public/galaxy
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

v1.4.0: Auto-compact streaming, Bedrock summarization support, subagent orchestration, and Galaxy rebrand continuation

Browse Source 
Major features:
- Auto-compact: triggers conversation summarization when context window >= 85%,
  compacts Bedrock message history to a summary pair, and tracks live context tokens
- Bedrock summarization: plumbs `is_summarization` flag through translator/client/response
  pipeline, handles SummarizeConversation input type, and marks `summarized` in metadata
- Session restore: rebuilds bedrock_message_history from persisted task messages via
  newly-public `convert_proto_message`, preventing empty history on reconnect
- Subagent orchestration: adds SubagentQuestion/Answer/CompletionSummary event types,
  parent-child question routing with depth limits, retry counting, and drain methods
- Summarization UI: inline SummarizationView in AI blocks with progress/finished states

Refactors:
- Rename WarpTheme → GalaxyTheme across ~100 files (rebrand continuation)
- Rename warp_home_config_dir → galaxy_home_config_dir and related path functions
- Predefined rules: replace "System Defined Rule #N" with descriptive names
  (e.g. "Correctness Over Speed", "Never Guess") and add lookup helpers
- Usage view: replace cumulative input/output token display with live context tokens,
  cache hit rate calculation, and separate cache read/write stats
- Telemetry: remove verbose doc comments, simplify trait definitions
- Facts view: simplify delete permission check (always allow local deletion)
- Remove warp_managed_paths_watcher.rs (dead code)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Ryan Ward
2026-05-21 11:59:37 -05:00
co-authored by Claude Opus 4.6
parent eaa2ddc75e
commit 6f54e2cb30
229 changed files with 2506 additions and 2634 deletions
Download Patch File Download Diff File Expand all files Collapse all files
app/src/ai/blocklist/controller.rs
+126 -5
View File
@@ -1995,6 +1995,9 @@ impl BlocklistAIController {
request_params.parent_agent_id = parent_agent_id;
request_params.agent_name = agent_name;
request_params.bedrock_message_history = bedrock_history;
request_params.is_summarization = request_input
.all_inputs()
.any(|input| matches!(input, AIAgentInput::SummarizeConversation { .. }));
let server_conversation_token_for_identifiers =
conversation_data.server_conversation_token.clone();
@@ -2020,9 +2023,13 @@ impl BlocklistAIController {
let input_contains_user_query = request_input
.all_inputs()
.any(|input| input.is_user_query());
let input_is_summarization = request_input
.all_inputs()
.any(|input| matches!(input, AIAgentInput::SummarizeConversation { .. }));
ctx.subscribe_to_model(&response_stream, move |me, event, ctx| {
me.handle_response_stream_event(
input_contains_user_query,
input_is_summarization,
event,
&response_stream_clone,
ctx,
@@ -2212,6 +2219,7 @@ impl BlocklistAIController {
fn handle_response_stream_event(
&mut self,
did_input_contain_user_query: bool,
is_summarization_request: bool,
event: &ResponseStreamEvent,
response_stream: &ModelHandle<ResponseStream>,
ctx: &mut ModelContext<Self>,
@@ -2319,11 +2327,88 @@ impl BlocklistAIController {
let history_model = BlocklistAIHistoryModel::handle(ctx);
history_model.update(ctx, |history_model, _| {
if let Some(conversation) = history_model.conversation_mut(&conversation_id) {
*conversation.bedrock_message_history_mut() = new_history;
log::info!(
"[bedrock] Updated conversation bedrock history: {} messages",
conversation.bedrock_message_history().len()
);
// If this was a summarization request, compact the
// history to just the summary instead of keeping
// the full message list. This is what actually
// frees up context window space.
let is_summarization = is_summarization_request;
if is_summarization {
// Extract the assistant's summary from the last
// message in the history (the response).
let summary_text = new_history
.iter()
.rev()
.find_map(|msg| {
use crate::ai::bedrock::convert::{
MessageContent, MessageRole,
};
if msg.role == MessageRole::Assistant {
if let MessageContent::Text(text) =
&msg.content
{
Some(text.clone())
} else {
None
}
} else {
None
}
});
if let Some(summary) = summary_text {
use crate::ai::bedrock::convert::{
ConversationMessage, MessageContent,
MessageRole,
};
let assistant_reply = "Understood. I have the context from our previous conversation. How can I help you next?";
let user_msg = format!(
"Here is a summary of our conversation so far:\n\n{summary}"
);
let compacted = vec![
ConversationMessage {
role: MessageRole::User,
content: MessageContent::Text(user_msg.clone()),
},
ConversationMessage {
role: MessageRole::Assistant,
content: MessageContent::Text(
assistant_reply.to_string()
),
},
];
log::info!(
"[bedrock] Compacted conversation history from {} messages to {} (summary)",
new_history.len(),
compacted.len()
);
*conversation.bedrock_message_history_mut() =
compacted;
// Estimate new context size from the compacted content.
// ~4 chars per token is a reasonable approximation.
let estimated_tokens = ((user_msg.len() + assistant_reply.len()) / 4) as u32;
let max_context = crate::ai::bedrock::response_translator::context_window_for_model("claude-opus-4-6-20250514[1m]");
let new_usage = estimated_tokens as f32 / max_context as f32;
conversation.set_context_window_usage(new_usage);
conversation.set_current_context_tokens(estimated_tokens);
log::info!(
"[bedrock] Post-compact context estimate: ~{} tokens ({:.1}% of context window)",
estimated_tokens,
new_usage * 100.0
);
} else {
*conversation.bedrock_message_history_mut() =
new_history;
}
} else {
*conversation.bedrock_message_history_mut() =
new_history;
log::info!(
"[bedrock] Updated conversation bedrock history: {} messages",
conversation.bedrock_message_history().len()
);
}
}
});
}
@@ -2800,6 +2885,42 @@ impl BlocklistAIController {
});
ctx.emit(BlocklistAIControllerEvent::FreeTierLimitCheckTriggered);
}
// Auto-compact: trigger summarization when context window usage >= 85%.
let should_auto_compact = {
let history_model = BlocklistAIHistoryModel::as_ref(ctx);
history_model
.conversation(&conversation_id)
.is_some_and(|conversation| {
let is_summarization_request = conversation
.latest_exchange()
.is_some_and(|exchange| {
exchange
.input
.iter()
.any(|i| matches!(i, AIAgentInput::SummarizeConversation { .. }))
});
conversation.context_window_usage() >= 0.85
&& !conversation.has_pending_auto_compact()
&& !is_summarization_request
})
};
if should_auto_compact {
log::info!(
"[auto-compact] Context window usage >= 85% for conversation {:?}, triggering summarization",
conversation_id
);
BlocklistAIHistoryModel::handle(ctx).update(ctx, |history_model, _| {
if let Some(conversation) = history_model.conversation_mut(&conversation_id) {
conversation.set_has_pending_auto_compact(true);
}
});
self.send_slash_command_request(
SlashCommandRequest::Summarize { prompt: None },
ctx,
);
}
}
}