mirror of
https://github.com/fawney19/Aether.git
synced 2026-10-05 00:47:48 +08:00
Merge branch 'main' of https://github.com/zhefox/Aether
This commit is contained in:
@@ -410,6 +410,7 @@ enum ClaudeOpenBlock {
|
||||
struct ClaudeClientToolState {
|
||||
call_id: String,
|
||||
name: String,
|
||||
buffered_arguments: String,
|
||||
}
|
||||
|
||||
#[derive(Default)]
|
||||
@@ -460,18 +461,47 @@ impl ClaudeClientEmitter {
|
||||
let Some(open_block) = self.open_block.take() else {
|
||||
return Ok(Vec::new());
|
||||
};
|
||||
let mut out = Vec::new();
|
||||
let block_index = match open_block {
|
||||
ClaudeOpenBlock::Text { block_index } => block_index,
|
||||
ClaudeOpenBlock::Thinking { block_index } => block_index,
|
||||
ClaudeOpenBlock::Tool { block_index, .. } => block_index,
|
||||
ClaudeOpenBlock::Tool {
|
||||
tool_index,
|
||||
block_index,
|
||||
} => {
|
||||
if let Some(state) = self.tool_states.get_mut(&tool_index) {
|
||||
if state.name == "Read" && !state.buffered_arguments.is_empty() {
|
||||
let arguments = remove_empty_pages_from_tool_arguments(
|
||||
&state.name,
|
||||
&state.buffered_arguments,
|
||||
);
|
||||
state.buffered_arguments.clear();
|
||||
if !arguments.is_empty() {
|
||||
out.extend(encode_json_sse(
|
||||
Some("content_block_delta"),
|
||||
&json!({
|
||||
"type": "content_block_delta",
|
||||
"index": block_index,
|
||||
"delta": {
|
||||
"type": "input_json_delta",
|
||||
"partial_json": arguments,
|
||||
}
|
||||
}),
|
||||
)?);
|
||||
}
|
||||
}
|
||||
}
|
||||
block_index
|
||||
}
|
||||
};
|
||||
encode_json_sse(
|
||||
out.extend(encode_json_sse(
|
||||
Some("content_block_stop"),
|
||||
&json!({
|
||||
"type": "content_block_stop",
|
||||
"index": block_index,
|
||||
}),
|
||||
)
|
||||
)?);
|
||||
Ok(out)
|
||||
}
|
||||
|
||||
fn ensure_text_block(&mut self) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
@@ -688,23 +718,33 @@ impl ClaudeClientEmitter {
|
||||
Ok(out)
|
||||
}
|
||||
CanonicalStreamEvent::ToolCallArgumentsDelta { index, arguments } => {
|
||||
let arguments = remove_empty_pages_from_tool_arguments(&arguments);
|
||||
let (call_id, name) = {
|
||||
let state = self.tool_states.entry(index).or_default();
|
||||
let call_id = if state.call_id.is_empty() {
|
||||
format!("tool_{index}")
|
||||
} else {
|
||||
state.call_id.clone()
|
||||
};
|
||||
let name = if state.name.is_empty() {
|
||||
"unknown".to_string()
|
||||
} else {
|
||||
state.name.clone()
|
||||
};
|
||||
(call_id, name)
|
||||
};
|
||||
if arguments.is_empty() {
|
||||
return Ok(Vec::new());
|
||||
}
|
||||
let mut out = self.ensure_started()?;
|
||||
let state = self.tool_states.entry(index).or_default();
|
||||
let call_id = if state.call_id.is_empty() {
|
||||
format!("tool_{index}")
|
||||
} else {
|
||||
state.call_id.clone()
|
||||
};
|
||||
let name = if state.name.is_empty() {
|
||||
"unknown".to_string()
|
||||
} else {
|
||||
state.name.clone()
|
||||
};
|
||||
out.extend(self.ensure_tool_block(index, &call_id, &name)?);
|
||||
if name == "Read" {
|
||||
self.tool_states
|
||||
.entry(index)
|
||||
.or_default()
|
||||
.buffered_arguments
|
||||
.push_str(&arguments);
|
||||
return Ok(out);
|
||||
}
|
||||
let block_index = match self.open_block {
|
||||
Some(ClaudeOpenBlock::Tool { block_index, .. }) => block_index,
|
||||
_ => return Ok(out),
|
||||
@@ -1230,7 +1270,7 @@ mod tests {
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn claude_client_emitter_removes_empty_pages_from_tool_arguments() {
|
||||
fn claude_client_emitter_removes_empty_pages_from_read_tool_arguments() {
|
