alchemy 0.3.1 → 0.4.1

package/src/internal/tutorial.ts

@@ -1,392 +0,0 @@
-import type { CoreMessage } from "ai";
-import { Approve, Document, Review } from "../ai";
-import type { ModelConfig } from "../ai/client";
-import { alchemy } from "../alchemy";
-import type { Context } from "../context";
-import type { Folder } from "../fs";
-import { Resource } from "../resource";
-
-/**
- * Properties for creating or updating a Tutorial
- */
-export interface TutorialProps {
-  /**
-   * The output directory for the tutorial document.
-   */
-  path: string | Folder;
-
-  /**
-   * The title of the tutorial.
-   */
-  title: string;
-
-  /**
-   * The prompt to use for generating the tutorial.
-   * This should include specific details about the tool or framework.
-   */
-  prompt: string;
-
-  /**
-   * Optional difficulty level for the tutorial.
-   * Defaults to "beginner".
-   */
-  difficulty?: "beginner" | "intermediate" | "advanced";
-
-  /**
-   * Optional estimated time to complete the tutorial in minutes.
-   * Defaults to 30.
-   */
-  estimatedTime?: number;
-
-  /**
-   * Optional model configuration for generating the tutorial.
-   * Defaults to Claude 3.5 Sonnet.
-   */
-  model?: ModelConfig;
-
-  /**
-   * Maximum number of review iterations.
-   * Defaults to 3.
-   */
-  maxIterations?: number;
-
-  /**
-   * Initial message history for the conversation.
-   * If not provided, a new conversation will be started.
-   */
-  messages?: CoreMessage[];
-}
-
-/**
- * Output returned after Tutorial creation/update
- */
-export interface Tutorial extends TutorialProps, Resource<"docs::Tutorial"> {
-  /**
-   * The Document resource
-   */
-  document: Document;
-
-  /**
-   * The content of the tutorial
-   */
-  content: string;
-
-  /**
-   * Message history from the conversation.
-   */
-  messages: CoreMessage[];
-
-  /**
-   * Time at which the tutorial was created
-   */
-  createdAt: number;
-
-  /**
-   * Time at which the tutorial was last updated
-   */
-  updatedAt: number;
-}
-
-/**
- * Creates a tutorial document with iterative review and improvement
- *
- * @example
- * // Create a basic tutorial
- * const tutorial = await Tutorial("getting-started", {
- *   path: "docs/tutorials/getting-started",
- *   title: "Getting Started with Alchemy",
- *   prompt: "Create a tutorial for getting started with the Alchemy framework",
- *   difficulty: "beginner",
- *   estimatedTime: 30
- * });
- *
- * @example
- * // Create an advanced tutorial with custom model and continue from existing messages
- * const advancedTutorial = await Tutorial("advanced-features", {
- *   path: "docs/tutorials/advanced-features",
- *   title: "Advanced Alchemy Features",
- *   prompt: "Create a tutorial covering advanced features of the Alchemy framework",
- *   difficulty: "advanced",
- *   estimatedTime: 60,
- *   model: {
- *     id: "claude-3-7-sonnet-latest",
- *     provider: "anthropic"
- *   },
- *   maxIterations: 5,
- *   messages: basicTutorial.messages
- * });
- */
-export const Tutorial = Resource(
-  "docs::Tutorial",
-  async function (
-    this: Context<Tutorial>,
-    id: string,
-    props: TutorialProps
-  ): Promise<Tutorial> {
-    const {
-      path: outFile,
-      title,
-      prompt,
-      difficulty = "beginner",
-      estimatedTime = 30,
-      model = {
-        id: "claude-3-5-sonnet-latest",
-        provider: "anthropic",
-      },
-      maxIterations = 3,
-      messages: initialMessages = [],
-    } = props;
-
-    console.log(`Tutorial: Starting creation of "${title}" (ID: ${id})`);
-
-    // Handle deletion phase
-    if (this.phase === "delete") {
-      console.log(`Tutorial: Deleting tutorial "${title}" (ID: ${id})`);
-      return this.destroy();
-    }
-
-    // System prompts
-    const docSystemPrompt = `You are a technical writer creating a comprehensive tutorial.
