feat(summary): add AI-powered message summary feature

Add new ThunderAI summary functionality that generates concise summaries
for email messages using the existing ThunderAI infrastructure. Includes
a new content script that creates a summary pane in the message display,
associated CSS styling, and backend integration for AI summary generation.

The feature shows a loading indicator while generating the summary and
falls back to showing truncated message content if the AI generation fails.

messageDisplay/message-content-script.js

@@ -0,0 +1,105 @@
+async function showSummaryPane() {
+    // Create the summary pane element
+    const summaryPane = document.createElement("div");
+    summaryPane.className = "thunderai-summary-pane";
+
+    // Create the title element
+    const summaryTitle = document.createElement("div");
+    summaryTitle.className = "thunderai-summary-title";
+    summaryTitle.innerText = "ThunderAI Summary";
+
+    // Create a loading indicator
+    const loadingIndicator = document.createElement("div");
+    loadingIndicator.className = "thunderai-summary-content";
+    loadingIndicator.innerText = "Generating AI summary...";
+
+    // Create the content element (initially hidden)
+    const summaryContent = document.createElement("div");
+    summaryContent.className = "thunderai-summary-content";
+    summaryContent.style.display = 'none';
+
+    // Add title and loading indicator to the pane
+    summaryPane.appendChild(summaryTitle);
+    summaryPane.appendChild(loadingIndicator);
+    summaryPane.appendChild(summaryContent);
+
+    // Insert it as the very first element in the message
+    document.body.insertBefore(summaryPane, document.body.firstChild);
+
+    // Get the message content and generate summary
+    try {
+        const messageContent = getMessageContent();
+        const aiSummary = await generateAISummary(messageContent);
+
+        // Update the UI with the AI summary
+        loadingIndicator.style.display = 'none';
+        summaryContent.innerText = aiSummary;
+        summaryContent.style.display = 'block';
+    } catch (error) {
+        console.error("Error generating AI summary:", error);
+        loadingIndicator.innerText = "Failed to generate AI summary. Showing message preview instead.";
+        loadingIndicator.style.color = '#d70022';
+
+        // Fallback to showing truncated message content
+        const messageContent = getMessageContent();
+        loadingIndicator.innerText += "\n\n" + truncateMessageContent(messageContent);
+    }
+}
+
+function getMessageContent() {
+    // Get the main message content from the page
+    // This selects the main message body content
+    const messageBody = document.querySelector('.moz-text-flowed, .moz-text-plain, body');
+    if (messageBody) {
+        return messageBody.textContent || messageBody.innerText || '';
+    }
+
+    // Fallback: get the entire body content
+    return document.body.textContent || document.body.innerText || '';
+}
+
+function truncateMessageContent(content) {
+    // Clean up the content by removing excessive whitespace and newlines
+    const cleanedContent = content.replace(/\s+/g, ' ').trim();
+
+    // Truncate to a reasonable length for preview
+    const maxLength = 500;
+    if (cleanedContent.length <= maxLength) {
+        return cleanedContent;
+    }
+
+    return cleanedContent.substring(0, maxLength) + '...';
+}
+
+async function generateAISummary(messageContent) {
+    // Clean up the message content
+    const cleanedContent = messageContent.replace(/\s+/g, ' ').trim();
+
+    // Create a simple summary prompt
+    const summaryPrompt = `Please provide a concise summary of the following email message. The summary should be 3-5 sentences maximum and capture the main points:
+
+${cleanedContent}
+
+Summary:`;
+
+    // Request AI summary from the background script
+    return new Promise((resolve, reject) => {
+        // Send message to background script to get AI summary
+        browser.runtime.sendMessage({
+            command: "generate_summary",
+            content: cleanedContent,
+            prompt: summaryPrompt
+        }, (response) => {
+            if (response && response.summary) {
+                resolve(response.summary);
+            } else if (response && response.error) {
+                reject(new Error(response.error));
+            } else {
+                reject(new Error("Failed to get AI summary"));
+            }
+        });
+    });
+}
+
+// Call the function to show the pane
+showSummaryPane();

messageDisplay/message-content-styles.css

@@ -0,0 +1,20 @@
+.thunderai-summary-pane {
+  background-color: #f0f0f0;
+  color: #333;
+  font-weight: 400;
+  padding: 0.5rem;
+  margin-bottom: 1rem;
+  border-radius: 4px;
+  border: 1px solid #ddd;
+}
+
+.thunderai-summary-title {
+  font-weight: bold;
+  margin-bottom: 0.5rem;
+  color: #d70022;
+}
+
+.thunderai-summary-content {
+  font-size: 0.9rem;
+  line-height: 1.4;
+}