||||
let mut emitter = ClaudeClientEmitter::default();
|
||||
let mut bytes = emitter
|
||||
.emit(CanonicalStreamFrame {
|
||||
@@ -1257,11 +1297,60 @@ mod tests {
|
||||
.expect("tool delta should encode"),
|
||||
);
|
||||
|
||||
let pending_sse = String::from_utf8(bytes.clone()).expect("sse should be utf8");
|
||||
assert!(!pending_sse.contains("\\\"pages\\\":\\\"\\\""));
|
||||
|
||||
bytes.extend(
|
||||
emitter
|
||||
.emit(CanonicalStreamFrame {
|
||||
id: "msg_123".to_string(),
|
||||
model: "claude-sonnet-4-5".to_string(),
|
||||
event: CanonicalStreamEvent::Finish {
|
||||
finish_reason: Some("tool_calls".to_string()),
|
||||
usage: None,
|
||||
},
|
||||
})
|
||||
.expect("finish should close read tool block"),
|
||||
);
|
||||
|
||||
let sse = String::from_utf8(bytes).expect("sse should be utf8");
|
||||
assert!(sse.contains("\"partial_json\":\"{\\\"file_path\\\":\\\"/tmp/a.txt\\\",\\\"offset\\\":1,\\\"limit\\\":20}\""));
|
||||
assert!(!sse.contains("\\\"pages\\\":\\\"\\\""));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn claude_client_emitter_preserves_empty_pages_for_other_tool_arguments() {
|
||||
let mut emitter = ClaudeClientEmitter::default();
|
||||
let mut bytes = emitter
|
||||
.emit(CanonicalStreamFrame {
|
||||
id: "msg_123".to_string(),
|
||||
model: "claude-sonnet-4-5".to_string(),
|
||||
event: CanonicalStreamEvent::ToolCallStart {
|
||||
index: 0,
|
||||
call_id: "toolu_search".to_string(),
|
||||
name: "Search".to_string(),
|
||||
},
|
||||
})
|
||||
.expect("tool start should encode");
|
||||
bytes.extend(
|
||||
emitter
|
||||
.emit(CanonicalStreamFrame {
|
||||
id: "msg_123".to_string(),
|
||||
model: "claude-sonnet-4-5".to_string(),
|
||||
event: CanonicalStreamEvent::ToolCallArgumentsDelta {
|
||||
index: 0,
|
||||
arguments: r#"{"query":"","pages":""}"#.to_string(),
|
||||
},
|
||||
})
|
||||
.expect("tool delta should encode"),
|
||||
);
|
||||
|
||||
let sse = String::from_utf8(bytes).expect("sse should be utf8");
|
||||
assert!(
|
||||
sse.contains("\"partial_json\":\"{\\\"query\\\":\\\"\\\",\\\"pages\\\":\\\"\\\"}\"")
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn claude_client_emitter_injects_default_usage_into_finish_events() {
|
||||
let mut emitter = ClaudeClientEmitter::default();
|
||||
|
||||
@@ -29,6 +29,7 @@ struct OpenAIResponsesProviderToolState {
|
||||
call_id: String,
|
||||
name: String,
|
||||
arguments: String,
|
||||
emitted_arguments_len: usize,
|
||||
started_emitted: bool,
|
||||
}
|
||||
|
||||
@@ -510,6 +511,74 @@ impl OpenAIResponsesProviderState {
|
||||
});
|
||||
}
|
||||
|
||||
fn emit_ready_tool_call(
|
||||
&mut self,
|
||||
report_context: &Value,
|
||||
out: &mut Vec<CanonicalStreamFrame>,
|
||||
index: usize,
|
||||
) {
|
||||
let (id, model) = self.identity(report_context);
|
||||
let Some(state) = self.tool_calls.get_mut(&index) else {
|
||||
return;
|
||||
};
|
||||
if state.name.is_empty() {
|
||||
return;
|
||||
}
|
||||
if !state.started_emitted {
|
||||
out.push(CanonicalStreamFrame {
|
||||
id: id.clone(),
|
||||
model: model.clone(),
|
||||
event: CanonicalStreamEvent::ToolCallStart {
|
||||
index,
|
||||
call_id: if state.call_id.is_empty() {
|
||||
build_generated_tool_call_id(index)
|
||||
} else {
|
||||
state.call_id.clone()
|
||||
},
|
||||
name: state.name.clone(),
|
||||
},
|
||||
});
|
||||
state.started_emitted = true;
|
||||
}
|
||||
if state.emitted_arguments_len > state.arguments.len() {
|
||||
state.emitted_arguments_len = 0;
|
||||
}
|
||||
let pending = state
|
||||
.arguments
|
||||
.get(state.emitted_arguments_len..)
