From 4791cd23af08da70895204a3a7fbaf889d9af2d5 Mon Sep 17 00:00:00 2001 From: "A.J. Shulman" Date: Sat, 7 Sep 2024 12:43:05 -0400 Subject: completely restructured, added comments, and significantly reduced the length of the prompt (~72% shorter and cheaper) --- src/client/views/nodes/chatbot/tools/BaseTool.ts | 24 ++++ .../views/nodes/chatbot/tools/CalculateTool.ts | 26 ++++ .../views/nodes/chatbot/tools/CreateCSVTool.ts | 51 ++++++++ .../nodes/chatbot/tools/CreateCollectionTool.ts | 36 ++++++ .../views/nodes/chatbot/tools/DataAnalysisTool.ts | 59 +++++++++ .../views/nodes/chatbot/tools/GetDocsTool.ts | 29 +++++ src/client/views/nodes/chatbot/tools/NoTool.ts | 18 +++ src/client/views/nodes/chatbot/tools/RAGTool.ts | 138 +++++++++++++++++++++ src/client/views/nodes/chatbot/tools/SearchTool.ts | 54 ++++++++ .../nodes/chatbot/tools/WebsiteInfoScraperTool.ts | 83 +++++++++++++ .../views/nodes/chatbot/tools/WikipediaTool.ts | 37 ++++++ 11 files changed, 555 insertions(+) create mode 100644 src/client/views/nodes/chatbot/tools/BaseTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/CalculateTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/CreateCSVTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/GetDocsTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/NoTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/RAGTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/SearchTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts create mode 100644 src/client/views/nodes/chatbot/tools/WikipediaTool.ts (limited to 'src/client/views/nodes/chatbot/tools') diff --git a/src/client/views/nodes/chatbot/tools/BaseTool.ts b/src/client/views/nodes/chatbot/tools/BaseTool.ts new file mode 100644 index 000000000..b57f1c8e4 --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/BaseTool.ts @@ -0,0 +1,24 @@ +import { Tool } from '../types/types'; + +export abstract class BaseTool = Record> implements Tool { + constructor( + public name: string, + public description: string, + public parameters: Record, + public citationRules: string, + public briefSummary: string + ) {} + + abstract execute(args: T): Promise; + + getActionRule(): Record { + return { + [this.name]: { + name: this.name, + citationRules: this.citationRules, + description: this.description, + parameters: this.parameters, + }, + }; + } +} diff --git a/src/client/views/nodes/chatbot/tools/CalculateTool.ts b/src/client/views/nodes/chatbot/tools/CalculateTool.ts new file mode 100644 index 000000000..74b7ca27b --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/CalculateTool.ts @@ -0,0 +1,26 @@ +import { BaseTool } from './BaseTool'; + +export class CalculateTool extends BaseTool<{ expression: string }> { + constructor() { + super( + 'calculate', + 'Perform a calculation', + { + expression: { + type: 'string', + description: 'The mathematical expression to evaluate', + required: 'true', + max_inputs: '1', + }, + }, + 'Provide a mathematical expression to calculate that would work with JavaScript eval().', + 'Runs a calculation and returns the number - uses JavaScript so be sure to use floating point syntax if necessary' + ); + } + + async execute(args: { expression: string }): Promise { + // Note: Using eval() can be dangerous. Consider using a safer alternative. + const result = eval(args.expression); + return [{ type: 'text', text: result.toString() }]; + } +} diff --git a/src/client/views/nodes/chatbot/tools/CreateCSVTool.ts b/src/client/views/nodes/chatbot/tools/CreateCSVTool.ts new file mode 100644 index 000000000..55015846b --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/CreateCSVTool.ts @@ -0,0 +1,51 @@ +import { BaseTool } from './BaseTool'; +import { Networking } from '../../../../Network'; + +export class CreateCSVTool extends BaseTool<{ csvData: string; filename: string }> { + private _handleCSVResult: (url: string, filename: string, id: string, data: string) => void; + + constructor(handleCSVResult: (url: string, title: string, id: string, data: string) => void) { + super( + 'createCSV', + 'Creates a CSV file from raw CSV data and saves it to the server', + { + type: 'object', + properties: { + csvData: { + type: 'string', + description: 'A string of comma-separated values representing the CSV data.', + }, + filename: { + type: 'string', + description: 'The base name of the CSV file to be created. Should end in ".csv".', + }, + }, + required: ['csvData', 'filename'], + }, + 'Provide a CSV string and a filename to create a CSV file.', + 'Creates a CSV file from the provided CSV string and saves it to the server with a unique identifier, returning the file URL and UUID.' + ); + this._handleCSVResult = handleCSVResult; + } + + async execute(args: { csvData: string; filename: string }): Promise { + try { + console.log('Creating CSV file:', args.filename, ' with data:', args.csvData); + // Post the raw CSV data to the createCSV endpoint on the server + const { fileUrl, id } = await Networking.PostToServer('/createCSV', { filename: args.filename, data: args.csvData }); + + // Handle the result by invoking the callback + this._handleCSVResult(fileUrl, args.filename, id, args.csvData); + + return [ + { + type: 'text', + text: `File successfully created: ${fileUrl}. \nNow a CSV file with this data and the name ${args.filename} is available as a user doc.`, + }, + ]; + } catch (error) { + console.error('Error creating CSV file:', error); + throw new Error('Failed to create CSV file.'); + } + } +} diff --git a/src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts b/src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts new file mode 100644 index 000000000..573428179 --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts @@ -0,0 +1,36 @@ +import { DocCast } from '../../../../../fields/Types'; +import { DocServer } from '../../../../DocServer'; +import { Docs } from '../../../../documents/Documents'; +import { DocumentView } from '../../DocumentView'; +import { OpenWhere } from '../../OpenWhere'; +import { BaseTool } from './BaseTool'; + +export class GetDocsContentTool extends BaseTool<{ title: string; document_ids: string[] }> { + private _docView: DocumentView; + constructor(docView: DocumentView) { + super( + 'retrieveDocs', + 'Retrieves the contents of all Documents that the user is interacting with in Dash ', + { + title: { + type: 'string', + description: 'the title of the collection that you will be making', + required: 'true', + max_inputs: '1', + }, + }, + 'Provide a mathematical expression to calculate that would work with JavaScript eval().', + 'Runs a calculation and returns the number - uses JavaScript so be sure to use floating point syntax if necessary' + ); + this._docView = docView; + } + + async execute(args: { title: string; document_ids: string[] }): Promise { + // Note: Using eval() can be dangerous. Consider using a safer alternative. + const docs = args.document_ids.map(doc_id => DocCast(DocServer.GetCachedRefField(doc_id))); + const collection = Docs.Create.FreeformDocument(docs, { title: args.title }); + this._docView._props.addDocTab(collection, OpenWhere.addRight); //in future, create popup prompting user where to add + return [{ type: 'text', text: 'Collection created in Dash called ' + args.title }]; + } +} +//export function create_collection(docView: DocumentView, document_ids: string[], title: string): string {} diff --git a/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts b/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts new file mode 100644 index 000000000..a12ee46e5 --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts @@ -0,0 +1,59 @@ +import { BaseTool } from './BaseTool'; + +export class DataAnalysisTool extends BaseTool<{ csv_file_name: string | string[] }> { + private csv_files_function: () => { filename: string; id: string; text: string }[]; + + constructor(csv_files: () => { filename: string; id: string; text: string }[]) { + super( + 'dataAnalysis', + 'Analyzes, and provides insights, from one or more CSV files', + { + csv_file_name: { + type: 'string', + description: 'Name(s) of the CSV file(s) to analyze', + required: 'true', + max_inputs: '3', + }, + }, + 'Provide the name(s) of up to 3 CSV files to analyze based on the user query and whichever available CSV files may be relevant.', + 'Provides the full CSV file text for your analysis based on the user query and the available CSV file(s). ' + ); + this.csv_files_function = csv_files; + } + + getFileContent(filename: string): string | undefined { + const files = this.csv_files_function(); + const file = files.find(f => f.filename === filename); + return file?.text; + } + + getFileID(filename: string): string | undefined { + const files = this.csv_files_function(); + const file = files.find(f => f.filename === filename); + return file?.id; + } + + async execute(args: { csv_file_name: string | string[] }): Promise { + const filenames = Array.isArray(args.csv_file_name) ? args.csv_file_name : [args.csv_file_name]; + const results = []; + + for (const filename of filenames) { + const fileContent = this.getFileContent(filename); + const fileID = this.getFileID(filename); + + if (fileContent && fileID) { + results.push({ + type: 'text', + text: `${fileContent}`, + }); + } else { + results.push({ + type: 'text', + text: `File not found: ${filename}`, + }); + } + } + + return results; + } +} diff --git a/src/client/views/nodes/chatbot/tools/GetDocsTool.ts b/src/client/views/nodes/chatbot/tools/GetDocsTool.ts new file mode 100644 index 000000000..f970ca8ee --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/GetDocsTool.ts @@ -0,0 +1,29 @@ +import { DocCast } from '../../../../../fields/Types'; +import { DocServer } from '../../../../DocServer'; +import { Docs } from '../../../../documents/Documents'; +import { DocumentView } from '../../DocumentView'; +import { OpenWhere } from '../../OpenWhere'; +import { BaseTool } from './BaseTool'; + +export class GetDocsTool extends BaseTool<{ title: string; document_ids: string[] }> { + private _docView: DocumentView; + constructor(docView: DocumentView) { + super( + 'retrieveDocs', + 'Retrieves the contents of all Documents that the user is interacting with in Dash', + {}, + 'No need to provide anything. Just run the tool and it will retrieve the contents of all Documents that the user is interacting with in Dash.', + 'Returns the the documents in Dash in JSON form. This will include the title of the document, the location in the FreeFormDocument, and the content of the document, any applicable data fields, the layout of the document, etc.' + ); + this._docView = docView; + } + + async execute(args: { title: string; document_ids: string[] }): Promise { + // Note: Using eval() can be dangerous. Consider using a safer alternative. + const docs = args.document_ids.map(doc_id => DocCast(DocServer.GetCachedRefField(doc_id))); + const collection = Docs.Create.FreeformDocument(docs, { title: args.title }); + this._docView._props.addDocTab(collection, OpenWhere.addRight); //in future, create popup prompting user where to add + return [{ type: 'text', text: 'Collection created in Dash called ' + args.title }]; + } +} +//export function create_collection(docView: DocumentView, document_ids: string[], title: string): string {} diff --git a/src/client/views/nodes/chatbot/tools/NoTool.ts b/src/client/views/nodes/chatbot/tools/NoTool.ts new file mode 100644 index 000000000..1f0830a77 --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/NoTool.ts @@ -0,0 +1,18 @@ +// tools/NoTool.ts +import { BaseTool } from './BaseTool'; + +export class NoTool extends BaseTool<{}> { + constructor() { + super( + 'no_tool', + 'Use this when no external tool or action is required to answer the question.', + {}, + 'When using the "no_tool" action, simply provide an empty element. The observation will always be "No tool used. Proceed with answering the question."', + 'Use when no external tool or action is required to answer the question.' + ); + } + + async execute(args: {}): Promise { + return [{ type: 'text', text: 'No tool used. Proceed with answering the question.' }]; + } +} diff --git a/src/client/views/nodes/chatbot/tools/RAGTool.ts b/src/client/views/nodes/chatbot/tools/RAGTool.ts new file mode 100644 index 000000000..c24306dcd --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/RAGTool.ts @@ -0,0 +1,138 @@ +import { BaseTool } from './BaseTool'; +import { Vectorstore } from '../vectorstore/Vectorstore'; +import { RAGChunk } from '../types/types'; +import * as fs from 'fs'; +import { Networking } from '../../../../Network'; +import { file } from 'jszip'; +import { ChatCompletion, ChatCompletionContentPart, ChatCompletionMessageParam } from 'openai/resources'; + +export class RAGTool extends BaseTool { + constructor(private vectorstore: Vectorstore) { + super( + 'rag', + 'Perform a RAG search on user documents', + { + hypothetical_document_chunk: { + type: 'string', + description: + "Detailed version of the prompt that is effectively a hypothetical document chunk that would be ideal to embed and compare to the vectors of real document chunks to fetch the most relevant document chunks to answer the user's query", + required: 'true', + }, + }, + ` + Your task is to provide a comprehensive response to the user's prompt based on the given chunks and chat history. Follow these structural guidelines meticulously: + + 1. Overall Structure: + + [Main content with grounded_text tags interspersed with normal plain text (information that is not derived from chunks' information)] + + [Individual citation tags] + + + [Three question tags] + + + + 2. Grounded Text Tag Structure: + - Basic format: + + [Your generated text based on information from