-    
-    Your task is to create a detailed tutorial about ${title} with difficulty level: ${difficulty} and estimated time to complete: ${estimatedTime} minutes.
-    
-    The tutorial should take users from zero knowledge to a working understanding of the tool or framework.
-    
-    The tutorial should be structured as follows:
-    
-    # ${title}
-    
-    ## Overview
-    
-    (Provide a brief introduction to the tool/framework, its purpose, and what users will learn in this tutorial)
-    
-    ## Prerequisites
-    
-    (List any prerequisites, tools, or knowledge required before starting the tutorial)
-    
-    ## Setup
-    
-    ### Installation
-    
-    \`\`\`bash
-    # Installation commands
-    \`\`\`
-    
-    ### Configuration
-    
-    (Explain any necessary configuration steps)
-    
-    \`\`\`bash
-    # Configuration commands or code
-    \`\`\`
-    
-    ## Step 1: [First Step Title]
-    
-    (Explain the first step in detail)
-    
-    \`\`\`bash
-    # Commands or code for the first step
-    \`\`\`
-    
-    (Explain what the code does and why it's important)
-    
-    ## Step 2: [Second Step Title]
-    
-    (Explain the second step in detail)
-    
-    \`\`\`bash
-    # Commands or code for the second step
-    \`\`\`
-    
-    (Explain what the code does and why it's important)
-    
-    ## Step 3: [Third Step Title]
-    
-    (Explain the third step in detail)
-    
-    \`\`\`bash
-    # Commands or code for the third step
-    \`\`\`
-    
-    (Explain what the code does and why it's important)
-    
-    ## Step 4: [Fourth Step Title]
-    
-    (Explain the fourth step in detail)
-    
-    \`\`\`bash
-    # Commands or code for the fourth step
-    \`\`\`
-    
-    (Explain what the code does and why it's important)
-    
-    ## Step 5: [Fifth Step Title]
-    
-    (Explain the fifth step in detail)
-    
-    \`\`\`bash
-    # Commands or code for the fifth step
-    \`\`\`
-    
-    (Explain what the code does and why it's important)
-    
-    ## Testing Your Work
-    
-    (Explain how to verify that everything is working correctly)
-    
-    \`\`\`bash
-    # Testing commands or code
-    \`\`\`
-    
-    ## Troubleshooting
-    
-    (Common issues users might encounter and how to resolve them)
-    
-    ## Next Steps
-    
-    (Suggest additional resources, advanced topics, or projects to try next)
-    
-    ## Additional Resources
-    
-    (Links to documentation, community resources, or related tutorials)`;
-
-    // Initial message if none provided
-    const startingMessages =
-      initialMessages.length > 0
-        ? initialMessages
-        : [
-            {
-              role: "user" as const,
-              content: `Create a comprehensive tutorial about ${title}.\n\n${prompt}\n\nDifficulty level: ${difficulty}\nEstimated time to complete: ${estimatedTime} minutes`,
-            },
-          ];
-
-    // Generate the initial tutorial
-    console.log(`Tutorial: Generating initial tutorial for "${title}"`);
-    let tutorial = await Document(`document`, {
-      title: title,
-      path: typeof outFile === "string" ? outFile : outFile.path,
-      model,
-      messages: startingMessages,
-      system: docSystemPrompt,
-    });
-    console.log(
-      `Tutorial: Initial tutorial generated for "${title}" (${tutorial.content.length} chars)`
-    );
-
-    // Review and improve the tutorial in a loop
-    let iteration = 0;
-    let approved = false;
-    let finalMessages = tutorial.messages;
-
-    while (!approved && iteration < maxIterations) {
-      iteration++;
-      console.log(
-        `Tutorial: Starting review iteration ${iteration}/${maxIterations} for "${title}"`
-      );
-
-      // Review the tutorial using system prompt
-      console.log(`Tutorial: Requesting review for iteration ${iteration}`);
-      const review = await Review(`review-${iteration}`, {
-        messages: finalMessages,
-        system:
-          await alchemy`Please review this tutorial for correctness, conciseness, logical flow, and accurateness. 
