Merge branch 'nathan-starter' of https://github.com/brown-dash/Dash-Web into nathan-starter

11 files changed, 504 insertions, 0 deletions

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..a77f567a5
--- /dev/null
+++ b/src/client/views/nodes/chatbot/tools/BaseTool.ts
@@ -0,0 +1,32 @@
+/**
+ * @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<T extends Record<string, unknown> = Record<string, unknown>> implements Tool<T> {
+    constructor(
+        public name: string,
+        public description: string,
+        public parameters: Record<string, unknown>,
+        public citationRules: string,
+        public briefSummary: string
+    ) {}
+
+    abstract execute(args: T): Promise<unknown>;
+
+    getActionRule(): Record<string, unknown> {
+        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..77ab1b39b
--- /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<unknown> {
+        // 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..d3ded0de0
--- /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<unknown> {
+        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..1e479a62c
--- /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<unknown> {
+        // 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..2e663fed1
--- /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<unknown> {
+        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: `<chunk chunk_id=${fileID} chunk_type=csv>${fileContent}</chunk>`,
+                });
+            } 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..903f3f69c
--- /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<unknown> {
+        // 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..edd3160ec
--- /dev/null
+++ b/src/client/views/nodes/chatbot/tools/NoTool.ts
@@ -0,0 +1,19 @@
+// tools/NoTool.ts
+import { BaseTool } from './BaseTool';
+
+export class NoTool extends BaseTool<Record<string, unknown>> {
+    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 <action_input> 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.'
+        );
+    }
+
+    // eslint-disable-next-line @typescript-eslint/no-unused-vars
+    async execute(args: object): Promise<unknown> {
+        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..4cc2f26ff
--- /dev/null
+++ b/src/client/views/nodes/chatbot/tools/RAGTool.ts
@@ -0,0 +1,79 @@
+import { Networking } from '../../../../Network';
+import { RAGChunk } from '../types/types';
+import { Vectorstore } from '../vectorstore/Vectorstore';
+import { BaseTool } from './BaseTool';
+
+export class RAGTool extends BaseTool {
+    constructor(private vectorstore: Vectorstore) {
+        super(
+            'rag',
+            'Perform a RAG search on user documents',
+            {
+                hypothetical_document_chunk: {
+                    type: 'string',
+                    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',
+                },
+            },
+            `
+            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. **Grounded Text Guidelines**:
+                - Each <grounded_text> 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 <grounded_text> tags for each.
+
+            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.
+
+            **Example**:
+
+            <answer>
+                <grounded_text citation_index="1">
+                    Artificial Intelligence is revolutionizing various sectors, with healthcare seeing transformations in diagnosis and treatment planning.
+                </grounded_text>
+                <grounded_text citation_index="2">
+                    Based on recent data, AI has drastically improved mammogram analysis, achieving 99% accuracy at a rate 30 times faster than human radiologists.
+                </grounded_text>
+
+                <citations>
+                    <citation index="1" chunk_id="abc123" type="text">Artificial Intelligence is revolutionizing various industries, especially in healthcare.</citation>
+                    <citation index="2" chunk_id="abc124" type="table"></citation>
+                </citations>
+
+                <follow_up_questions>
+                    <question>How can AI enhance patient outcomes in fields outside radiology?</question>
+                    <question>What are the challenges in implementing AI systems across different hospitals?</question>
+                    <question>How might AI-driven advancements impact healthcare costs?</question>
+                </follow_up_questions>
+            </answer>
+            `,
+
+            `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<unknown> {
+        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<unknown> {
+        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..3a4668422
--- /dev/null
+++ b/src/client/views/nodes/chatbot/tools/SearchTool.ts
@@ -0,0 +1,53 @@
+import { v4 as uuidv4 } from 'uuid';
+import { Networking } from '../../../../Network';
+import { BaseTool } from './BaseTool';
+
+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<unknown> {
+        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: `<chunk chunk_id="${id}" chunk_type="text">
+                                <url>${result.url}</url>
+                                <overview>${result.snippet}</overview>
+                            </chunk>`,
+                    };
+                });
+                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..1efb389b8
--- /dev/null
+++ b/src/client/views/nodes/chatbot/tools/WebsiteInfoScraperTool.ts
@@ -0,0 +1,84 @@
+import { v4 as uuidv4 } from 'uuid';
+import { Networking } from '../../../../Network';
+import { BaseTool } from './BaseTool';
+
+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 <grounded_text> tags.
