feat: enhance recharts and report agents with new configurations and processors

Author: ssdeanx

src/mastra/agents/recharts.ts

@@ -1,6 +1,8 @@
 import { Agent } from '@mastra/core/agent'
 import { InternalSpans } from '@mastra/core/ai-tracing'
 import { googleAI, googleAIFlashLite, pgMemory } from '../config'
+import { grokAI } from '../config/copilot';
+import { BatchPartsProcessor, UnicodeNormalizer } from '@mastra/core/processors/processors';
 
 export const rechartsTypes = new Agent({
     id: 'recharts-types',
@@ -18,18 +20,18 @@ export const rechartsTypes = new Agent({
             ## Style Guide
 
     `,
-            providerOptions: {
-                google: {
-                    thinkingConfig: {
-                        thinkingLevel: 'medium',
-                        includeThoughts: true,
-                        thinkingBudget: -1,
-                    }
-                }
-            }
+          ///  providerOptions: {
+            //    google: {
+            //        thinkingConfig: {
+            //            thinkingLevel: 'medium',
+            //            includeThoughts: true,
+            //            thinkingBudget: -1,
+            //        }
+        //        }
+        //    }
         }
     },
-    model: googleAIFlashLite,
+    model: grokAI,
     memory: pgMemory,
     options: { tracingPolicy: { internal: InternalSpans.AGENT } },
 })
@@ -53,7 +55,11 @@ export const rechartsIssueLabeler = new Agent({
                         thinkingLevel: 'medium',
                         includeThoughts: true,
                         thinkingBudget: -1,
-                    }
+                    },
+                    mediaResolution: 'MEDIA_RESOLUTION_MEDIUM',
+                    maxOutputTokens: 64000,
+                    temperature: 0.2,
+                    topP: 1.0,
                 }
             }
         }
@@ -88,7 +94,11 @@ export const rechartsLinkChecker = new Agent({
                         thinkingLevel: 'low',
                         includeThoughts: true,
                         thinkingBudget: -1,
-                    }
+                    },
+                    mediaResolution: 'MEDIA_RESOLUTION_MEDIUM',
+                    maxOutputTokens: 64000,
+                    temperature: 0.2,
+                    topP: 1.0
                 }
             }
         }
@@ -124,7 +134,11 @@ export const chartdesigner = new Agent({
                         thinkingLevel: 'medium',
                         includeThoughts: true,
                         thinkingBudget: -1,
-                    }
+                    },
+                    mediaResolution: 'MEDIA_RESOLUTION_MEDIUM',
+                    maxOutputTokens: 64000,
+                    temperature: 0.2,
+                    topP: 1.0
                 }
             }
         }
@@ -161,7 +175,11 @@ export const rechartsMaster = new Agent({
                         thinkingLevel: 'high',
                         includeThoughts: true,
                         thinkingBudget: -1,
-                    }
+                    },
+                    mediaResolution: 'MEDIA_RESOLUTION_MEDIUM',
+                    maxOutputTokens: 64000,
+                    temperature: 0.2,
+                    topP: 1.0
                 }
             }
         }
@@ -172,4 +190,26 @@ export const rechartsMaster = new Agent({
 
     },
     options: { tracingPolicy: { internal: InternalSpans.AGENT } },
+    scorers: {
+        // Add any custom scorers here
+    },
+    workflows: {
+        // Add any custom workflows here
+    },
+    inputProcessors: [
+        new UnicodeNormalizer({
+            stripControlChars: false,
+            collapseWhitespace: true,
+            preserveEmojis: true,
+            trim: true,
+        }),
+      ],
+      outputProcessors: [
+        new BatchPartsProcessor({
+          batchSize: 15,
+          maxWaitTime: 50,
+          emitOnNonText: true,
+        }),
+      ],
+    maxRetries: 5
 })

src/mastra/agents/reportAgent.ts

@@ -95,7 +95,11 @@ export const reportAgent = new Agent({
                         thinkingLevel: 'medium',
                         includeThoughts: true,
                         thinkingBudget: -1,
-                    }
+                    },
+                    mediaResolution: 'MEDIA_RESOLUTION_MEDIUM',
+                    maxOutputTokens: 64000,
+                    temperature: 0.2,
+                    topP: 1.0
                 }
             }
         }

src/mastra/config/copilot.ts

@@ -3,6 +3,7 @@ import { createGitHubCopilotOpenAICompatible } from '@opeoginni/github-copilot-o
 const githubCopilot = createGitHubCopilotOpenAICompatible({
   baseURL: 'https://api.githubcopilot.com',
   name: 'githubcopilot',
+  apiKey: process.env.COPILOT_TOKEN,
   headers: {
     Authorization: `Bearer ${process.env.COPILOT_TOKEN}`,
     "Copilot-Integration-Id": "vscode-chat", // These configs must be provided