a subset of a chunk (a citation's direct text)] + + + 3. Citation Tag Structure: + + [For text: relevant subset of original chunk] + [For image/table: leave empty] + + + 4. Detailed Grounded Text Guidelines: + a. Wrap all information derived from chunks in grounded_text tags. + b. DO NOT PUT ANYTHING THAT IS NOT DIRECTLY DERIVED FROM INFORMATION FROM CHUNKS (EITHER IMAGE, TABLE, OR TEXT) IN GROUNDED_TEXT TAGS. + c. Use a single grounded_text tag for suquential and closely related information that references the same citation. If other citations' information are used sequentially, create new grounded_text tags. + d. Ensure every grounded_text tag has up to a few corresponding citations (should not be more than 3 and only 1 is fine). Multiple citation indices should be separated by commas. + e. Grounded text can be as short as a few words or as long as several sentences. + f. Avoid overlapping or nesting grounded_text tags; instead, use sequential tags. + + 5. Detailed Citation Guidelines: + a. Create a unique citation for each distinct piece of information from the chunks that is used to support grounded_text. + b. ALL TEXT CITATIONS must have direct text in its element content (e.g. DIRECT TEXT HERE) that is a relevant SUBSET of the original text chunk that is being cited specifically. + c. DO NOT paraphrase or summarize the text; use the original text as much as possible. + d. DO NOT USE THE FULL TEXT CHUNK as the citation content; only use the relevant subset of the text that the grounded_text is base. AS SHORT AS POSSIBLE WHILE PROVIDING INFORMATION (ONE TO TWO SENTENCES USUALLY)! + e. Ensure each citation has a unique index number. + f. Specify the correct type: "text", "image", or "table". + g. For text chunks, the content of the citation should ALWAYS have the relevant subset of the original text that the grounded_text is based on. + h. For image/table chunks, leave the citation content empty. + i. One citation can be used for multiple grounded_text tags if they are based on the same chunk information. + j. !!!DO NOT OVERCITE - only include citations for information that is directly relevant to the grounded_text. + + 6. Structural Integrity Checks: + a. Ensure all opening tags have corresponding closing tags. + b. Verify that all grounded_text tags have valid citation_index attributes (they should be equal to the associated citation(s) index field—not their chunk_id field). + c. Check that all cited indices in grounded_text tags have corresponding citations. + + Example of grounded_text usage: + + + + Artificial Intelligence (AI) is revolutionizing various sectors, with healthcare experiencing significant transformations in areas such as diagnosis and treatment planning. + + + In the field of medical diagnosis, AI has shown remarkable capabilities, particularly in radiology. For instance, AI systems have drastically improved mammogram analysis, achieving 99% accuracy at a rate 30 times faster than human radiologists. + + + This advancement not only enhances the efficiency of healthcare systems but also significantly reduces the occurrence of false positives, leading to fewer unnecessary biopsies and reduced patient stress. + + + + Beyond diagnosis, AI is playing a crucial role in drug discovery and development. By analyzing vast amounts of genetic and molecular data, AI algorithms can identify potential drug candidates much faster than traditional methods. + + + This could potentially reduce the time and cost of bringing new medications to market, especially for rare diseases that have historically received less attention due to limited market potential. + + + [... rest of the content ...] + + + Artificial Intelligence is revolutionizing various industries, with healthcare being one of the most profoundly affected sectors. + AI has shown particular promise in the field of radiology, enhancing the accuracy and speed of image analysis. + According to recent studies, AI systems have achieved 99% accuracy in mammogram analysis, performing the task 30 times faster than human radiologists. + The improvement in mammogram accuracy has led to a significant reduction in false positives, decreasing the need for unnecessary biopsies and reducing patient anxiety. + AI is accelerating the drug discovery process by analyzing complex molecular and genetic data to identify potential drug candidates. + The use of AI in drug discovery could significantly reduce the time and cost associated with bringing new medications to market, particularly for rare diseases. + + + + How might AI-driven personalized medicine impact the cost and accessibility of healthcare in the future? + What measures can be taken to ensure that AI systems in healthcare are free from biases and equally effective for diverse populations? + How could the role of healthcare professionals evolve as AI becomes more integrated into medical practices? + + + `, + + `Performs a RAG (Retrieval-Augmented Generation) search on user documents and returns a + set of document chunks (either images or text) that can be used to provide a grounded response based on + user documents` + ); + } + + async execute(args: { hypothetical_document_chunk: string }): Promise { + const relevantChunks = await this.vectorstore.retrieve(args.hypothetical_document_chunk); + const formatted_chunks = await this.getFormattedChunks(relevantChunks); + return formatted_chunks; + } + + async getFormattedChunks(relevantChunks: RAGChunk[]): Promise { + try { + const { formattedChunks } = await Networking.PostToServer('/formatChunks', { relevantChunks }); + + if (!formattedChunks) { + throw new Error('Failed to format chunks'); + } + + return formattedChunks; + } catch (error) { + console.error('Error formatting chunks:', error); + throw error; + } + } +} diff --git a/src/client/views/nodes/chatbot/tools/SearchTool.ts b/src/client/views/nodes/chatbot/tools/SearchTool.ts new file mode 100644 index 000000000..b926cbadc --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/SearchTool.ts @@ -0,0 +1,54 @@ +import { max } from 'lodash'; +import { Networking } from '../../../../Network'; +import { BaseTool } from './BaseTool'; +import { v4 as uuidv4 } from 'uuid'; + +export class SearchTool extends BaseTool<{ query: string | string[] }> { + private _addLinkedUrlDoc: (url: string, id: string) => void; + private _max_results: number; + constructor(addLinkedUrlDoc: (url: string, id: string) => void, max_results: number = 5) { + super( + 'searchTool', + 'Search the web to find a wide range of websites related to a query or multiple queries', + { + query: { + type: 'string', + description: 'The search query or queries to use for finding websites', + required: 'true', + max_inputs: '3', + }, + }, + 'Provide up to 3 search queries to find a broad range of websites. This tool is intended to help you identify relevant websites, but not to be used for providing the final answer. Use this information to determine which specific website to investigate further.', + 'Returns a list of websites and their overviews based on the search queries, helping to identify which websites might contain relevant information.' + ); + this._addLinkedUrlDoc = addLinkedUrlDoc; + this._max_results = max_results; + } + + async execute(args: { query: string | string[] }): Promise { + const queries = Array.isArray(args.query) ? args.query : [args.query]; + const allResults = []; + + for (const query of queries) { + try { + const { results } = await Networking.PostToServer('/getWebSearchResults', { query, max_results: this._max_results }); + const data: { type: string; text: string }[] = results.map((result: { url: string; snippet: string }) => { + const id = uuidv4(); + return { + type: 'text', + text: ` + ${result.url} + ${result.snippet} + `, + }; + }); + allResults.push(...data); + } catch (error) { + console.log(error); + allResults.push({ type: 'text', text: `An error occurred while performing the web search for query: ${query}` }); + } + } + + return allResults; + } +} diff --git a/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts b/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts new file mode 100644 index 000000000..2118218f6 --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts @@ -0,0 +1,83 @@ +import { Networking } from '../../../../Network'; +import { BaseTool } from './BaseTool'; +import { v4 as uuidv4 } from 'uuid'; + +export class WebsiteInfoScraperTool extends BaseTool<{ url: string | string[] }> { + private _addLinkedUrlDoc: (url: string, id: string) => void; + + constructor(addLinkedUrlDoc: (url: string, id: string) => void) { + super( + 'websiteInfoScraper', + 'Scrape detailed information from specific websites relevant to the user query', + { + url: { + type: 'string', + description: 'The URL(s) of the website(s) to scrape', + required: true, + max_inputs: 3, + }, + }, + ` + Your task is to provide a comprehensive response to the user's prompt using the content scraped from relevant websites. Ensure you follow these guidelines for structuring your response: + + 1. Grounded Text Tag Structure: + - Wrap all text derived from the scraped website(s) in tags. + - **Do not include non-sourced information** in tags. + - Use a single tag for content derived from a single website. If citing multiple websites, create new tags for each. + - Ensure each tag has a citation index corresponding to the scraped URL. + + 2. Citation Tag Structure: + - Create a tag for each distinct piece of information used from the website(s). + - Each tag must reference a URL chunk using the chunk_id attribute. + - For URL-based citations, leave the citation content empty, but reference the chunk_id and type as 'url'. + + 3. Structural Integrity Checks: + - Ensure all opening and closing tags are matched properly. + - Verify that all citation_index attributes in tags correspond to valid citations. + - Do not over-cite—cite only the most relevant parts of the websites. + + Example Usage: + + + + Based on data from the World Bank, economic growth has stabilized in recent years, following a surge in investments. + + + According to information retrieved from the International Monetary Fund, the inflation rate has been gradually decreasing since 2020. + + + + + + + + + What are the long-term economic impacts of increased investments on GDP? + How might inflation trends affect future monetary policy? + Are there additional factors that could influence economic growth beyond investments and inflation? + + + `, + 'Returns the text content of the webpages for further analysis and grounding.' + ); + this._addLinkedUrlDoc = addLinkedUrlDoc; + } + + async execute(args: { url: string | string[] }): Promise { + const urls = Array.isArray(args.url) ? args.url : [args.url]; + const results = []; + + for (const url of urls) { + try { + const { website_plain_text } = await Networking.PostToServer('/scrapeWebsite', { url }); + const id = uuidv4(); + this._addLinkedUrlDoc(url, id); + results.push({ type: 'text', text: `\n${website_plain_text}\n\n` }); + } catch (error) { + results.push({ type: 'text', text: `An error occurred while scraping the website: ${url}` }); + } + } + + return results; + } +} diff --git a/src/client/views/nodes/chatbot/tools/WikipediaTool.ts b/src/client/views/nodes/chatbot/tools/WikipediaTool.ts new file mode 100644 index 000000000..143d91d80 --- /dev/null +++ b/src/client/views/nodes/chatbot/tools/WikipediaTool.ts @@ -0,0 +1,37 @@ +import { title } from 'process'; +import { Networking } from '../../../../Network'; +import { BaseTool } from './BaseTool'; +import axios from 'axios'; +import { v4 as uuidv4 } from 'uuid'; + +export class WikipediaTool extends BaseTool<{ title: string }> { + private _addLinkedUrlDoc: (url: string, id: string) => void; + constructor(addLinkedUrlDoc: (url: string, id: string) => void) { + super( + 'wikipedia', + 'Search Wikipedia and return a summary', + { + title: { + type: 'string', + description: 'The title of the Wikipedia article to search', + required: true, + }, + }, + 'Provide simply the title you want to search on Wikipedia and nothing more. If re-using this tool, try a different title for different information.', + 'Returns a summary from searching an article title on Wikipedia' + ); + this._addLinkedUrlDoc = addLinkedUrlDoc; + } + + async execute(args: { title: string }): Promise { + try { + const { text } = await Networking.PostToServer('/getWikipediaSummary', { title: args.title }); + const id = uuidv4(); + const url = `https://en.wikipedia.org/wiki/${args.title.replace(/ /g, '_')}`; + this._addLinkedUrlDoc(url, id); + return [{ type: 'text', text: ` ${text} ` }]; + } catch (error) { + return [{ type: 'text', text: 'An error occurred while fetching the article.' }]; + } + } +} -- cgit v1.2.3-70-g09d2 From ba0520baaa1f84d9fb08d3b2880c68302d28350a Mon Sep 17 00:00:00 2001 From: "A.J. Shulman" Date: Sat, 7 Sep 2024 13:07:33 -0400 Subject: added clarifying structural info to ReAct prompt (still ~69% shorter than previous prompt) and shortened the RAG prompt --- .../views/nodes/chatbot/agentsystem/prompts.ts | 28 +++++- .../nodes/chatbot/chatboxcomponents/ChatBox.tsx | 7 +- src/client/views/nodes/chatbot/tools/RAGTool.ts | 104 +++++---------------- 3 files changed, 52 insertions(+), 87 deletions(-) (limited to 'src/client/views/nodes/chatbot/tools') diff --git a/src/client/views/nodes/chatbot/agentsystem/prompts.ts b/src/client/views/nodes/chatbot/agentsystem/prompts.ts index 9daabc35f..7000d8634 100644 --- a/src/client/views/nodes/chatbot/agentsystem/prompts.ts +++ b/src/client/views/nodes/chatbot/agentsystem/prompts.ts @@ -26,12 +26,33 @@ export function getReactPrompt(tools: Tool[], summaries: () => string, chatHisto