+               - **Do not include non-sourced information** in <grounded_text> tags.
+               - Use a single <grounded_text> tag for content derived from a single website. If citing multiple websites, create new <grounded_text> tags for each.
+               - Ensure each <grounded_text> tag has a citation index corresponding to the scraped URL.
+
+            2. Citation Tag Structure:
+               - Create a <citation> tag for each distinct piece of information used from the website(s).
+               - Each <citation> 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 <grounded_text> tags correspond to valid citations.
+               - Do not over-cite—cite only the most relevant parts of the websites.
+
+            Example Usage:
+
+            <answer>
+                <grounded_text citation_index="1">
+                Based on data from the World Bank, economic growth has stabilized in recent years, following a surge in investments.
+                </grounded_text>
+                <grounded_text citation_index="2">
+                According to information retrieved from the International Monetary Fund, the inflation rate has been gradually decreasing since 2020.
+                </grounded_text>
+
+                <citations>
+                    <citation index="1" chunk_id="1234" type="url"></citation>
+                    <citation index="2" chunk_id="5678" type="url"></citation>
+                </citations>
+
+                <follow_up_questions>
+                    <question>What are the long-term economic impacts of increased investments on GDP?</question>
+                    <question>How might inflation trends affect future monetary policy?</question>
+                    <question>Are there additional factors that could influence economic growth beyond investments and inflation?</question>
+                </follow_up_questions>
+            </answer>
+            `,
+            'Returns the text content of the webpages for further analysis and grounding.'
+        );
+        this._addLinkedUrlDoc = addLinkedUrlDoc;
+    }
+
+    async execute(args: { url: string | string[] }): Promise<unknown> {
+        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: `<chunk chunk_id=${id} chunk_type=url>\n${website_plain_text}\n</chunk>\n` });
+            } catch (error) {
+                console.log(error);
+                results.push({ type: 'text', text: `An error occurred while scraping the website: ${url}` });
+            }
+        }
+
+        return results;
+    }
+}
diff --git a/src/client/views/nodes/chatbot/tools/WikipediaTool.ts b/src/client/views/nodes/chatbot/tools/WikipediaTool.ts
new file mode 100644
index 000000000..692dff749
--- /dev/null
+++ b/src/client/views/nodes/chatbot/tools/WikipediaTool.ts
@@ -0,0 +1,36 @@
+import { v4 as uuidv4 } from 'uuid';
+import { Networking } from '../../../../Network';
+import { BaseTool } from './BaseTool';
+
+export class WikipediaTool extends BaseTool<{ title: string }> {
+    private _addLinkedUrlDoc: (url: string, id: string) => void;
+    constructor(addLinkedUrlDoc: (url: string, id: string) => void) {
+        super(
+            'wikipedia',
+            'Search Wikipedia and return a summary',
+            {
+                title: {
+                    type: 'string',
+                    description: 'The title of the Wikipedia article to search',
+                    required: true,
+                },
+            },
+            'Provide simply the title you want to search on Wikipedia and nothing more. If re-using this tool, try a different title for different information.',
+            'Returns a summary from searching an article title on Wikipedia'
+        );
+        this._addLinkedUrlDoc = addLinkedUrlDoc;
+    }
+
+    async execute(args: { title: string }): Promise<unknown> {
+        try {
+            const { text } = await Networking.PostToServer('/getWikipediaSummary', { title: args.title });
+            const id = uuidv4();
+            const url = `https://en.wikipedia.org/wiki/${args.title.replace(/ /g, '_')}`;
+            this._addLinkedUrlDoc(url, id);
+            return [{ type: 'text', text: `<chunk chunk_id=${id} chunk_type=csv}> ${text} </chunk>` }];
+        } catch (error) {
+            console.log(error);
+            return [{ type: 'text', text: 'An error occurred while fetching the article.' }];
+        }
+    }
+}