
}
if (
event.event === 'on_chain_end' &&
event.name === 'FinalResponseGenerator'
) {
emitter.emit('end');
}
}
emitter.emit('end');
}
getTools({
llm,
emitter,
}: {
llm: BaseLanguageModel;
emitter: EventEmitter;
}): DynamicStructuredTool[] {
const searchToolInputSchema = z.object({
query: z.string().describe('The query to search the web for.'),
links: z
.array(z.string().describe('The link to get data from'))
.describe(
'A list of links (if shared by user) to generate an answer from.',
),
});
const searchTool = tool(
async (input: any) => {
if (input.links.length > 0) {
let docs: Document[] = [];
const linkDocs = await getDocumentsFromLinks({ links: input.links });
const docGroups: Document[] = [];
linkDocs.map((doc) => {
const URLDocExists = docGroups.find(
(d) =>
d.metadata.url === doc.metadata.url &&
d.metadata.totalDocs < 10,
);
if (!URLDocExists) {
docGroups.push({
...doc,
metadata: {
...doc.metadata,
totalDocs: 1,
},
});
}
const docIndex = docGroups.findIndex(
(d) =>
d.metadata.url === doc.metadata.url &&
d.metadata.totalDocs < 10,
);
if (docIndex !== -1) {
docGroups[docIndex].pageContent =
docGroups[docIndex].pageContent + `\n\n` + doc.pageContent;
docGroups[docIndex].metadata.totalDocs += 1;
}
});
const URLSourcePrompt = `
You are a web search question answerer, tasked with finding relevant information from web documents to answer questions. Your job is to extract and summarize the most relevant parts of a document that can help answer the user's query.
- **Find relevant sections**: Identify parts of the document that directly relate to the question
- **Extract key information**: Pull out specific facts, data, or explanations that answer the query
- **Summarize concisely**: Provide a focused summary of the relevant information found
- **Stay on topic**: Only include information that helps answer the specific question asked
The document text will be shared inside the \`text\` XML tag, and the query inside the \`query\` XML tag.
Extract and summarize the relevant information from the document that answers the query.
`;
const URLSourceChatPrompt = ChatPromptTemplate.fromMessages([
['system', URLSourcePrompt],
[
'human',
`
<text>
Docker is a set of platform-as-a-service products that use OS-level virtualization to deliver software in packages called containers.
It was first released in 2013 and is developed by Docker, Inc. Docker is designed to make it easier to create, deploy, and run applications
by using containers.
</text>
<query>
What is Docker and how does it work?
</query>
`,
],
[
'assistant',
`
Docker is a revolutionary platform-as-a-service product developed by Docker, Inc., that uses container technology to make application
deployment more efficient. It allows developers to package their software with all necessary dependencies, making it easier to run in
any environment. Released in 2013, Docker has transformed the way applications are built, deployed, and managed.
`,
],
[
'human',
`
<text>
Rising global temperatures are causing significant changes to Earth's climate system. Ice sheets in Greenland and Antarctica are melting at accelerated rates, contributing to sea level rise. Ocean temperatures are increasing, leading to coral bleaching events. Weather patterns are becoming more extreme, with increased frequency of hurricanes, droughts, and flooding. The Arctic is warming twice as fast as the global average, a phenomenon known as Arctic amplification.
</text>
<query>
What are the main effects of climate change on the environment?
</query>
`,
],
[
'assistant',
`
Climate change is causing accelerated melting of ice sheets in Greenland and Antarctica, leading to sea level rise. Rising ocean temperatures are causing coral bleaching, while weather patterns are becoming more extreme with increased hurricanes, droughts, and flooding. The Arctic is experiencing particularly rapid warming at twice the global average rate.
`,
],
[
'human',
`
<text>
Pasta is a staple food of Italian cuisine made from wheat flour and water. Common shapes include spaghetti, penne, and fusilli. It can be served with various sauces like marinara, alfredo, or pesto. Pasta cooking involves boiling water with salt and cooking until al dente. Different regions of Italy have their own traditional pasta dishes and preparation methods.