Ensure that **ALL answers follow the answer structure**: grounded text wrapped in tags with corresponding citations, normal text in tags, and three follow-up questions at the end. + + + + Always provide a thought before each action to explain why you are choosing the next step or tool. This helps clarify your reasoning for the action you will take. + + + + + + + + Always describe what the action will do in the tag. Be clear about how the tool will process the input and why it is appropriate for this stage. + + + + Provide the actual inputs for the action in the tag. Ensure that each input is specific to the tool being used. Inputs should match the expected parameters for the tool (e.g., a search term for the website scraper, document references for RAG). + + + + + - All information derived from tools or user documents must be wrapped in these tags with proper citation. - - Use this tag for text not derived from tools or user documents. + - Use this tag for text not derived from tools or user documents. It should only be for narrative-like text or extremely common knowledge information. - - Provide proper citations for each , referencing the tool or document chunk used. + - Provide proper citations for each , referencing the tool or document chunk used. ENSURE THAT THERE IS A CITATION WHOSE INDEX MATCHES FOR EVERY GROUNDED TEXT CITATION INDEX. - Provide exactly three user-perspective follow-up questions. - Summarize the actions and tools used in the conversation. @@ -41,7 +62,8 @@ export function getReactPrompt(tools: Tool[], summaries: () => string, chatHisto **Wrap ALL tool-based information** in tags and provide citations. Use separate tags for distinct information or when switching to a different tool or document. - Ensure that **EVERY** tag includes a citation index referencing the source of the information. + Ensure that **EVERY** tag includes a citation index aligned with a citation that you provide that references the source of the information. + There should be a one-to-one relationship between tags and citations. Over-citing is discouraged—only cite the information that is directly relevant to the user's query. diff --git a/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx b/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx index 6dc691798..28bfbeae3 100644 --- a/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx +++ b/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx @@ -51,8 +51,7 @@ export class ChatBox extends ViewBoxAnnotatableComponent() { private vectorstore_id: string; private vectorstore: Vectorstore; private agent: Agent; - private _oldWheel: HTMLDivElement | null = null; - private messagesRef: React.RefObject; + private messagesRef: React.RefObject; /** * Static method that returns the layout string for the field. @@ -246,12 +245,12 @@ export class ChatBox extends ViewBoxAnnotatableComponent() { * @param event The form submission event. */ @action - askGPT = async (event: React.FormEvent): Promise => { + askGPT = async (event: React.FormEvent): Promise => { event.preventDefault(); this.inputValue = ''; // Extract the user's message - const textInput = event.currentTarget.elements.namedItem('messageInput') as HTMLInputElement; + const textInput = (event.currentTarget as HTMLFormElement).elements.namedItem('messageInput') as HTMLInputElement; const trimmedText = textInput.value.trim(); if (trimmedText) { diff --git a/src/client/views/nodes/chatbot/tools/RAGTool.ts b/src/client/views/nodes/chatbot/tools/RAGTool.ts index c24306dcd..f4b7b42ea 100644 --- a/src/client/views/nodes/chatbot/tools/RAGTool.ts +++ b/src/client/views/nodes/chatbot/tools/RAGTool.ts @@ -14,114 +14,58 @@ export class RAGTool extends BaseTool { { hypothetical_document_chunk: { type: 'string', - description: - "Detailed version of the prompt that is effectively a hypothetical document chunk that would be ideal to embed and compare to the vectors of real document chunks to fetch the most relevant document chunks to answer the user's query", + description: "A detailed prompt representing an ideal chunk to embed and compare against document vectors to retrieve the most relevant content for answering the user's query.", required: 'true', }, }, ` - Your task is to provide a comprehensive response to the user's prompt based on the given chunks and chat history. Follow these structural guidelines meticulously: + When using the RAG tool, the structure must adhere to the format described in the ReAct prompt. Below are additional guidelines specifically for RAG-based responses: - 1. Overall Structure: - - [Main content with grounded_text tags interspersed with normal plain text (information that is not derived from chunks' information)] - - [Individual citation tags] - - - [Three question tags] - - - - 2. Grounded Text Tag Structure: - - Basic format: - - [Your generated text based on information from a subset of a chunk (a citation's direct text)] - + 1. **Grounded Text Guidelines**: + - Each tag must correspond to exactly one citation, ensuring a one-to-one relationship. + - Always cite a **subset** of the chunk, never the full text. The citation should be as short as possible while providing the relevant information (typically one to two sentences). + - Do not paraphrase the chunk text in the citation; use the original subset directly from the chunk. + - If multiple citations are needed for different sections of the response, create new tags for each. - 3. Citation Tag Structure: - - [For text: relevant subset of original chunk] - [For image/table: leave empty] - + 2. **Citation Guidelines**: + - The citation must include only the relevant excerpt from the chunk being referenced. + - Use unique citation indices and reference the chunk_id for the source of the information. + - For text chunks, the citation content must reflect the **exact subset** of the original chunk that is relevant to the grounded_text tag. - 4. Detailed Grounded Text Guidelines: - a. Wrap all information derived from chunks in grounded_text tags. - b. DO NOT PUT ANYTHING THAT IS NOT DIRECTLY DERIVED FROM INFORMATION FROM CHUNKS (EITHER IMAGE, TABLE, OR TEXT) IN GROUNDED_TEXT TAGS. - c. Use a single grounded_text tag for suquential and closely related information that references the same citation. If other citations' information are used sequentially, create new grounded_text tags. - d. Ensure every grounded_text tag has up to a few corresponding citations (should not be more than 3 and only 1 is fine). Multiple citation indices should be separated by commas. - e. Grounded text can be as short as a few words or as long as several sentences. - f. Avoid overlapping or nesting grounded_text tags; instead, use sequential tags. - - 5. Detailed Citation Guidelines: - a. Create a unique citation for each distinct piece of information from the chunks that is used to support grounded_text. - b. ALL TEXT CITATIONS must have direct text in its element content (e.g. DIRECT TEXT HERE) that is a relevant SUBSET of the original text chunk that is being cited specifically. - c. DO NOT paraphrase or summarize the text; use the original text as much as possible. - d. DO NOT USE THE FULL TEXT CHUNK as the citation content; only use the relevant subset of the text that the grounded_text is base. AS SHORT AS POSSIBLE WHILE PROVIDING INFORMATION (ONE TO TWO SENTENCES USUALLY)! - e. Ensure each citation has a unique index number. - f. Specify the correct type: "text", "image", or "table". - g. For text chunks, the content of the citation should ALWAYS have the relevant subset of the original text that the grounded_text is based on. - h. For image/table chunks, leave the citation content empty. - i. One citation can be used for multiple grounded_text tags if they are based on the same chunk information. - j. !!!DO NOT OVERCITE - only include citations for information that is directly relevant to the grounded_text. - - 6. Structural Integrity Checks: - a. Ensure all opening tags have corresponding closing tags. - b. Verify that all grounded_text tags have valid citation_index attributes (they should be equal to the associated citation(s) index field—not their chunk_id field). - c. Check that all cited indices in grounded_text tags have corresponding citations. - - Example of grounded_text usage: + **Example**: - - Artificial Intelligence (AI) is revolutionizing various sectors, with healthcare experiencing significant transformations in areas such as diagnosis and treatment planning. - - - In the field of medical diagnosis, AI has shown remarkable capabilities, particularly in radiology. For instance, AI systems have drastically improved mammogram analysis, achieving 99% accuracy at a rate 30 times faster than human radiologists. + + Artificial Intelligence is revolutionizing various sectors, with healthcare seeing transformations in diagnosis and treatment planning. - - This advancement not only enhances the efficiency of healthcare systems but also significantly reduces the occurrence of false positives, leading to fewer unnecessary biopsies and reduced patient stress. + + Based on recent data, AI has drastically improved mammogram analysis, achieving 99% accuracy at a rate 30 times faster than human radiologists. - - - Beyond diagnosis, AI is playing a crucial role in drug discovery and development. By analyzing vast amounts of genetic and molecular data, AI algorithms can identify potential drug candidates much faster than traditional methods. - - - This could potentially reduce the time and cost of bringing new medications to market, especially for rare diseases that have historically received less attention due to limited market potential. - - - [... rest of the content ...] - Artificial Intelligence is revolutionizing various industries, with healthcare being one of the most profoundly affected sectors. - AI has shown particular promise in the field of radiology, enhancing the accuracy and speed of image analysis. - According to recent studies, AI systems have achieved 99% accuracy in mammogram analysis, performing the task 30 times faster than human radiologists. - The improvement in mammogram accuracy has led to a significant reduction in false positives, decreasing the need for unnecessary biopsies and reducing patient anxiety. - AI is accelerating the drug discovery process by analyzing complex molecular and genetic data to identify potential drug candidates. - The use of AI in drug discovery could significantly reduce the time and cost associated with bringing new medications to market, particularly for rare diseases. + Artificial Intelligence is revolutionizing various industries, especially in healthcare. + - How might AI-driven personalized medicine impact the cost and accessibility of healthcare in the future? - What measures can be taken to ensure that AI systems in healthcare are free from biases and equally effective for diverse populations? - How could the role of healthcare professionals evolve as AI becomes more integrated into medical practices? + How can AI enhance patient outcomes in fields outside radiology? + What are the challenges in implementing AI systems across different hospitals? + How might AI-driven advancements impact healthcare costs? `, - `Performs a RAG (Retrieval-Augmented Generation) search on user documents and returns a - set of document chunks (either images or text) that can be used to provide a grounded response based on - user documents` + `Performs a RAG (Retrieval-Augmented Generation) search on user documents and returns a set of document chunks (text or images) to provide a grounded response based on user documents.` ); } - async execute(args: { hypothetical_document_chunk: string }): Promise { + async execute(args: { hypothetical_document_chunk: string }): Promise { const relevantChunks = await this.vectorstore.retrieve(args.hypothetical_document_chunk); const formatted_chunks = await this.getFormattedChunks(relevantChunks); return formatted_chunks; } - async getFormattedChunks(relevantChunks: RAGChunk[]): Promise { + async getFormattedChunks(relevantChunks: RAGChunk[]): Promise { try { const { formattedChunks } = await Networking.PostToServer('/formatChunks', { relevantChunks }); -- cgit v1.2.3-70-g09d2 From 139a3cb0b3b081c270187e9b4ca281d04ca923bf Mon Sep 17 00:00:00 2001 From: bobzel Date: Mon, 30 Sep 2024 12:19:22 -0400 Subject: upate AJ assistant from master and fix some lint errors --- eslint.config.mjs | 2 +- src/client/documents/Documents.ts | 9 -- src/client/util/CurrentUserUtils.ts | 8 +- src/client/util/Scripting.ts | 13 +- src/client/views/MainView.tsx | 7 +- .../collections/CollectionNoteTakingViewColumn.tsx | 5 +- src/client/views/nodes/PDFBox.tsx | 21 ++- .../views/nodes/RecordingBox/ProgressBar.tsx | 2 - .../views/nodes/chatbot/agentsystem/Agent.ts | 33 ++-- .../nodes/chatbot/chatboxcomponents/ChatBox.tsx | 61 +++---- .../nodes/chatbot/response_parsers/AnswerParser.ts | 9 +- .../response_parsers/StreamedAnswerParser.ts | 2 - src/client/views/nodes/chatbot/tools/BaseTool.ts | 8 +- .../views/nodes/chatbot/tools/CalculateTool.ts | 2 +- .../views/nodes/chatbot/tools/CreateCSVTool.ts | 2 +- .../nodes/chatbot/tools/CreateCollectionTool.ts | 2 +- .../views/nodes/chatbot/tools/DataAnalysisTool.ts | 2 +- .../views/nodes/chatbot/tools/GetDocsTool.ts | 2 +- src/client/views/nodes/chatbot/tools/NoTool.ts | 5 +- src/client/views/nodes/chatbot/tools/RAGTool.ts | 13 +- src/client/views/nodes/chatbot/tools/SearchTool.ts | 5 +- .../nodes/chatbot/tools/WebsiteInfoScraperTool.ts | 5 +- .../views/nodes/chatbot/tools/WikipediaTool.ts | 7 +- src/client/views/nodes/chatbot/types/types.ts | 25 ++- .../views/nodes/chatbot/vectorstore/Vectorstore.ts | 13 +- src/client/views/pdf/PDFViewer.tsx | 46 ++---- src/server/ApiManagers/AssistantManager.ts | 178 ++++++++++----------- 27 files changed, 215 insertions(+), 272 deletions(-) (limited to 'src/client/views/nodes/chatbot/tools') diff --git a/eslint.config.mjs b/eslint.config.mjs index aebdc20d0..f7063caa5 100644 --- a/eslint.config.mjs +++ b/eslint.config.mjs @@ -48,7 +48,7 @@ export default [ 'no-return-assign': 'error', 'no-await-in-loop': 'error', 'no-loop-func': 'error', - '@typescript-eslint/no-cond-assign': 'error', + 'no-cond-assign': 'error', 'no-use-before-define': 'error', '@typescript-eslint/no-explicit-any': 'error', 'no-restricted-globals': ['error', 'event'], diff --git a/src/client/documents/Documents.ts b/src/client/documents/Documents.ts index e0a9918f4..d77f76b81 100644 --- a/src/client/documents/Documents.ts +++ b/src/client/documents/Documents.ts @@ -1,5 +1,3 @@ -/* eslint-disable prefer-destructuring */ -/* eslint-disable default-param-last */ /* eslint-disable no-use-before-define */ import { reaction } from 'mobx'; import { basename } from 'path'; @@ -671,7 +669,6 @@ export namespace Docs { * only when creating a DockDocument from the current user's already existing * main document. */ - // eslint-disable-next-line default-param-last function InstanceFromProto(proto: Doc, data: FieldType | undefined, options: DocumentOptions, delegId?: string, fieldKey: string = 'data', protoId?: string, placeholderDocIn?: Doc, noView?: boolean) { const placeholderDoc = placeholderDocIn; const viewKeys = ['x', 'y', 'isSystem']; // keys that should be addded to the view document even though they don't begin with an "_" @@ -732,7 +729,6 @@ export namespace Docs { return dataDoc; } - // eslint-disable-next-line default-param-last export function ImageDocument(url: string | ImageField, options: DocumentOptions = {}, overwriteDoc?: Doc) { const imgField = url instanceof ImageField ? url : url ? new ImageField(url) : undefined; return InstanceFromProto(Prototypes.get(DocumentType.IMG), imgField, { title: basename(imgField?.url.href ?? '-no image-'), ...options }, undefined, undefined, undefined, overwriteDoc); @@ -751,7 +747,6 @@ export namespace Docs { * @param fieldKey the field that the compiled script is written into. * @returns the Scripting Doc */ - // eslint-disable-next-line default-param-last export function ScriptingDocument(script: Opt | null, options: DocumentOptions = {}, fieldKey?: string) { return InstanceFromProto(Prototypes.get(DocumentType.SCRIPTING), script || undefined, { ...options, layout: fieldKey ? `` /* ScriptingBox.LayoutString(fieldKey) */ : undefined }); } @@ -759,7 +754,6 @@ export namespace Docs { export function ChatDocument(options?: DocumentOptions) { return InstanceFromProto(Prototypes.get(DocumentType.CHAT), undefined, { ...(options || {}) }); } - // eslint-disable-next-line default-param-last export function VideoDocument(url: string, options: DocumentOptions = {}, overwriteDoc?: Doc) { return InstanceFromProto(Prototypes.get(DocumentType.VID), new VideoField(url), options, undefined, undefined, undefined, overwriteDoc); } @@ -779,7 +773,6 @@ export namespace Docs { return InstanceFromProto(Prototypes.get(DocumentType.DIAGRAM), undefined, options); } - // eslint-disable-next-line default-param-last export function AudioDocument(url: string, options: DocumentOptions = {}, overwriteDoc?: Doc) { return InstanceFromProto(Prototypes.get(DocumentType.AUDIO), new AudioField(url), options, undefined, undefined, undefined, overwriteDoc); } @@ -839,7 +832,6 @@ export namespace Docs { return InstanceFromProto(Prototypes.get(DocumentType.RTF), field, options, undefined, fieldKey); } - // eslint-disable-next-line default-param-last export function LinkDocument(source: Doc, target: Doc, options: DocumentOptions = {}, id?: string) { const linkDoc = InstanceFromProto( Prototypes.get(DocumentType.LINK), @@ -883,7 +875,6 @@ export namespace Docs { return ink; } - // eslint-disable-next-line default-param-last export function PdfDocument(url: string, options: DocumentOptions = {}, overwriteDoc?: Doc) { const width = options._width || undefined; const height = options._height || undefined; diff --git a/src/client/util/CurrentUserUtils.ts b/src/client/util/CurrentUserUtils.ts index 96b30c429..09adf70f5 100644 --- a/src/client/util/CurrentUserUtils.ts +++ b/src/client/util/CurrentUserUtils.ts @@ -2,7 +2,7 @@ import { reaction, runInAction } from "mobx"; import * as rp from 'request-promise'; import { ClientUtils, OmitKeys } from "../../ClientUtils"; -import { Doc, DocListCast, DocListCastAsync, FieldType, Opt, StrListCast } from "../../fields/Doc"; +import { Doc, DocListCast, DocListCastAsync, FieldType, Opt } from "../../fields/Doc"; import { DocData } from "../../fields/DocSymbols"; import { InkTool } from "../../fields/InkField"; import { List } from "../../fields/List"; @@ -366,11 +366,11 @@ pie title Minerals in my tap water {key: "Button", creator: Docs.Create.ButtonDocument, opts: { _width: 150, _height: 50, _xPadding: 10, _yPadding: 10, title_custom: true, waitForDoubleClickToClick: 'never'}, scripts: {onClick: FollowLinkScript()?.script.originalScript ?? ""}}, {key: "Script", creator: opts => Docs.Create.ScriptingDocument(null, opts), opts: { _width: 200, _height: 250, }}, {key: "DataViz", creator: opts => Docs.Create.DataVizDocument("/users/rz/Downloads/addresses.csv", opts), opts: { _width: 300, _height: 300 }}, - {key: "Chat", creator: Docs.Create.ChatDocument, opts: { _width: 500, _height: 500, }}, + {key: "Chat", creator: Docs.Create.ChatDocument, opts: { _width: 500, _height: 500, }}, {key: "Header", creator: headerTemplate, opts: { _width: 300, _height: 120, _header_pointerEvents: "all", _header_height: 50, _header_fontSize: 9,_layout_autoHeightMargins: 50, _layout_autoHeight: true, treeView_HideUnrendered: true}}, {key: "ViewSlide", creator: slideView, opts: { _width: 400, _height: 300, _xMargin: 3, _yMargin: 3,}}, {key: "Trail", creator: Docs.Create.PresDocument, opts: { _width: 400, _height: 30, _type_collection: CollectionViewType.Stacking, _layout_dontCenter:'xy', dropAction: dropActionType.embed, treeView_HideTitle: true, _layout_fitWidth:true, layout_boxShadow: "0 0" }}, - {key: "Tab", creator: opts => Docs.Create.FreeformDocument([], opts), opts: { _width: 500, _height: 800, _layout_fitWidth: true, _freeform_backgroundGrid: true, }}, + {key: "Tab", creator: opts => Docs.Create.FreeformDocument([], opts), opts: { _width: 500, _height: 800, _layout_fitWidth: true, _freeform_backgroundGrid: true, }}, {key: "Slide", creator: opts => Docs.Create.TreeDocument([], opts), opts: { _width: 300, _height: 200, _type_collection: CollectionViewType.Tree, treeView_HasOverlay: true, _text_fontSize: "20px", _layout_autoHeight: true, dropAction:dropActionType.move, treeView_Type: TreeViewType.outline, @@ -801,7 +801,7 @@ pie title Minerals in my tap water { title: "Num", icon:"", toolTip: "Frame # (click to toggle edit mode)",btnType: ButtonType.TextButton, expertMode: true, toolType:CollectionViewType.Freeform, funcs: {hidden: '!SelectedDocType(this.toolType, this.expertMode)', buttonText: 'selectedDocs()?.lastElement()?.currentFrame?.toString()'}, width: 20, scripts: { onClick: '{ return curKeyFrame(_readOnly_);}'}}, { title: "Fwd", icon: "chevron-right", toolTip: "Next Animation Frame", btnType: ButtonType.ClickButton, expertMode: true, toolType:CollectionViewType.Freeform, funcs: {hidden: '!SelectedDocType(this.toolType, this.expertMode)'}, width: 30, scripts: { onClick: 'nextKeyFrame(_readOnly_)'}}, - { title: "Filter", icon: "=", toolTip: "Filter cards by tags", subMenu: CurrentUserUtils.tagGroupTools(),ignoreClick:true, toolType:DocumentType.COL, funcs: {hidden: '!SelectedDocType(this.toolType, this.expertMode)'}, btnType: ButtonType.MultiToggleButton, width: 30, backgroundColor: doc.userVariantColor}, + { title: "Filter", icon: "=", toolTip: "Filter cards by tags", subMenu: CurrentUserUtils.tagGroupTools(),ignoreClick:true, toolType:DocumentType.COL, funcs: {hidden: '!SelectedDocType(this.toolType, this.expertMode)'}, btnType: ButtonType.MultiToggleButton, width: 30, backgroundColor: doc.userVariantColor as string}, { title: "Text", icon: "Text", toolTip: "Text functions", subMenu: CurrentUserUtils.textTools(), expertMode: false, toolType:DocumentType.RTF, funcs: { linearView_IsOpen: `SelectedDocType(this.toolType, this.expertMode)`} }, // Always available { title: "Ink", icon: "Ink", toolTip: "Ink functions", subMenu: CurrentUserUtils.inkTools(), expertMode: false, toolType:DocumentType.INK, funcs: {hidden: `IsExploreMode()`, linearView_IsOpen: `SelectedDocType(this.toolType, this.expertMode)`}, scripts: { onClick: 'setInkToolDefaults()'} }, // Always available { title: "Doc", icon: "Doc", toolTip: "Freeform Doc tools", subMenu: CurrentUserUtils.freeTools(), expertMode: false, toolType:CollectionViewType.Freeform, funcs: {hidden: `!SelectedDocType(this.toolType, this.expertMode, true)`, linearView_IsOpen: `SelectedDocType(this.toolType, this.expertMode)`} }, // Always available diff --git a/src/client/util/Scripting.ts b/src/client/util/Scripting.ts index 3e7a2df02..c7b86815a 100644 --- a/src/client/util/Scripting.ts +++ b/src/client/util/Scripting.ts @@ -1,7 +1,7 @@ // export const ts = (window as any).ts; // import * as typescriptlib from '!!raw-loader!../../../node_modules/typescript/lib/lib.d.ts' // import * as typescriptes5 from '!!raw-loader!../../../node_modules/typescript/lib/lib.es5.d.ts' -// import typescriptlib from 'type_decls.d'; +import typescriptlib from 'type_decls.d'; import * as ts from 'typescript'; import { Doc, FieldType } from '../../fields/Doc'; import { RefField } from '../../fields/RefField'; @@ -29,7 +29,7 @@ export interface CompiledScript { readonly compiled: true; readonly originalScript: string; // eslint-disable-next-line no-use-before-define - readonly options: Readonly; + readonly options: Readonly; run(args?: { [name: string]: unknown }, onError?: (res: string) => void, errorVal?: unknown): ScriptResult; } @@ -60,7 +60,6 @@ function Run(script: string | undefined, customParams: string[], diagnostics: ts // let params: any[] = [Docs, ...fieldTypes]; const compiledFunction = (() => { try { - // eslint-disable-next-line no-new-func return new Function(...paramNames, `return ${script}`); } catch (e) { console.log(e); @@ -69,10 +68,8 @@ function Run(script: string | undefined, customParams: string[], diagnostics: ts })(); if (!compiledFunction) return { compiled: false, errors }; const { capturedVariables = {} } = options; - // eslint-disable-next-line default-param-last const run = (args: { [name: string]: unknown } = {}, onError?: (e: string) => void, errorVal?: ts.Diagnostic): ScriptResult => { const argsArray: unknown[] = []; - // eslint-disable-next-line no-restricted-syntax for (const name of customParams) { if (name !== 'this') { argsArray.push(name in args ? args[name] : capturedVariables[name]); @@ -154,7 +151,7 @@ class ScriptingCompilerHost { export type Traverser = (node: ts.Node, indentation: string) => boolean | void; export type TraverserParam = Traverser | { onEnter: Traverser; onLeave: Traverser }; export type Transformer = { - transformer: ts.TransformerFactory; + transformer: ts.TransformerFactory; getVars?: () => { [name: string]: FieldType }; }; export interface ScriptOptions { @@ -224,7 +221,6 @@ export function CompileScript(script: string, options: ScriptOptions = {}): Comp if ('this' in params || 'this' in capturedVariables) { paramNames.push('this'); } - // eslint-disable-next-line no-restricted-syntax for (const key in params) { if (key !== 'this') { paramNames.push(key); @@ -234,7 +230,6 @@ export function CompileScript(script: string, options: ScriptOptions = {}): Comp const val = params[key]; return `${key}: ${val}`; }); - // eslint-disable-next-line no-restricted-syntax for (const key in capturedVariables) { if (key !== 'this') { const val = capturedVariables[key]; @@ -248,7 +243,7 @@ export function CompileScript(script: string, options: ScriptOptions = {}): Comp const funcScript = `(function(${paramString})${reqTypes} { ${body} })`; host.writeFile('file.ts', funcScript); - // if (typecheck) host.writeFile('node_modules/typescript/lib/lib.d.ts', typescriptlib); + if (typecheck) host.writeFile('node_modules/typescript/lib/lib.d.ts', typescriptlib); const program = ts.createProgram(['file.ts'], {}, host); const testResult = program.emit(); const outputText = host.readFile('file.js'); diff --git a/src/client/views/MainView.tsx b/src/client/views/MainView.tsx index 8b8f85dfb..abe154de4 100644 --- a/src/client/views/MainView.tsx +++ b/src/client/views/MainView.tsx @@ -75,7 +75,7 @@ import { AnchorMenu } from './pdf/AnchorMenu'; import { GPTPopup } from './pdf/GPTPopup/GPTPopup'; import { TopBar } from './topbar/TopBar'; -// eslint-disable-next-line @typescript-eslint/no-var-requires, @typescript-eslint/no-require-imports +// eslint-disable-next-line @typescript-eslint/no-require-imports const { LEFT_MENU_WIDTH, TOPBAR_HEIGHT } = require('./global/globalCssVariables.module.scss'); // prettier-ignore @observer @@ -1059,10 +1059,7 @@ export class MainView extends ObservableReactComponent { docView={DocButtonState.Instance.LinkEditorDocView} /> ) : null} - {LinkInfo.Instance?.LinkInfo ? ( - // eslint-disable-next-line react/jsx-props-no-spreading - - ) : null} + {LinkInfo.Instance?.LinkInfo ? : null} {((page: string) => { // prettier-ignore switch (page) { diff --git a/src/client/views/collections/CollectionNoteTakingViewColumn.tsx b/src/client/views/collections/CollectionNoteTakingViewColumn.tsx index fc5f5cb71..226d06f37 100644 --- a/src/client/views/collections/CollectionNoteTakingViewColumn.tsx +++ b/src/client/views/collections/CollectionNoteTakingViewColumn.tsx @@ -252,10 +252,7 @@ export class CollectionNoteTakingViewColumn extends ObservableReactComponent
