valknar/llmx
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: context compaction (#3446)

Browse Source 
## Compact feature:
1. Stops the model when the context window become too large
2. Add a user turn, asking for the model to summarize
3. Build a bridge that contains all the previous user message + the
summary. Rendered from a template
4. Start sampling again from a clean conversation with only that bridge
This commit is contained in:
jif-oai
2025-09-12 13:07:10 -07:00
committed by GitHub
parent d4848e558b
commit ea225df22e
14 changed files with 1243 additions and 326 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
codex-rs/core/src/config.rs
View File

@@ -55,6 +55,9 @@ pub struct Config {
/// Maximum number of output tokens.
pub model_max_output_tokens: Option<u64>,
/// Token usage threshold triggering auto-compaction of conversation history.
pub model_auto_compact_token_limit: Option<i64>,
/// Key into the model_providers map that specifies which provider to use.
pub model_provider_id: String,
@@ -519,6 +522,9 @@ pub struct ConfigToml {
/// Maximum number of output tokens.
pub model_max_output_tokens: Option<u64>,
/// Token usage threshold triggering auto-compaction of conversation history.
pub model_auto_compact_token_limit: Option<i64>,
/// Default approval policy for executing commands.
pub approval_policy: Option<AskForApproval>,
@@ -877,6 +883,11 @@ impl Config {
.as_ref()
.map(|info| info.max_output_tokens)
});
let model_auto_compact_token_limit = cfg.model_auto_compact_token_limit.or_else(|| {
openai_model_info
.as_ref()
.and_then(|info| info.auto_compact_token_limit)
});
let experimental_resume = cfg.experimental_resume;
@@ -896,6 +907,7 @@ impl Config {
model_family,
model_context_window,
model_max_output_tokens,
model_auto_compact_token_limit,
model_provider_id,
model_provider,
cwd: resolved_cwd,
@@ -1430,6 +1442,7 @@ model_verbosity = "high"
model_family: find_family_for_model("o3").expect("known model slug"),
model_context_window: Some(200_000),
model_max_output_tokens: Some(100_000),
model_auto_compact_token_limit: None,
model_provider_id: "openai".to_string(),
model_provider: fixture.openai_provider.clone(),
approval_policy: AskForApproval::Never,
@@ -1486,6 +1499,7 @@ model_verbosity = "high"
model_family: find_family_for_model("gpt-3.5-turbo").expect("known model slug"),
model_context_window: Some(16_385),
model_max_output_tokens: Some(4_096),
model_auto_compact_token_limit: None,
model_provider_id: "openai-chat-completions".to_string(),
model_provider: fixture.openai_chat_completions_provider.clone(),
approval_policy: AskForApproval::UnlessTrusted,
@@ -1557,6 +1571,7 @@ model_verbosity = "high"
model_family: find_family_for_model("o3").expect("known model slug"),
model_context_window: Some(200_000),
model_max_output_tokens: Some(100_000),
model_auto_compact_token_limit: None,
model_provider_id: "openai".to_string(),
model_provider: fixture.openai_provider.clone(),
approval_policy: AskForApproval::OnFailure,
@@ -1614,6 +1629,7 @@ model_verbosity = "high"
model_family: find_family_for_model("gpt-5").expect("known model slug"),
model_context_window: Some(272_000),
model_max_output_tokens: Some(128_000),
model_auto_compact_token_limit: None,
model_provider_id: "openai".to_string(),
model_provider: fixture.openai_provider.clone(),
approval_policy: AskForApproval::OnFailure,