</text>
<query>
How do solar panels generate electricity?
</query>
`,
],
[
'assistant',
`
The provided document about pasta and Italian cuisine does not contain information about how solar panels generate electricity. No relevant information found to answer the query.
`,
],
[
'human',
`
<text>
{content}
</text>
<query>
{query}
</query>
`,
],
]);
await Promise.all(
docGroups.map(async (doc) => {
const formattedPrompt = await URLSourceChatPrompt.formatMessages({
query: input.query,
content: doc.pageContent,
});
const llmResponse = await llm.invoke(formattedPrompt);
const document = new Document({
pageContent: llmResponse.content as string,
metadata: {
title: doc.metadata.title,
url: doc.metadata.url,
},
});
docs.push(document);
}),
);
emitter.emit('data', JSON.stringify({ type: 'sources', data: docs }));
return this.processDocs(docs);
} else {
const res = await searchSearxng(input.query, {
language: 'en',
engines: this.config.activeEngines,
});
const documents = res.results.map(
(result) =>
new Document({
pageContent:
result.content ||
(this.config.activeEngines.includes('youtube')
? result.title
: '') /* Todo: Implement transcript grabbing using Youtubei (source: https://www.npmjs.com/package/youtubei) */,
metadata: {
title: result.title,
url: result.url,
...(result.img_src && { img_src: result.img_src }),
},
}),
);
emitter.emit(
'data',
JSON.stringify({ type: 'sources', data: documents }),
);
return this.processDocs(documents);
}
},
{
name: 'search_web',
schema: searchToolInputSchema,
description: 'This tool allows you to search the web for information.',
},
);
return [searchTool];
}
async searchAndAnswer(
@ -475,21 +448,94 @@ class MetaSearchAgent implements MetaSearchAgentType {
) {
const emitter = new eventEmitter();
const answeringChain = await this.createAnsweringChain(
const tools = this.getTools({
emitter,
llm,
fileIds,
embeddings,
optimizationMode,
systemInstructions,
);
});
const stream = answeringChain.streamEvents(
const shouldContinue = (state: typeof MessagesAnnotation.State) => {
const lastMessage = state.messages[
state.messages.length - 1
] as AIMessage;
if (lastMessage.tool_calls && lastMessage.tool_calls.length) {
return 'tools';
}
return '__end__';
};
const callTools = async (
state: typeof MessagesAnnotation.State,
): Promise<Partial<typeof MessagesAnnotation.State>> => {
const lastMessage = state.messages[
state.messages.length - 1
] as AIMessage;
const toolResults: BaseMessage[] = [];
if (lastMessage.tool_calls && lastMessage.tool_calls.length) {
await Promise.all(
lastMessage.tool_calls.map(async (t) => {
const toolToCall = tools.find((i) => i.name === t.name);
const result = await toolToCall?.invoke(t.args)!;
toolResults.push(
new ToolMessage({
content: result,
tool_call_id: t.id!,
}),
);
}),
);
}
return {
messages: [...toolResults],
};
};
const boundModel = llm.bindTools?.(tools)!;
const callModel = async (
state: typeof MessagesAnnotation.State,
): Promise<Partial<typeof MessagesAnnotation.State>> => {
const { messages } = state;
const res = await boundModel?.invoke(messages);
return {
messages: [res!],
};
};
const workflow = new StateGraph(MessagesAnnotation)
.addNode('agent', callModel)
.addNode(
'tools',
RunnableLambda.from(callTools).withConfig({
tags: ['nostream'],
}),
)
.addEdge('__start__', 'agent')
.addEdge('tools', 'agent')
.addConditionalEdges('agent', shouldContinue);
const app = workflow.compile();
const filledPrompt = await PromptTemplate.fromTemplate(
this.config.responsePrompt,
).format({
systemInstructions: systemInstructions,
date: Date.now(),
});
const stream = await app.stream(
{
chat_history: history,
query: message,
messages: [['system', filledPrompt], ...history, ['human', message]],
},
{
version: 'v1',
streamMode: 'messages',
},
);