- { - // eslint-disable-next-line react/jsx-props-no-spreading - - } +
) : null} diff --git a/src/client/views/nodes/PDFBox.tsx b/src/client/views/nodes/PDFBox.tsx index 4616ec057..596975062 100644 --- a/src/client/views/nodes/PDFBox.tsx +++ b/src/client/views/nodes/PDFBox.tsx @@ -10,7 +10,7 @@ import { DocData } from '../../../fields/DocSymbols'; import { Id } from '../../../fields/FieldSymbols'; import { InkTool } from '../../../fields/InkField'; import { ComputedField } from '../../../fields/ScriptField'; -import { Cast, FieldValue, ImageCast, NumCast, StrCast, toList } from '../../../fields/Types'; +import { Cast, FieldValue, NumCast, StrCast, toList } from '../../../fields/Types'; import { ImageField, PdfField } from '../../../fields/URLField'; import { TraceMobx } from '../../../fields/util'; import { emptyFunction } from '../../../Utils'; @@ -43,14 +43,14 @@ export class PDFBox extends ViewBoxAnnotatableComponent() { public static openSidebarWidth = 250; public static sidebarResizerWidth = 5; private _searchString: string = ''; - private _initialScrollTarget: Opt; + private _initialScrollTarget: Opt; private _pdfViewer: PDFViewer | undefined; private _searchRef = React.createRef(); private _disposers: { [name: string]: IReactionDisposer } = {}; private _sidebarRef = React.createRef(); @observable private _searching: boolean = false; - @observable private _pdf: Opt = undefined; + @observable private _pdf: Opt = undefined; @observable private _pageControls = false; @computed get pdfUrl() { @@ -242,13 +242,13 @@ export class PDFBox extends ViewBoxAnnotatableComponent() { options.didMove = true; this.toggleSidebar(false); } - return new Promise(res => { + return new Promise>(res => { DocumentView.addViewRenderedCb(doc, dv => res(dv)); }); }; getAnchor = (addAsAnnotation: boolean, pinProps?: PinProps) => { - let ele: Opt; + let ele: Opt; if (this._pdfViewer?.selectionContent()) { ele = document.createElement('div'); ele.append(this._pdfViewer.selectionContent()!); @@ -327,7 +327,7 @@ export class PDFBox extends ViewBoxAnnotatableComponent() { this._initialScrollTarget = undefined; } }; - searchStringChanged = (e: React.ChangeEvent) => { + searchStringChanged = (e: React.ChangeEvent) => { this._searchString = e.currentTarget.value; }; @@ -365,8 +365,8 @@ export class PDFBox extends ViewBoxAnnotatableComponent() { } ); }; - @observable _previewNativeWidth: Opt = undefined; - @observable _previewWidth: Opt = undefined; + @observable _previewNativeWidth: Opt = undefined; + @observable _previewWidth: Opt = undefined; toggleSidebar = action((preview: boolean = false) => { const nativeWidth = NumCast(this.layoutDoc[this.fieldKey + '_nativeWidth']); const sideratio = ((!this.layoutDoc.nativeWidth || this.layoutDoc.nativeWidth === nativeWidth ? PDFBox.openSidebarWidth : 0) + nativeWidth) / nativeWidth; @@ -540,7 +540,6 @@ export class PDFBox extends ViewBoxAnnotatableComponent() { return ComponentTag === CollectionStackingView ? ( () { ) : (
setupMoveUpEvents(this, e, returnFalse, emptyFunction, () => this._props.select(false), true)}> () { top: 0, }}> () { } static pdfcache = new Map(); - static pdfpromise = new Map(); + static pdfpromise = new Map>(); render() { TraceMobx(); const pdfView = !this._pdf ? null : this.renderPdfView; diff --git a/src/client/views/nodes/RecordingBox/ProgressBar.tsx b/src/client/views/nodes/RecordingBox/ProgressBar.tsx index 62798bc2f..7e91df7ab 100644 --- a/src/client/views/nodes/RecordingBox/ProgressBar.tsx +++ b/src/client/views/nodes/RecordingBox/ProgressBar.tsx @@ -1,5 +1,3 @@ -/* eslint-disable react/no-array-index-key */ -/* eslint-disable react/require-default-props */ import * as React from 'react'; import { useEffect, useState, useRef } from 'react'; import './ProgressBar.scss'; diff --git a/src/client/views/nodes/chatbot/agentsystem/Agent.ts b/src/client/views/nodes/chatbot/agentsystem/Agent.ts index 180d05cf3..ccf9caf15 100644 --- a/src/client/views/nodes/chatbot/agentsystem/Agent.ts +++ b/src/client/views/nodes/chatbot/agentsystem/Agent.ts @@ -1,20 +1,19 @@ +import dotenv from 'dotenv'; +import { XMLBuilder, XMLParser } from 'fast-xml-parser'; import OpenAI from 'openai'; -import { Tool, AgentMessage, AssistantMessage, TEXT_TYPE, CHUNK_TYPE, ASSISTANT_ROLE, ProcessingInfo, PROCESSING_TYPE } from '../types/types'; -import { getReactPrompt } from './prompts'; -import { XMLParser, XMLBuilder } from 'fast-xml-parser'; -import { Vectorstore } from '../vectorstore/Vectorstore'; import { ChatCompletionMessageParam } from 'openai/resources'; -import dotenv from 'dotenv'; -import { CalculateTool } from '../tools/CalculateTool'; -import { RAGTool } from '../tools/RAGTool'; -import { DataAnalysisTool } from '../tools/DataAnalysisTool'; -import { WebsiteInfoScraperTool } from '../tools/WebsiteInfoScraperTool'; -import { SearchTool } from '../tools/SearchTool'; -import { NoTool } from '../tools/NoTool'; -import { v4 as uuidv4 } from 'uuid'; import { AnswerParser } from '../response_parsers/AnswerParser'; import { StreamedAnswerParser } from '../response_parsers/StreamedAnswerParser'; +import { CalculateTool } from '../tools/CalculateTool'; import { CreateCSVTool } from '../tools/CreateCSVTool'; +import { DataAnalysisTool } from '../tools/DataAnalysisTool'; +import { NoTool } from '../tools/NoTool'; +import { RAGTool } from '../tools/RAGTool'; +import { SearchTool } from '../tools/SearchTool'; +import { WebsiteInfoScraperTool } from '../tools/WebsiteInfoScraperTool'; +import { AgentMessage, AssistantMessage, PROCESSING_TYPE, ProcessingInfo, Tool } from '../types/types'; +import { Vectorstore } from '../vectorstore/Vectorstore'; +import { getReactPrompt } from './prompts'; dotenv.config(); @@ -25,7 +24,7 @@ dotenv.config(); export class Agent { // Private properties private client: OpenAI; - private tools: Record>; + private tools: Record>; // bcz: need a real type here private messages: AgentMessage[] = []; private interMessages: AgentMessage[] = []; private vectorstore: Vectorstore; @@ -102,7 +101,7 @@ export class Agent { ignoreAttributes: false, attributeNamePrefix: '@_', textNodeName: '_text', - isArray: (name, jpath, isLeafNode, isAttribute) => ['query', 'url'].indexOf(name) !== -1, + isArray: (name /* , jpath, isLeafNode, isAttribute */) => ['query', 'url'].indexOf(name) !== -1, }); const builder = new XMLBuilder({ ignoreAttributes: false, attributeNamePrefix: '@_' }); @@ -167,7 +166,7 @@ export class Agent { if (currentAction) { try { // Process the action with its input - const observation = await this.processAction(currentAction, actionInput.inputs); + const observation = (await this.processAction(currentAction, actionInput.inputs)) as any; // bcz: really need a type here const nextPrompt = [{ type: 'text', text: ` ` }, ...observation, { type: 'text', text: '' }]; console.log(observation); this.interMessages.push({ role: 'user', content: nextPrompt }); @@ -214,7 +213,7 @@ export class Agent { // Process each chunk of the streamed response for await (const chunk of stream) { - let content = chunk.choices[0]?.delta?.content || ''; + const content = chunk.choices[0]?.delta?.content || ''; fullResponse += content; // Parse the streamed content character by character @@ -267,7 +266,7 @@ export class Agent { * @param actionInput The inputs for the action. * @returns The result of the action. */ - private async processAction(action: string, actionInput: any): Promise { + private async processAction(action: string, actionInput: unknown): Promise { if (!(action in this.tools)) { throw new Error(`Unknown action: ${action}`); } diff --git a/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx b/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx index d4deff78b..613cb7078 100644 --- a/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx +++ b/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx @@ -1,30 +1,29 @@ -import { action, computed, makeObservable, observable, observe, reaction, runInAction, ObservableSet } from 'mobx'; +import dotenv from 'dotenv'; +import { ObservableSet, action, computed, makeObservable, observable, observe, reaction, runInAction } from 'mobx'; import { observer } from 'mobx-react'; import OpenAI, { ClientOptions } from 'openai'; import * as React from 'react'; +import { v4 as uuidv4 } from 'uuid'; +import { ClientUtils } from '../../../../../ClientUtils'; import { Doc, DocListCast } from '../../../../../fields/Doc'; +import { DocData, DocViews } from '../../../../../fields/DocSymbols'; import { CsvCast, DocCast, PDFCast, RTFCast, StrCast } from '../../../../../fields/Types'; +import { Networking } from '../../../../Network'; +import { DocUtils } from '../../../../documents/DocUtils'; import { DocumentType } from '../../../../documents/DocumentTypes'; import { Docs } from '../../../../documents/Documents'; +import { DocumentManager } from '../../../../util/DocumentManager'; import { LinkManager } from '../../../../util/LinkManager'; import { ViewBoxAnnotatableComponent } from '../../../DocComponent'; +import { DocumentView } from '../../DocumentView'; import { FieldView, FieldViewProps } from '../../FieldView'; +import { PDFBox } from '../../PDFBox'; +import { Agent } from '../agentsystem/Agent'; +import { ASSISTANT_ROLE, AssistantMessage, CHUNK_TYPE, Citation, ProcessingInfo, SimplifiedChunk, TEXT_TYPE } from '../types/types'; +import { Vectorstore } from '../vectorstore/Vectorstore'; import './ChatBox.scss'; import MessageComponentBox from './MessageComponent'; -import { ASSISTANT_ROLE, AssistantMessage, Citation, CHUNK_TYPE, TEXT_TYPE, SimplifiedChunk, ProcessingInfo } from '../types/types'; -import { Vectorstore } from '../vectorstore/Vectorstore'; -import { Agent } from '../agentsystem/Agent'; -import dotenv from 'dotenv'; -import { DocData, DocViews } from '../../../../../fields/DocSymbols'; -import { DocumentManager } from '../../../../util/DocumentManager'; -import { v4 as uuidv4 } from 'uuid'; -import { DocUtils } from '../../../../documents/DocUtils'; -import { ClientUtils } from '../../../../../ClientUtils'; import { ProgressBar } from './ProgressBar'; -import { DocumentView } from '../../DocumentView'; -import { Networking } from '../../../../Network'; -import { PDFViewer } from '../../../pdf/PDFViewer'; -import { PDFBox } from '../../PDFBox'; dotenv.config(); @@ -247,7 +246,7 @@ export class ChatBox extends ViewBoxAnnotatableComponent() { * @param event The form submission event. */ @action - askGPT = async (event: React.FormEvent): Promise => { + askGPT = async (event: React.FormEvent): Promise => { event.preventDefault(); this.inputValue = ''; @@ -413,7 +412,7 @@ export class ChatBox extends ViewBoxAnnotatableComponent() { const chunkId = citation.chunk_id; // Loop through the linked documents to find the matching chunk and handle its display - for (let doc of currentLinkedDocs) { + for (const doc of currentLinkedDocs) { if (doc.chunk_simpl) { const docChunkSimpl = JSON.parse(StrCast(doc.chunk_simpl)) as { chunks: SimplifiedChunk[] }; const foundChunk = docChunkSimpl.chunks.find(chunk => chunk.chunkId === chunkId); @@ -422,24 +421,26 @@ export class ChatBox extends ViewBoxAnnotatableComponent() { switch (foundChunk.chunkType) { case CHUNK_TYPE.IMAGE: case CHUNK_TYPE.TABLE: - const values = foundChunk.location?.replace(/[\[\]]/g, '').split(','); + { + const values = foundChunk.location?.replace(/[[\]]/g, '').split(','); - if (values?.length !== 4) { - console.error('Location string must contain exactly 4 numbers'); - return; - } + if (values?.length !== 4) { + console.error('Location string must contain exactly 4 numbers'); + return; + } - const x1 = parseFloat(values[0]) * Doc.NativeWidth(doc); - const y1 = parseFloat(values[1]) * Doc.NativeHeight(doc) + foundChunk.startPage * Doc.NativeHeight(doc); - const x2 = parseFloat(values[2]) * Doc.NativeWidth(doc); - const y2 = parseFloat(values[3]) * Doc.NativeHeight(doc) + foundChunk.startPage * Doc.NativeHeight(doc); + const x1 = parseFloat(values[0]) * Doc.NativeWidth(doc); + const y1 = parseFloat(values[1]) * Doc.NativeHeight(doc) + foundChunk.startPage * Doc.NativeHeight(doc); + const x2 = parseFloat(values[2]) * Doc.NativeWidth(doc); + const y2 = parseFloat(values[3]) * Doc.NativeHeight(doc) + foundChunk.startPage * Doc.NativeHeight(doc); - const annotationKey = Doc.LayoutFieldKey(doc) + '_annotations'; + const annotationKey = Doc.LayoutFieldKey(doc) + '_annotations'; - const existingDoc = DocListCast(doc[DocData][annotationKey]).find(d => d.citation_id === citation.citation_id); - const highlightDoc = existingDoc ?? this.createImageCitationHighlight(x1, y1, x2, y2, citation, annotationKey, doc); + const existingDoc = DocListCast(doc[DocData][annotationKey]).find(d => d.citation_id === citation.citation_id); + const highlightDoc = existingDoc ?? this.createImageCitationHighlight(x1, y1, x2, y2, citation, annotationKey, doc); - DocumentManager.Instance.showDocument(highlightDoc, { willZoomCentered: true }, () => {}); + DocumentManager.Instance.showDocument(highlightDoc, { willZoomCentered: true }, () => {}); + } break; case CHUNK_TYPE.TEXT: this.citationPopup = { text: citation.direct_text ?? 'No text available', visible: true }; @@ -695,7 +696,7 @@ export class ChatBox extends ViewBoxAnnotatableComponent() {
)}
-

{this.userName()}'s AI Assistant

+

{this.userName()}'s AI Assistant