mzta-background.js

@@ -99,6 +99,13 @@ browser.composeScripts.register({
 // Register the message display script for all newly opened message tabs.
 messenger.messageDisplayScripts.register({
     js: [{ file: "js/mzta-compose-script.js" }],
+    css: [{ file: "messageDisplay/message-content-styles.css" }]
+});
+
+// Register our new ThunderAI summary script
+messenger.messageDisplayScripts.register({
+    js: [{ file: "messageDisplay/message-content-script.js" }],
+    css: [{ file: "messageDisplay/message-content-styles.css" }]
 });
 
 // Inject script and CSS in all already open message tabs.
@@ -114,6 +121,13 @@ for (let messageTab of messageTabs) {
         await browser.tabs.executeScript(messageTab.id, {
             file: "js/mzta-compose-script.js"
         })
+        // Inject our ThunderAI summary script
+        await browser.tabs.executeScript(messageTab.id, {
+            file: "messageDisplay/message-content-script.js"
+        })
+        await browser.tabs.insertCSS(messageTab.id, {
+            file: "messageDisplay/message-content-styles.css"
+        });
     } catch (error) {
         console.error("[ThunderAI] Error injecting message display script:", error);
         console.error("[ThunderAI] Message tab:", messageTab.url);
@@ -220,6 +234,31 @@ messenger.runtime.onMessage.addListener((message, sender, sendResponse) => {
     // handler function.
     if (message && message.hasOwnProperty("command")){
         switch (message.command) {
+            case 'generate_summary':
+                async function _generate_summary(message) {
+                    try {
+                        // Get user preferences for AI connection
+                        let prefs = await browser.storage.sync.get({
+                            connection_type: prefs_default.connection_type,
+                            chatgpt_model: prefs_default.chatgpt_model,
+                            chatgpt_api_key: prefs_default.chatgpt_api_key,
+                            do_debug: prefs_default.do_debug
+                        });
+
+                        // Use the existing ThunderAI infrastructure
+                        const summary = await generateAISummaryUsingThunderAIInfrastructure(
+                            message.content,
+                            message.prompt,
+                            prefs
+                        );
+
+                        return { summary: summary };
+                    } catch (error) {
+                        console.error("[ThunderAI] Error generating summary:", error);
+                        return { error: "Failed to generate AI summary: " + error.message };
+                    }
+                }
+                return _generate_summary(message);
             // case 'chatgpt_open':
             //         openChatGPT(message.prompt,message.action,message.tabId);
             //         return true;
@@ -1177,3 +1216,48 @@ try {
     taLog.log("Using browser.messages.onNewMailReceived.addListener with one agrument for Thunderbird 115.");
     browser.messages.onNewMailReceived.addListener(newEmailListener);
 }
+
+/**
+ * AI summary generation function using ThunderAI infrastructure
+ */
+async function generateAISummaryUsingThunderAIInfrastructure(content, prompt, prefs) {
+    // Import the special command class
+    const { mzta_specialCommand } = await import('./js/mzta-special-commands.js');
+
+    // Determine which LLM to use based on user preferences
+    const llmType = getConnectionType(prefs.connection_type, {}, '');
+
+    // Create a special command instance
+    const summaryCommand = new mzta_specialCommand({
+        prompt: prompt,
+        llm: llmType,
+        custom_model: prefs.chatgpt_model,
+        do_debug: prefs.do_debug
+    });
+
+    // Initialize the worker
+    await summaryCommand.initWorker();
+
+    // Send the prompt and get the AI response
+    const aiResponse = await summaryCommand.sendPrompt();
+
+    // Clean up the response - extract just the summary content
+    const cleanedResponse = cleanAISummaryResponse(aiResponse);
+
+    return cleanedResponse;
+}
+
+/**
+ * Helper function to clean AI response
+ */
+function cleanAISummaryResponse(response) {
+    // Remove any markdown formatting or code blocks
+    let cleaned = response.replace(/```[\s\S]*?```/g, '');
+    cleaned = cleaned.replace(/[\*#_~`]/g, '');
+    cleaned = cleaned.replace(/\s+/g, ' ').trim();
+
+    // Remove any "Summary:" prefixes that the AI might add
+    cleaned = cleaned.replace(/^Summary:\s*/i, '');
+
+    return cleaned;
+}