src/mastra/config/google.ts

@@ -108,11 +108,23 @@ export const googleNanoBanana = google('gemini-3-pro-image-preview')
 //  systemInstruction: 'This is model has cache',
 //  displayName: 'cacheContent',
 //});
-
+/*
+   * imageGen: Gemini Imagen 4.0 Generate model for standard image generation
+    * When to use: This model is suitable for standard image generation tasks, offering a balance between quality and computational efficiency. It is ideal for applications ranging from simple illustrations to basic visualizations.
+    * Why use: Choose this model when you need to generate images quickly and efficiently, balancing the need for decent quality with the desire to keep costs down.
+*/
 export const imageGen = google.image('imagen-4.0-generate-001');
-
+/*
+ * imageUltra: Gemini Imagen 4.0 Ultra model for high-resolution image generation
+    * When to use: This model is ideal for applications that require ultra-high-resolution images with exceptional detail and quality, such as professional design work, marketing materials, and high-end visual content creation.
+    * Why use: Choose this model when image quality is a top priority and you need the best possible resolution and detail for your visual assets, despite the higher computational costs involved.
+*/
 export const imageUltra = google.image('imagen-4.0-ultra-generate-001');
-
+/*
+ * imageFast: Gemini Imagen 4.0 Fast model for rapid image generation
+    * When to use: This model is designed for scenarios where quick turnaround times are essential, such as real-time applications, rapid prototyping, or situations where speed is prioritized over ultra-high resolution.
+    * Why use: Choose this model when you need to generate images swiftly without significantly compromising on quality, making it ideal for dynamic content creation and interactive applications.
+*/
 export const imageFast = google.image('imagen-4.0-fast-generate-001');
 
 /*
@@ -132,4 +144,60 @@ export const googleWithMemory = withSupermemory(googleAI,"mastra", {
    verbose: true,
    mode: "full",
    addMemory: "always"
+});
+/*
+   * googleWithMemoryLite: Google Generative AI Flash Lite model integrated with Supermemory for enhanced conversational capabilities
+         * When to use: This configuration is ideal for budget-conscious applications that still require context-aware interactions, such as lightweight chatbots and virtual assistants. It leverages Supermemory to provide relevant historical context in conversations while minimizing costs.
+         * Why use: Utilize this setup when you want to enhance user interactions by incorporating past conversations and relevant information while keeping costs low.
+         * Example
+         * use cases:
+            * A budget-friendly customer support chatbot that remembers previous interactions with users to provide tailored assistance.
+            * A lightweight virtual assistant that maintains context across multiple user requests for more coherent interactions.
+            * An educational tutor bot that recalls past lessons and user progress to adapt its teaching approach on a budget.
+            * A travel planning assistant that considers past queries about destinations and preferences to suggest personalized itineraries without incurring high costs.
+            * A medical consultation tool that takes into account patient history and previous consultations to offer more informed advice while being cost-effective.
+         * Configuration details: This setup uses the 'mastra' container tag for memory search, includes a conversation ID for grouping messages, and operates in 'full' mode to maximize context retrieval. It is configured to always add relevant memories to the prompts while utilizing the cost-effective Flash Lite model.
+*/
+export const googleWithMemoryLite = withSupermemory(googleAIFlashLite,"mastra", {
+   conversationId: "mastra-conversation",
+   verbose: true,
+   mode: "full",
+   addMemory: "always"
+});
+/*
+   * superGoogle: Google Generative AI model integrated with Supermemory for enhanced conversational capabilities
+         * When to use: This configuration is ideal for applications that require context-aware interactions, such as chatbots, virtual assistants, and customer support systems. It leverages Supermemory to provide relevant historical context in conversations.
+         * Why use: Utilize this setup when you want to enhance user interactions by incorporating past conversations and relevant information, leading to more personalized and accurate responses.
+         * Example use cases:
+            * A customer support chatbot that remembers previous interactions with users to provide tailored assistance.
+            * A virtual assistant that maintains context across multiple user requests for more coherent interactions.
+            * An educational tutor bot that recalls past lessons and user progress to adapt its teaching approach.
+            * A travel planning assistant that considers past queries about destinations and preferences to suggest personalized itineraries.
+            * A medical consultation tool that takes into account patient history and previous consultations to offer more informed advice.
+         * Configuration details: This setup uses the 'mastra' container tag for memory search, includes a conversation ID for grouping messages, and operates in 'full' mode to maximize context retrieval. It is configured to always add relevant memories to the prompts.
+*/
+export const superGoogle = withSupermemory(googleAI,"mastra", {
+   conversationId: "mastra-conversation",
+   verbose: true,
+   mode: "full",
+   addMemory: "always"
+});
+/*
+   * superGoogleLite: Google Generative AI Flash Lite model integrated with Supermemory for enhanced conversational capabilities
+         * When to use: This configuration is ideal for budget-conscious applications that still require context-aware interactions, such as lightweight chatbots and virtual assistants. It leverages Supermemory to provide relevant historical context in conversations while minimizing costs.
+         * Why use: Utilize this setup when you want to enhance user interactions by incorporating past conversations and relevant information while keeping costs low.
+         * Example
+         * use cases:
+            * A budget-friendly customer support chatbot that remembers previous interactions with users to provide tailored assistance.
+            * A lightweight virtual assistant that maintains context across multiple user requests for more coherent interactions.
+            * An educational tutor bot that recalls past lessons and user progress to adapt its teaching approach on a budget.
+            * A travel planning assistant that considers past queries about destinations and preferences to suggest personalized itineraries without incurring high costs.
+            * A medical consultation tool that takes into account patient history and previous consultations to offer more informed advice while being cost-effective.
+         * Configuration details: This setup uses the 'mastra' container tag for memory search, includes a conversation ID for grouping messages, and operates in 'full' mode to maximize context retrieval. It is configured to always add relevant memories to the prompts while utilizing the cost-effective Flash Lite model.
+*/
+export const superGoogleLite = withSupermemory(googleAIFlashLite,"mastra", {
+   conversationId: "mastra-conversation",
+   verbose: true,
+   mode: "full",
+   addMemory: "always"
 });