{this.history.map((message, index) => ( diff --git a/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts b/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts index 3b4fdb6f5..1ac753790 100644 --- a/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts +++ b/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts @@ -1,5 +1,5 @@ -import { ASSISTANT_ROLE, AssistantMessage, Citation, CHUNK_TYPE, TEXT_TYPE, getChunkType, ProcessingInfo } from '../types/types'; import { v4 as uuid } from 'uuid'; +import { ASSISTANT_ROLE, AssistantMessage, Citation, ProcessingInfo, TEXT_TYPE, getChunkType } from '../types/types'; export class AnswerParser { static parse(xml: string, processingInfo: ProcessingInfo[]): AssistantMessage { @@ -22,8 +22,8 @@ export class AnswerParser { } let rawTextContent = answerMatch[1].trim(); - let content: AssistantMessage['content'] = []; - let citations: Citation[] = []; + const content: AssistantMessage['content'] = []; + const citations: Citation[] = []; let contentIndex = 0; // Remove citations and follow-up questions from rawTextContent @@ -43,6 +43,7 @@ export class AnswerParser { if (citationsMatch) { const citationsContent = citationsMatch[1]; while ((citationMatch = citationRegex.exec(citationsContent)) !== null) { + // eslint-disable-next-line @typescript-eslint/no-unused-vars const [_, index, chunk_id, type, direct_text] = citationMatch; const citation_id = uuid(); citationMap.set(index, citation_id); @@ -102,7 +103,7 @@ export class AnswerParser { } } - let followUpQuestions: string[] = []; + const followUpQuestions: string[] = []; if (followUpQuestionsMatch) { const questionsText = followUpQuestionsMatch[1]; let questionMatch; diff --git a/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts b/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts index 3585cab4a..4149f3da9 100644 --- a/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts +++ b/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts @@ -1,5 +1,3 @@ -import { threadId } from 'worker_threads'; - enum ParserState { Outside, InGroundedText, diff --git a/src/client/views/nodes/chatbot/tools/BaseTool.ts b/src/client/views/nodes/chatbot/tools/BaseTool.ts index b57f1c8e4..10780617b 100644 --- a/src/client/views/nodes/chatbot/tools/BaseTool.ts +++ b/src/client/views/nodes/chatbot/tools/BaseTool.ts @@ -1,17 +1,17 @@ import { Tool } from '../types/types'; -export abstract class BaseTool = Record> implements Tool { +export abstract class BaseTool = Record> implements Tool { constructor( public name: string, public description: string, - public parameters: Record, + public parameters: Record, public citationRules: string, public briefSummary: string ) {} - abstract execute(args: T): Promise; + abstract execute(args: T): Promise; - getActionRule(): Record { + getActionRule(): Record { return { [this.name]: { name: this.name, diff --git a/src/client/views/nodes/chatbot/tools/CalculateTool.ts b/src/client/views/nodes/chatbot/tools/CalculateTool.ts index 74b7ca27b..77ab1b39b 100644 --- a/src/client/views/nodes/chatbot/tools/CalculateTool.ts +++ b/src/client/views/nodes/chatbot/tools/CalculateTool.ts @@ -18,7 +18,7 @@ export class CalculateTool extends BaseTool<{ expression: string }> { ); } - async execute(args: { expression: string }): Promise { + async execute(args: { expression: string }): Promise { // Note: Using eval() can be dangerous. Consider using a safer alternative. const result = eval(args.expression); return [{ type: 'text', text: result.toString() }]; diff --git a/src/client/views/nodes/chatbot/tools/CreateCSVTool.ts b/src/client/views/nodes/chatbot/tools/CreateCSVTool.ts index 55015846b..d3ded0de0 100644 --- a/src/client/views/nodes/chatbot/tools/CreateCSVTool.ts +++ b/src/client/views/nodes/chatbot/tools/CreateCSVTool.ts @@ -28,7 +28,7 @@ export class CreateCSVTool extends BaseTool<{ csvData: string; filename: string this._handleCSVResult = handleCSVResult; } - async execute(args: { csvData: string; filename: string }): Promise { + async execute(args: { csvData: string; filename: string }): Promise { try { console.log('Creating CSV file:', args.filename, ' with data:', args.csvData); // Post the raw CSV data to the createCSV endpoint on the server diff --git a/src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts b/src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts index 573428179..1e479a62c 100644 --- a/src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts +++ b/src/client/views/nodes/chatbot/tools/CreateCollectionTool.ts @@ -25,7 +25,7 @@ export class GetDocsContentTool extends BaseTool<{ title: string; document_ids: this._docView = docView; } - async execute(args: { title: string; document_ids: string[] }): Promise { + async execute(args: { title: string; document_ids: string[] }): Promise { // Note: Using eval() can be dangerous. Consider using a safer alternative. const docs = args.document_ids.map(doc_id => DocCast(DocServer.GetCachedRefField(doc_id))); const collection = Docs.Create.FreeformDocument(docs, { title: args.title }); diff --git a/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts b/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts index a12ee46e5..2e663fed1 100644 --- a/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts +++ b/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts @@ -33,7 +33,7 @@ export class DataAnalysisTool extends BaseTool<{ csv_file_name: string | string[ return file?.id; } - async execute(args: { csv_file_name: string | string[] }): Promise { + async execute(args: { csv_file_name: string | string[] }): Promise { const filenames = Array.isArray(args.csv_file_name) ? args.csv_file_name : [args.csv_file_name]; const results = []; diff --git a/src/client/views/nodes/chatbot/tools/GetDocsTool.ts b/src/client/views/nodes/chatbot/tools/GetDocsTool.ts index f970ca8ee..903f3f69c 100644 --- a/src/client/views/nodes/chatbot/tools/GetDocsTool.ts +++ b/src/client/views/nodes/chatbot/tools/GetDocsTool.ts @@ -18,7 +18,7 @@ export class GetDocsTool extends BaseTool<{ title: string; document_ids: string[ this._docView = docView; } - async execute(args: { title: string; document_ids: string[] }): Promise { + async execute(args: { title: string; document_ids: string[] }): Promise { // Note: Using eval() can be dangerous. Consider using a safer alternative. const docs = args.document_ids.map(doc_id => DocCast(DocServer.GetCachedRefField(doc_id))); const collection = Docs.Create.FreeformDocument(docs, { title: args.title }); diff --git a/src/client/views/nodes/chatbot/tools/NoTool.ts b/src/client/views/nodes/chatbot/tools/NoTool.ts index 1f0830a77..edd3160ec 100644 --- a/src/client/views/nodes/chatbot/tools/NoTool.ts +++ b/src/client/views/nodes/chatbot/tools/NoTool.ts @@ -1,7 +1,7 @@ // tools/NoTool.ts import { BaseTool } from './BaseTool'; -export class NoTool extends BaseTool<{}> { +export class NoTool extends BaseTool> { constructor() { super( 'no_tool', @@ -12,7 +12,8 @@ export class NoTool extends BaseTool<{}> { ); } - async execute(args: {}): Promise { + // eslint-disable-next-line @typescript-eslint/no-unused-vars + async execute(args: object): Promise { return [{ type: 'text', text: 'No tool used. Proceed with answering the question.' }]; } } diff --git a/src/client/views/nodes/chatbot/tools/RAGTool.ts b/src/client/views/nodes/chatbot/tools/RAGTool.ts index f4b7b42ea..4cc2f26ff 100644 --- a/src/client/views/nodes/chatbot/tools/RAGTool.ts +++ b/src/client/views/nodes/chatbot/tools/RAGTool.ts @@ -1,10 +1,7 @@ -import { BaseTool } from './BaseTool'; -import { Vectorstore } from '../vectorstore/Vectorstore'; -import { RAGChunk } from '../types/types'; -import * as fs from 'fs'; import { Networking } from '../../../../Network'; -import { file } from 'jszip'; -import { ChatCompletion, ChatCompletionContentPart, ChatCompletionMessageParam } from 'openai/resources'; +import { RAGChunk } from '../types/types'; +import { Vectorstore } from '../vectorstore/Vectorstore'; +import { BaseTool } from './BaseTool'; export class RAGTool extends BaseTool { constructor(private vectorstore: Vectorstore) { @@ -59,13 +56,13 @@ export class RAGTool extends BaseTool { ); } - async execute(args: { hypothetical_document_chunk: string }): Promise { + async execute(args: { hypothetical_document_chunk: string }): Promise { const relevantChunks = await this.vectorstore.retrieve(args.hypothetical_document_chunk); const formatted_chunks = await this.getFormattedChunks(relevantChunks); return formatted_chunks; } - async getFormattedChunks(relevantChunks: RAGChunk[]): Promise { + async getFormattedChunks(relevantChunks: RAGChunk[]): Promise { try { const { formattedChunks } = await Networking.PostToServer('/formatChunks', { relevantChunks }); diff --git a/src/client/views/nodes/chatbot/tools/SearchTool.ts b/src/client/views/nodes/chatbot/tools/SearchTool.ts index b926cbadc..3a4668422 100644 --- a/src/client/views/nodes/chatbot/tools/SearchTool.ts +++ b/src/client/views/nodes/chatbot/tools/SearchTool.ts @@ -1,7 +1,6 @@ -import { max } from 'lodash'; +import { v4 as uuidv4 } from 'uuid'; import { Networking } from '../../../../Network'; import { BaseTool } from './BaseTool'; -import { v4 as uuidv4 } from 'uuid'; export class SearchTool extends BaseTool<{ query: string | string[] }> { private _addLinkedUrlDoc: (url: string, id: string) => void; @@ -25,7 +24,7 @@ export class SearchTool extends BaseTool<{ query: string | string[] }> { this._max_results = max_results; } - async execute(args: { query: string | string[] }): Promise { + async execute(args: { query: string | string[] }): Promise { const queries = Array.isArray(args.query) ? args.query : [args.query]; const allResults = []; diff --git a/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts b/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts index 2118218f6..1efb389b8 100644 --- a/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts +++ b/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts @@ -1,6 +1,6 @@ +import { v4 as uuidv4 } from 'uuid'; import { Networking } from '../../../../Network'; import { BaseTool } from './BaseTool'; -import { v4 as uuidv4 } from 'uuid'; export class WebsiteInfoScraperTool extends BaseTool<{ url: string | string[] }> { private _addLinkedUrlDoc: (url: string, id: string) => void; @@ -63,7 +63,7 @@ export class WebsiteInfoScraperTool extends BaseTool<{ url: string | string[] }> this._addLinkedUrlDoc = addLinkedUrlDoc; } - async execute(args: { url: string | string[] }): Promise { + async execute(args: { url: string | string[] }): Promise { const urls = Array.isArray(args.url) ? args.url : [args.url]; const results = []; @@ -74,6 +74,7 @@ export class WebsiteInfoScraperTool extends BaseTool<{ url: string | string[] }> this._addLinkedUrlDoc(url, id); results.push({ type: 'text', text: `\n${website_plain_text}\n\n` }); } catch (error) { + console.log(error); results.push({ type: 'text', text: `An error occurred while scraping the website: ${url}` }); } } diff --git a/src/client/views/nodes/chatbot/tools/WikipediaTool.ts b/src/client/views/nodes/chatbot/tools/WikipediaTool.ts index 143d91d80..692dff749 100644 --- a/src/client/views/nodes/chatbot/tools/WikipediaTool.ts +++ b/src/client/views/nodes/chatbot/tools/WikipediaTool.ts @@ -1,8 +1,6 @@ -import { title } from 'process'; +import { v4 as uuidv4 } from 'uuid'; import { Networking } from '../../../../Network'; import { BaseTool } from './BaseTool'; -import axios from 'axios'; -import { v4 as uuidv4 } from 'uuid'; export class WikipediaTool extends BaseTool<{ title: string }> { private _addLinkedUrlDoc: (url: string, id: string) => void; @@ -23,7 +21,7 @@ export class WikipediaTool extends BaseTool<{ title: string }> { this._addLinkedUrlDoc = addLinkedUrlDoc; } - async execute(args: { title: string }): Promise { + async execute(args: { title: string }): Promise { try { const { text } = await Networking.PostToServer('/getWikipediaSummary', { title: args.title }); const id = uuidv4(); @@ -31,6 +29,7 @@ export class WikipediaTool extends BaseTool<{ title: string }> { this._addLinkedUrlDoc(url, id); return [{ type: 'text', text: ` ${text} ` }]; } catch (error) { + console.log(error); return [{ type: 'text', text: 'An error occurred while fetching the article.' }]; } } diff --git a/src/client/views/nodes/chatbot/types/types.ts b/src/client/views/nodes/chatbot/types/types.ts index f5d14ad6a..2bc7f4952 100644 --- a/src/client/views/nodes/chatbot/types/types.ts +++ b/src/client/views/nodes/chatbot/types/types.ts @@ -52,15 +52,6 @@ export interface ProcessingInfo { content: string; } -export interface AssistantMessage { - role: ASSISTANT_ROLE; - content: MessageContent[]; - follow_up_questions?: string[]; - citations?: Citation[]; - processing_info: ProcessingInfo[]; - loop_summary?: string; -} - export interface MessageContent { index: number; type: TEXT_TYPE; @@ -75,6 +66,14 @@ export interface Citation { citation_id: string; url?: string; } +export interface AssistantMessage { + role: ASSISTANT_ROLE; + content: MessageContent[]; + follow_up_questions?: string[]; + citations?: Citation[]; + processing_info: ProcessingInfo[]; + loop_summary?: string; +} export interface RAGChunk { id: string; @@ -113,14 +112,14 @@ export interface AI_Document { type: string; } -export interface Tool = Record> { +export interface Tool = Record> { name: string; description: string; - parameters: Record; + parameters: Record; citationRules: string; briefSummary: string; - execute: (args: T) => Promise; - getActionRule: () => Record; + execute: (args: T) => Promise; + getActionRule: () => Record; } export interface AgentMessage { diff --git a/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts b/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts index 07a2b73bc..9575277f7 100644 --- a/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts +++ b/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts @@ -1,11 +1,11 @@ -import { Pinecone, Index, IndexList, PineconeRecord, RecordMetadata, QueryResponse } from '@pinecone-database/pinecone'; +import { Index, IndexList, Pinecone, PineconeRecord, QueryResponse, RecordMetadata } from '@pinecone-database/pinecone'; import { CohereClient } from 'cohere-ai'; import { EmbedResponse } from 'cohere-ai/api'; import dotenv from 'dotenv'; -import { RAGChunk, AI_Document, CHUNK_TYPE } from '../types/types'; import { Doc } from '../../../../../fields/Doc'; import { CsvCast, PDFCast, StrCast } from '../../../../../fields/Types'; import { Networking } from '../../../../Network'; +import { AI_Document, CHUNK_TYPE, RAGChunk } from '../types/types'; dotenv.config(); @@ -103,8 +103,8 @@ export class Vectorstore { const { jobId } = await Networking.PostToServer('/createDocument', { file_path: local_file_path }); // Poll the server for progress updates. - let inProgress: boolean = true; - let result: any = null; + const inProgress = true; + let result: (AI_Document & { doc_id: string }) | null = null; // bcz: is this the correct type?? while (inProgress) { // Polling interval for status updates. await new Promise(resolve => setTimeout(resolve, 2000)); @@ -127,6 +127,9 @@ export class Vectorstore { progressCallback(progress, step); } } + if (!result) { + throw new Error('no result received...'); // bcz: is this an Error? + } // Once completed, process the document and add it to the vectorstore. console.log('Document JSON:', result); @@ -175,7 +178,7 @@ export class Vectorstore { * Indexes the processed document by uploading the document's vector chunks to the Pinecone index. * @param document The processed document containing its chunks and metadata. */ - private async indexDocument(document: any) { + private async indexDocument(document: AI_Document) { console.log('Uploading vectors to content namespace...'); // Prepare Pinecone records for each chunk in the document. diff --git a/src/client/views/pdf/PDFViewer.tsx b/src/client/views/pdf/PDFViewer.tsx index 7543b3fb1..b5c69bff0 100644 --- a/src/client/views/pdf/PDFViewer.tsx +++ b/src/client/views/pdf/PDFViewer.tsx @@ -51,7 +51,7 @@ interface IViewerProps extends FieldViewProps { * Handles rendering and virtualization of the pdf */ @observer -export class PDFViewer extends ObservableReactComponent { +export class PDFViewer extends ObservableReactComponent { static _annotationStyle = addStyleSheet(); constructor(props: IViewerProps) { @@ -68,12 +68,12 @@ export class PDFViewer extends ObservableReactComponent { private _pdfViewer!: PDFJSViewer.PDFViewer; private _styleRule: number | undefined; // stylesheet rule for making hyperlinks clickable private _retries = 0; // number of times tried to create the PDF viewer - private _setPreviewCursor: undefined | ((x: number, y: number, drag: boolean, hide: boolean, doc: Opt) => void); + private _setPreviewCursor: undefined | ((x: number, y: number, drag: boolean, hide: boolean, doc: Opt) => void); private _marqueeref = React.createRef(); - private _annotationLayer: React.RefObject = React.createRef(); + private _annotationLayer: React.RefObject = React.createRef(); private _disposers: { [name: string]: IReactionDisposer } = {}; - private _viewer: React.RefObject = React.createRef(); - _mainCont: React.RefObject = React.createRef(); + private _viewer: React.RefObject = React.createRef(); + _mainCont: React.RefObject = React.createRef(); private _selectionText: string = ''; private _selectionContent: DocumentFragment | undefined; private _downX: number = 0; @@ -81,9 +81,9 @@ export class PDFViewer extends ObservableReactComponent { private _lastSearch = false; private _viewerIsSetup = false; private _ignoreScroll = false; - private _initialScroll: { loc: Opt; easeFunc: 'linear' | 'ease' | undefined } | undefined; + private _initialScroll: { loc: Opt; easeFunc: 'linear' | 'ease' | undefined } | undefined; private _forcedScroll = true; - _getAnchor: (savedAnnotations: Opt, addAsAnnotation: boolean) => Opt = () => undefined; + _getAnchor: (savedAnnotations: Opt>, addAsAnnotation: boolean) => Opt = () => undefined; selectionText = () => this._selectionText; selectionContent = () => this._selectionContent; @@ -179,7 +179,7 @@ export class PDFViewer extends ObservableReactComponent { // otherwise it will scroll smoothly. scrollFocus = (doc: Doc, scrollTop: number, options: FocusViewOptions) => { const mainCont = this._mainCont.current; - let focusSpeed: Opt; + let focusSpeed: Opt; if (doc !== this._props.Document && mainCont) { const windowHeight = this._props.PanelHeight() / (this._props.NativeDimScaling?.() || 1); const scrollTo = ClientUtils.scrollIntoView(scrollTop, doc[Height](), NumCast(this._props.layoutDoc._layout_scrollTop), windowHeight, windowHeight * 0.1, this._scrollHeight); @@ -394,30 +394,6 @@ export class PDFViewer extends ObservableReactComponent { } }; - // @action - // createMarquee = (coords: [x1: number, x2: number, y1: number, y2: number]): void => { - // // const hit = document.elementFromPoint(e.clientX, e.clientY); - // // bcz: Change. drag selecting requires that preventDefault is NOT called. This used to happen in DocumentView, - // // but that's changed, so this shouldn't be needed. - // // if (hit && hit.localName === "span" && this.annotationsActive(true)) { // drag selecting text stops propagation - // // e.button === 0 && e.stopPropagation(); - // // } - // // if alt+left click, drag and annotate - // this._downX = coords[0]; - // this._downY = coords[2]; - - // if ((this._props.Document._freeform_scale || 1) !== 1) return; - // this._props.select(false); - // MarqueeAnnotator.clearAnnotations(this._savedAnnotations); - // this.isAnnotating = true; - // this._textSelecting = false; - // // if textLayer is hit, then we select text instead of using a marquee so clear out the marquee. - // this._styleRule = addStyleSheetRule(PDFViewer._annotationStyle, 'htmlAnnotation', { 'pointer-events': 'none' }); - - // this._marqueeref.current?.onInitiateSelection([coords[0], coords[2]]); - // this._marqueeref.current?.onTerminateSelection(); - // }; - @action finishMarquee = (/* x?: number, y?: number */) => { this._getAnchor = AnchorMenu.Instance?.GetAnchor; @@ -495,7 +471,7 @@ export class PDFViewer extends ObservableReactComponent { // e.stopPropagation(); // bcz: not sure why this was here. We need to allow the DocumentView to get clicks to process doubleClicks }; - setPreviewCursor = (func?: (x: number, y: number, drag: boolean, hide: boolean, doc: Opt) => void) => { + setPreviewCursor = (func?: (x: number, y: number, drag: boolean, hide: boolean, doc: Opt) => void) => { this._setPreviewCursor = func; }; @@ -520,7 +496,6 @@ export class PDFViewer extends ObservableReactComponent { return (
{inlineAnnos.map(anno => ( - // eslint-disable-next-line react/jsx-props-no-spreading ))}
@@ -534,7 +509,7 @@ export class PDFViewer extends ObservableReactComponent { panelHeight = () => this._props.PanelHeight() / (this._props.NativeDimScaling?.() || 1); transparentFilter = () => [...this._props.childFilters(), ClientUtils.TransparentBackgroundFilter]; opaqueFilter = () => [...this._props.childFilters(), ClientUtils.noDragDocsFilter, ...(SnappingManager.CanEmbed && this._props.isContentActive() ? [] : [ClientUtils.OpaqueBackgroundFilter])]; - childStyleProvider = (doc: Doc | undefined, props: Opt, property: string) => { + childStyleProvider = (doc: Doc | undefined, props: Opt, property: string) => { if (doc instanceof Doc && property === StyleProp.PointerEvents) { if (this.inlineTextAnnotations.includes(doc) || this._props.isContentActive() === false) return 'none'; const isInk = doc.layout_isSvg && !props?.LayoutTemplateString; @@ -554,7 +529,6 @@ export class PDFViewer extends ObservableReactComponent { pointerEvents: Doc.ActiveTool !== InkTool.None ? 'all' : undefined, }}> ({ + response.data.items?.map(item => ({ url: item.link, snippet: item.snippet, })) || []; res.send({ results }); - } catch (error: any) { + } catch (error) { console.error('Error performing web search:', error); res.status(500).send({ error: 'Failed to perform web search', - details: error.message, + details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -170,16 +166,16 @@ export default class AssistantManager extends ApiManager { * @param retries The number of retry attempts. * @param backoff Initial backoff time in milliseconds. */ - const fetchWithRetry = async (url: string, retries = 3, backoff = 300) => { + const fetchWithRetry = async (url: string, retries = 3, backoff = 300): Promise => { try { const response = await axiosInstance.get(url); return response.data; - } catch (error: any) { - if (retries > 0 && error.response?.status === 429) { + } catch (error) { + if (retries > 0 && (error as { response: { status: number } }).response?.status === 429) { // bcz: don't know the error type console.log(`Rate limited. Retrying in ${backoff}ms...`); await delay(backoff); return fetchWithRetry(url, retries - 1, backoff * 2); - } + } // prettier-ignore throw error; } }; @@ -199,11 +195,11 @@ export default class AssistantManager extends ApiManager { try { const data = await fetchWithRetry(url); res.send({ data }); - } catch (error: any) { + } catch (error) { console.error('Error fetching the URL:', error); res.status(500).send({ error: 'Failed to fetch the URL', - details: error.message, + details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -241,11 +237,11 @@ export default class AssistantManager extends ApiManager { } else { res.status(500).send({ error: 'Failed to extract readable content' }); } - } catch (error: any) { + } catch (error) { console.error('Error scraping website:', error); res.status(500).send({ error: 'Failed to scrape website', - details: error.message, + details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -267,15 +263,16 @@ export default class AssistantManager extends ApiManager { const jobId = uuid.v4(); // Spawn the Python process and track its progress/output + // eslint-disable-next-line no-use-before-define spawnPythonProcess(jobId, file_name, file_data); // Send the job ID back to the client for tracking res.send({ jobId }); - } catch (error: any) { + } catch (error) { console.error('Error initiating document creation:', error); res.status(500).send({ error: 'Failed to initiate document creation', - details: error.message, + details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -307,13 +304,13 @@ export default class AssistantManager extends ApiManager { const { jobId } = req.params; // Get the job ID from the URL parameters // Check if the job result is available if (jobResults[jobId]) { - const result = jobResults[jobId]; + const result = jobResults[jobId] as AI_Document & { status: string }; // If the result contains image or table chunks, save the base64 data as image files if (result.chunks && Array.isArray(result.chunks)) { for (const chunk of result.chunks) { if (chunk.metadata && (chunk.metadata.type === 'image' || chunk.metadata.type === 'table')) { - let files_directory = '/files/chunk_images/'; + const files_directory = '/files/chunk_images/'; const directory = path.join(publicDirectory, files_directory); // Ensure the directory exists or create it @@ -338,7 +335,7 @@ export default class AssistantManager extends ApiManager { } } } - result['status'] = 'completed'; + result.status = 'completed'; } else { result.status = 'pending'; } @@ -429,11 +426,11 @@ export default class AssistantManager extends ApiManager { // Send the file URL and UUID back to the client res.send({ fileUrl, id: uuidv4 }); - } catch (error: any) { + } catch (error) { console.error('Error creating CSV file:', error); res.status(500).send({ error: 'Failed to create CSV file.', - details: error.message, + details: (error as { message: string }).message ?? error, // bcz: don't know what the error type contains... }); } }, @@ -446,59 +443,6 @@ function spawnPythonProcess(jobId: string, file_name: string, file_data: string) const requirementsPath = path.join(__dirname, '../chunker/requirements.txt'); const pythonScriptPath = path.join(__dirname, '../chunker/pdf_chunker.py'); - // Check if venv exists - if (!fs.existsSync(venvPath)) { - console.log('Virtual environment not found. Creating and setting up...'); - - // Create venv - const createVenvProcess = spawn('python', ['-m', 'venv', venvPath]); - - createVenvProcess.on('close', code => { - if (code !== 0) { - console.error(`Failed to create virtual environment. Exit code: ${code}`); - return; - } - - console.log('Virtual environment created. Installing requirements...'); - - // Determine the pip path based on the OS - const pipPath = process.platform === 'win32' ? path.join(venvPath, 'Scripts', 'pip.exe') : path.join(venvPath, 'bin', 'pip3'); // Try 'pip3' for Unix-like systems - - if (!fs.existsSync(pipPath)) { - console.error(`pip executable not found at ${pipPath}`); - return; - } - - // Install requirements - const installRequirementsProcess = spawn(pipPath, ['install', '-r', requirementsPath]); - - installRequirementsProcess.stdout.on('data', data => { - console.log(`pip stdout: ${data}`); - }); - - installRequirementsProcess.stderr.on('data', data => { - console.error(`pip stderr: ${data}`); - }); - - installRequirementsProcess.on('error', error => { - console.error(`Error starting pip process: ${error}`); - }); - - installRequirementsProcess.on('close', code => { - if (code !