Initial work on Vision support in text conversations and Image Editing

This is a huge patch which does some major refactoring like:

- renaming "Image Generation" to "Image Creation" in most places,
  to better match its new command (`!bai image create`)

- relocating image creation command (`!bai image` -> `!bai image create`),
  so it wouldn't conflict with the new image editing command (`!bai image edit`)

- introducing a new image editing command (`!bai image edit`), which
  is meant to work only with the OpenAI provider, but doesn't fully work yet
  due to https://github.com/64bit/async-openai/issues/364, though a next patch will fix it

- adding support for reading images off of Matrix conversations and forwarding them to
  text conversations. Works for OpenAI, but not for Anthropic yet
  (requires custom patches) and not for OpenAI-Compat (no support for
  images there)

- relocating some utils around (base64, mime)

src/agent/provider/anthropic/controller.rs

@@ -7,12 +7,14 @@ use anthropic::types::ContentBlock;
 use super::super::ControllerTrait;
 use crate::agent::AgentPurpose;
 use crate::agent::provider::entity::{
-    ImageGenerationResult, PingResult, TextGenerationParams, TextGenerationResult,
-    TextToSpeechParams, TextToSpeechResult,
+    ImageGenerationResult, ImageEditResult, ImageSource, PingResult, TextGenerationParams,
+    TextGenerationResult, TextToSpeechParams, TextToSpeechResult,
+};
+use crate::agent::provider::{
+    ImageGenerationParams, ImageEditParams, SpeechToTextParams, SpeechToTextResult,
 };
-use crate::agent::provider::{ImageGenerationParams, SpeechToTextParams, SpeechToTextResult};
 use crate::conversation::llm::{
-    Author as LLMAuthor, Conversation as LLMConversation, Message as LLMMessage,
+    Author as LLMAuthor, Conversation as LLMConversation, Message as LLMMessage, MessageContent as LLMMessageContent,
     shorten_messages_list_to_context_size,
 };
 use crate::strings;
@@ -69,7 +71,7 @@ impl ControllerTrait for Controller {
 
         let messages = vec![LLMMessage {
             author: LLMAuthor::User,
-            message_text: "Hello!".to_string(),
+            content: LLMMessageContent::Text("Hello!".to_string()),
             timestamp: chrono::Utc::now(),
         }];
 
@@ -106,7 +108,7 @@ impl ControllerTrait for Controller {
         } else {
             Some(LLMMessage {
                 author: LLMAuthor::Prompt,
-                message_text: prompt_text,
+                content: LLMMessageContent::Text(prompt_text),
                 timestamp: chrono::Utc::now(),
             })
         };
@@ -145,7 +147,9 @@ impl ControllerTrait for Controller {
             .unwrap_or(text_generation_config.temperature);
 
         if let Some(prompt_message) = prompt_message {
-            request.system = prompt_message.message_text;
+            if let LLMMessageContent::Text(text) = &prompt_message.content {
+                request.system = text.clone();
+            }
         }
 
         request.model = text_generation_config.model_id.clone();
@@ -213,6 +217,15 @@ impl ControllerTrait for Controller {
         Err(anyhow::anyhow!("Image generation not supported"))
     }
 
+    async fn create_image_edit(
+        &self,
+        _prompt: &str,
+        _images: Vec<ImageSource>,
+        _params: ImageEditParams,
+    ) -> anyhow::Result<ImageEditResult> {
+        Err(anyhow::anyhow!("Image editing is not supported"))
+    }
+
     async fn text_to_speech(
         &self,
         _input: &str,

src/agent/provider/anthropic/utils.rs

@@ -1,6 +1,6 @@
 use anthropic::types::{ContentBlock, Message, MessagesRequest, MessagesRequestBuilder, Role};
 
-use crate::conversation::llm::{Author as LLMAuthor, Message as LLMMessage};
+use crate::conversation::llm::{Author as LLMAuthor, Message as LLMMessage, MessageContent as LLMMessageContent};
 
 pub(super) fn create_anthropic_message_request(llm_messages: Vec<LLMMessage>) -> MessagesRequest {
     let mut messages = vec![];
@@ -14,13 +14,28 @@ pub(super) fn create_anthropic_message_request(llm_messages: Vec<LLMMessage>) ->
             }
         };
 
-        let content = vec![ContentBlock::Text {
-            text: message.message_text,
-        }];
+        let content = match &message.content {
+            LLMMessageContent::Text(text) => Some(vec![ContentBlock::Text { text: text.clone() }]),
+            LLMMessageContent::Image(_image_details) => {
+                // This cannot be implemented yet, because the Anthropic library does not support it.
+                // The code below requires the library to be forked and changed a bit.
+                // vec![ContentBlock::Image {
+                //     source: ImageSource {
+                //         r#type: "base64".to_string(),
+                //         media_type: mime_type.to_string(),
+                //         data: crate::utils::base64::base64_encode(image_details.data),
+                //     },
+                // }]
+                tracing::warn!("Image content is not supported by the Anthropic library yet. Skipping it.");
+                None
+            }
+        };
 
-        let message = Message { role, content };
+        if let Some(content) = content {
+            let message = Message { role, content };
 
-        messages.push(message);
+            messages.push(message);
+        }
     }
 
     MessagesRequestBuilder::default()