-        
-      The tutorial should:
-      1. Be technically accurate and free of errors
-      2. Be concise without over-explaining
-      3. Have a logical flow from basic to advanced concepts
-      4. Be accurate in its descriptions and instructions
-      5. Be appropriate for the ${difficulty} difficulty level
-      6. Be completable within approximately ${estimatedTime} minutes
-      
-      Provide specific feedback on areas that need improvement.`,
-        model,
-      });
-
-      // Extract the actual review content from the last assistant message
-      const reviewContent =
-        review.messages[review.messages.length - 1]?.content || "";
-      console.log(
-        `Tutorial: Review received for iteration ${iteration} (${reviewContent.length} chars)`
-      );
-
-      // Check if the tutorial is approved using system prompt
-      console.log(`Tutorial: Evaluating approval for iteration ${iteration}`);
-      const approvalResult = await Approve(`approval-${iteration}`, {
-        messages: review.messages,
-        system:
-          await alchemy`Based on the review above, determine if the tutorial meets the quality standards.
-      The tutorial should be approved if:
-      1. It is technically accurate and free of errors
-      2. It is concise without over-explaining
-      3. It has a logical flow from basic to advanced concepts
-      4. It is accurate in its descriptions and instructions
-      5. It is appropriate for the ${difficulty} difficulty level
-      6. It is completable within approximately ${estimatedTime} minutes
-      
-      Respond with APPROVED or DENIED, followed by your explanation.`,
-      });
-      console.log(
-        `Tutorial: Approval result for iteration ${iteration}: ${approvalResult.approved ? "APPROVED" : "DENIED"}`
-      );
-
-      // Log approval status
-      if (approvalResult.approved) {
-        console.log(
-          `Tutorial: Final tutorial approved: ${approvalResult.explanation}`
-        );
-        approved = true;
-        finalMessages = approvalResult.messages;
-      } else {
-        console.log(
-          `Tutorial: Tutorial needs improvement: ${approvalResult.explanation}`
-        );
-
-        // Add improvement request and generate an improved tutorial
-        const improvementMessages = [
-          ...approvalResult.messages,
-          {
-            role: "user" as const,
-            content: `Please create an improved version of the tutorial that addresses the issues identified in the review: ${approvalResult.explanation}`,
-          },
-        ];
-
-        console.log(
-          `Tutorial: Regenerating tutorial based on feedback for iteration ${iteration}`
-        );
-
-        tutorial = await Document(`tutorial-${iteration}`, {
-          title: title,
-          path: typeof outFile === "string" ? outFile : outFile.path,
-          model,
-          messages: improvementMessages,
-          system: docSystemPrompt,
-        });
-        console.log(
-          `Tutorial: Improved tutorial generated for iteration ${iteration} (${tutorial.content.length} chars)`
-        );
-
-        finalMessages = tutorial.messages;
-      }
-    }
-
-    if (!approved) {
-      console.log(
-        `Tutorial: Tutorial not approved after ${maxIterations} iterations. Using the best version available.`
-      );
-    }
-
-    console.log(`Tutorial: Completed creation of "${title}" (ID: ${id})`);
-
-    // Return the tutorial resource
-    return this({
-      ...props,
-      document: tutorial,
-      content: tutorial.content,
-      messages: finalMessages,
-      createdAt: Date.now(),
-      updatedAt: Date.now(),
-    });
-  }
-);