== 0) { - console.error(`Failed to install requirements. Exit code: ${code}`); - return; - } - - console.log('Requirements installed. Running Python script...'); - runPythonScript(); - }); - }); - } else { - console.log('Virtual environment found. Running Python script...'); - runPythonScript(); - } - function runPythonScript() { const pythonPath = process.platform === 'win32' ? path.join(venvPath, 'Scripts', 'python') : path.join(venvPath, 'bin', 'python3'); @@ -530,7 +474,7 @@ function spawnPythonProcess(jobId: string, file_name: string, file_data: string) }; } } catch (err) { - console.error('Progress log from Python:', line); + console.error('Progress log from Python:', line, err); } } }); @@ -551,4 +495,56 @@ function spawnPythonProcess(jobId: string, file_name: string, file_data: string) } }); } + // Check if venv exists + if (!fs.existsSync(venvPath)) { + console.log('Virtual environment not found. Creating and setting up...'); + + // Create venv + const createVenvProcess = spawn('python', ['-m', 'venv', venvPath]); + + createVenvProcess.on('close', code => { + if (code !== 0) { + console.error(`Failed to create virtual environment. Exit code: ${code}`); + return; + } + + console.log('Virtual environment created. Installing requirements...'); + + // Determine the pip path based on the OS + const pipPath = process.platform === 'win32' ? path.join(venvPath, 'Scripts', 'pip.exe') : path.join(venvPath, 'bin', 'pip3'); // Try 'pip3' for Unix-like systems + + if (!fs.existsSync(pipPath)) { + console.error(`pip executable not found at ${pipPath}`); + return; + } + + // Install requirements + const installRequirementsProcess = spawn(pipPath, ['install', '-r', requirementsPath]); + + installRequirementsProcess.stdout.on('data', data => { + console.log(`pip stdout: ${data}`); + }); + + installRequirementsProcess.stderr.on('data', data => { + console.error(`pip stderr: ${data}`); + }); + + installRequirementsProcess.on('error', error => { + console.error(`Error starting pip process: ${error}`); + }); + + installRequirementsProcess.on('close', closecode => { + if (closecode !== 0) { + console.error(`Failed to install requirements. Exit code: ${closecode}`); + return; + } + + console.log('Requirements installed. Running Python script...'); + runPythonScript(); + }); + }); + } else { + console.log('Virtual environment found. Running Python script...'); + runPythonScript(); + } } -- cgit v1.2.3-70-g09d2 From d347fc59feefd91a796012892da57511787bb6d0 Mon Sep 17 00:00:00 2001 From: "A.J. Shulman" Date: Thu, 10 Oct 2024 11:39:19 -0400 Subject: added new file header comments and fixed some error handling --- .../views/nodes/chatbot/agentsystem/prompts.ts | 9 ++++++++- .../nodes/chatbot/chatboxcomponents/ChatBox.tsx | 9 +++++++++ .../chatbot/chatboxcomponents/MessageComponent.tsx | 9 +++++++++ .../chatbot/chatboxcomponents/ProgressBar.tsx | 8 ++++++++ .../nodes/chatbot/response_parsers/AnswerParser.ts | 8 ++++++++ .../response_parsers/StreamedAnswerParser.ts | 8 ++++++++ src/client/views/nodes/chatbot/tools/BaseTool.ts | 8 ++++++++ .../views/nodes/chatbot/vectorstore/Vectorstore.ts | 10 +++++++++- src/server/ApiManagers/AssistantManager.ts | 22 ++++++++++------------ 9 files changed, 77 insertions(+), 14 deletions(-) (limited to 'src/client/views/nodes/chatbot/tools') diff --git a/src/client/views/nodes/chatbot/agentsystem/prompts.ts b/src/client/views/nodes/chatbot/agentsystem/prompts.ts index 01c30d444..f5aec3130 100644 --- a/src/client/views/nodes/chatbot/agentsystem/prompts.ts +++ b/src/client/views/nodes/chatbot/agentsystem/prompts.ts @@ -1,4 +1,11 @@ -// prompts.ts +/** + * @file prompts.ts + * @description This file contains functions that generate prompts for various AI tasks, including + * generating system messages for structured AI assistant interactions and summarizing document chunks. + * It defines prompt structures to ensure the AI follows specific guidelines for response formatting, + * tool usage, and citation rules, with a rigid structure in mind for tasks such as answering user queries + * and summarizing content from provided text chunks. + */ import { Tool } from '../types/types'; diff --git a/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx b/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx index 613cb7078..44c231c87 100644 --- a/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx +++ b/src/client/views/nodes/chatbot/chatboxcomponents/ChatBox.tsx @@ -1,3 +1,12 @@ +/** + * @file ChatBox.tsx + * @description This file defines the ChatBox component, which manages user interactions with + * an AI assistant. It handles document uploads, chat history, message input, and integration + * with the OpenAI API. The ChatBox is MobX-observable and tracks the progress of tasks such as + * document analysis and AI-driven summaries. It also maintains real-time chat functionality + * with support for follow-up questions and citation management. + */ + import dotenv from 'dotenv'; import { ObservableSet, action, computed, makeObservable, observable, observe, reaction, runInAction } from 'mobx'; import { observer } from 'mobx-react'; diff --git a/src/client/views/nodes/chatbot/chatboxcomponents/MessageComponent.tsx b/src/client/views/nodes/chatbot/chatboxcomponents/MessageComponent.tsx index 801becb64..d48f46963 100644 --- a/src/client/views/nodes/chatbot/chatboxcomponents/MessageComponent.tsx +++ b/src/client/views/nodes/chatbot/chatboxcomponents/MessageComponent.tsx @@ -1,3 +1,12 @@ +/** + * @file MessageComponentBox.tsx + * @description This file defines the MessageComponentBox component, which renders the content + * of an AssistantMessage. It supports rendering various message types such as grounded text, + * normal text, and follow-up questions. The component uses React and MobX for state management + * and includes functionality for handling citation and follow-up actions, as well as displaying + * agent processing information. + */ + import React, { useState } from 'react'; import { observer } from 'mobx-react'; import { AssistantMessage, Citation, MessageContent, PROCESSING_TYPE, ProcessingInfo, TEXT_TYPE } from '../types/types'; diff --git a/src/client/views/nodes/chatbot/chatboxcomponents/ProgressBar.tsx b/src/client/views/nodes/chatbot/chatboxcomponents/ProgressBar.tsx index b9fd08742..240862f8b 100644 --- a/src/client/views/nodes/chatbot/chatboxcomponents/ProgressBar.tsx +++ b/src/client/views/nodes/chatbot/chatboxcomponents/ProgressBar.tsx @@ -1,3 +1,11 @@ +/** + * @file ProgressBar.tsx + * @description This file defines the ProgressBar component, which displays a loading spinner + * to indicate progress during ongoing tasks or processing. The animation consists of two + * bouncing elements that create a pulsating effect, providing a visual cue for active progress. + * The component is styled using the accompanying `ProgressBar.scss` for smooth animation. + */ + import React from 'react'; import './ProgressBar.scss'; diff --git a/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts b/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts index 1ac753790..ed78cc7cb 100644 --- a/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts +++ b/src/client/views/nodes/chatbot/response_parsers/AnswerParser.ts @@ -1,3 +1,11 @@ +/** + * @file AnswerParser.ts + * @description This file defines the AnswerParser class, which processes structured XML-like responses + * from the AI system, parsing grounded text, normal text, citations, follow-up questions, and loop summaries. + * The parser converts the XML response into an AssistantMessage format, extracting key information like + * citations and processing steps for further use in the assistant's workflow. + */ + import { v4 as uuid } from 'uuid'; import { ASSISTANT_ROLE, AssistantMessage, Citation, ProcessingInfo, TEXT_TYPE, getChunkType } from '../types/types'; diff --git a/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts b/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts index 4149f3da9..dbd568faa 100644 --- a/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts +++ b/src/client/views/nodes/chatbot/response_parsers/StreamedAnswerParser.ts @@ -1,3 +1,11 @@ +/** + * @file StreamedAnswerParser.ts + * @description This file defines the StreamedAnswerParser class, which parses incoming character streams + * to extract grounded or normal text based on the tags found in the input stream. It maintains state + * between grounded text and normal text sections, handling buffered input and ensuring proper text formatting + * for AI assistant responses. + */ + enum ParserState { Outside, InGroundedText, diff --git a/src/client/views/nodes/chatbot/tools/BaseTool.ts b/src/client/views/nodes/chatbot/tools/BaseTool.ts index 10780617b..a77f567a5 100644 --- a/src/client/views/nodes/chatbot/tools/BaseTool.ts +++ b/src/client/views/nodes/chatbot/tools/BaseTool.ts @@ -1,3 +1,11 @@ +/** + * @file BaseTool.ts + * @description This file defines the abstract BaseTool class, which serves as a blueprint + * for tool implementations in the AI assistant system. Each tool has a name, description, + * parameters, and citation rules. The BaseTool class provides a structure for executing actions + * and retrieving action rules for use within the assistant's workflow. + */ + import { Tool } from '../types/types'; export abstract class BaseTool = Record> implements Tool { diff --git a/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts b/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts index 9575277f7..f96f55997 100644 --- a/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts +++ b/src/client/views/nodes/chatbot/vectorstore/Vectorstore.ts @@ -1,3 +1,10 @@ +/** + * @file Vectorstore.ts + * @description This file defines the Vectorstore class, which integrates with Pinecone for vector-based document indexing and Cohere for text embeddings. + * It handles tasks such as AI document management, document chunking, and retrieval of relevant document sections based on user queries. + * The class supports adding documents to the vectorstore, managing document status, and querying Pinecone for document chunks matching a query. + */ + import { Index, IndexList, Pinecone, PineconeRecord, QueryResponse, RecordMetadata } from '@pinecone-database/pinecone'; import { CohereClient } from 'cohere-ai'; import { EmbedResponse } from 'cohere-ai/api'; @@ -128,7 +135,8 @@ export class Vectorstore { } } if (!result) { - throw new Error('no result received...'); // bcz: is this an Error? + console.error('Error processing document.'); + return; } // Once completed, process the document and add it to the vectorstore. diff --git a/src/server/ApiManagers/AssistantManager.ts b/src/server/ApiManagers/AssistantManager.ts index b4ebb1eae..b7d4191ca 100644 --- a/src/server/ApiManagers/AssistantManager.ts +++ b/src/server/ApiManagers/AssistantManager.ts @@ -1,3 +1,13 @@ +/** + * @file AssistantManager.ts + * @description This file defines the AssistantManager class, responsible for managing various + * API routes related to the Assistant functionality. It provides features such as file handling, + * web scraping, and integration with third-party APIs like OpenAI and Google Custom Search. + * It also handles job tracking and progress reporting for tasks like document creation and web scraping. + * Utility functions for path manipulation and file operations are included, along with + * a mechanism for handling retry logic during API calls. + */ + import { Readability } from '@mozilla/readability'; import axios from 'axios'; import { spawn } from 'child_process'; @@ -76,12 +86,6 @@ export default class AssistantManager extends ApiManager { * @param register The registration method to register routes and handlers. */ protected initialize(register: Registration): void { - // Initialize OpenAI API with client key - const openai = new OpenAI({ // bcz: is this needed? variable is never used... - apiKey: process.env._CLIENT_OPENAI_KEY, - dangerouslyAllowBrowser: true, - }); // prettier-ignore - // Initialize Google Custom Search API const customsearch = google.customsearch('v1'); @@ -107,7 +111,6 @@ export default class AssistantManager extends ApiManager { console.error('Error retrieving Wikipedia summary:', error); res.status(500).send({ error: 'Error retrieving article summary from Wikipedia.', - details: (error as { message: string }).message ?? error, // bcz: don't know what the error type contains... }); } }, @@ -140,7 +143,6 @@ export default class AssistantManager extends ApiManager { console.error('Error performing web search:', error); res.status(500).send({ error: 'Failed to perform web search', - details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -199,7 +201,6 @@ export default class AssistantManager extends ApiManager { console.error('Error fetching the URL:', error); res.status(500).send({ error: 'Failed to fetch the URL', - details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -241,7 +242,6 @@ export default class AssistantManager extends ApiManager { console.error('Error scraping website:', error); res.status(500).send({ error: 'Failed to scrape website', - details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -272,7 +272,6 @@ export default class AssistantManager extends ApiManager { console.error('Error initiating document creation:', error); res.status(500).send({ error: 'Failed to initiate document creation', - details: (error as { message: string }).message ?? error, // bcz: don't know wha tthe error type contains... }); } }, @@ -430,7 +429,6 @@ export default class AssistantManager extends ApiManager { console.error('Error creating CSV file:', error); res.status(500).send({ error: 'Failed to create CSV file.', - details: (error as { message: string }).message ?? error, // bcz: don't know what the error type contains... }); } }, -- cgit v1.2.3-70-g09d2