|
||||
.unwrap_or_default()
|
||||
.to_string();
|
||||
if pending.is_empty() {
|
||||
return;
|
||||
}
|
||||
state.emitted_arguments_len = state.arguments.len();
|
||||
out.push(CanonicalStreamFrame {
|
||||
id,
|
||||
model,
|
||||
event: CanonicalStreamEvent::ToolCallArgumentsDelta {
|
||||
index,
|
||||
arguments: pending,
|
||||
},
|
||||
});
|
||||
}
|
||||
|
||||
fn merge_tool_call_arguments(state: &mut OpenAIResponsesProviderToolState, arguments: &str) {
|
||||
if arguments.is_empty() {
|
||||
return;
|
||||
}
|
||||
if arguments.starts_with(&state.arguments) {
|
||||
state
|
||||
.arguments
|
||||
.push_str(&arguments[state.arguments.len()..]);
|
||||
} else if state.arguments != arguments {
|
||||
if state.emitted_arguments_len == 0 {
|
||||
state.arguments = arguments.to_string();
|
||||
} else {
|
||||
state.arguments.push_str(arguments);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
fn emit_tool_call_item(
|
||||
&mut self,
|
||||
report_context: &Value,
|
||||
@@ -527,7 +596,6 @@ impl OpenAIResponsesProviderState {
|
||||
.and_then(Value::as_str)
|
||||
.map(ToOwned::to_owned);
|
||||
let index = self.tool_index_for_key(key, output_index);
|
||||
let (id, model) = self.identity(report_context);
|
||||
let state = self.tool_calls.entry(index).or_default();
|
||||
state.call_id = item
|
||||
.get("call_id")
|
||||
@@ -540,50 +608,13 @@ impl OpenAIResponsesProviderState {
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or(state.name.as_str())
|
||||
.to_string();
|
||||
if !state.started_emitted {
|
||||
out.push(CanonicalStreamFrame {
|
||||
id: id.clone(),
|
||||
model: model.clone(),
|
||||
event: CanonicalStreamEvent::ToolCallStart {
|
||||
index,
|
||||
call_id: if state.call_id.is_empty() {
|
||||
build_generated_tool_call_id(index)
|
||||
} else {
|
||||
state.call_id.clone()
|
||||
},
|
||||
name: if state.name.is_empty() {
|
||||
"unknown".to_string()
|
||||
} else {
|
||||
state.name.clone()
|
||||
},
|
||||
},
|
||||
});
|
||||
state.started_emitted = true;
|
||||
}
|
||||
let completed_arguments = item
|
||||
.get("arguments")
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or_default()
|
||||
.to_string();
|
||||
let missing = if completed_arguments.starts_with(&state.arguments) {
|
||||
completed_arguments[state.arguments.len()..].to_string()
|
||||
} else if state.arguments == completed_arguments {
|
||||
String::new()
|
||||
} else {
|
||||
completed_arguments.clone()
|
||||
};
|
||||
if missing.is_empty() {
|
||||
return;
|
||||
}
|
||||
state.arguments.push_str(&missing);
|
||||
out.push(CanonicalStreamFrame {
|
||||
id,
|
||||
model,
|
||||
event: CanonicalStreamEvent::ToolCallArgumentsDelta {
|
||||
index,
|
||||
arguments: missing,
|
||||
},
|
||||
});
|
||||
Self::merge_tool_call_arguments(state, &completed_arguments);
|
||||
self.emit_ready_tool_call(report_context, out, index);
|
||||
}
|
||||
|
||||
fn emit_missing_tool_result(
|
||||
@@ -978,44 +1009,22 @@ impl OpenAIResponsesProviderState {
|
||||
.and_then(Value::as_u64)
|
||||
.map(|value| value as usize);
|
||||
let index = self.tool_index_for_key(key, output_index);
|
||||
let (id, model) = self.identity(report_context);
|
||||
let state = self.tool_calls.entry(index).or_default();
|
||||
state.call_id = value
|
||||
.get("item_id")
|
||||
.or_else(|| value.get("call_id"))
|
||||
if let Some(call_id) = value
|
||||
.get("call_id")
|
||||
.or_else(|| value.get("id"))
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or(state.call_id.as_str())
|
||||
.to_string();
|
||||
if !state.started_emitted {
|
||||
out.push(CanonicalStreamFrame {
|
||||
id: id.clone(),
|
||||
model: model.clone(),
|
||||
event: CanonicalStreamEvent::ToolCallStart {
|
||||
index,
|
||||
call_id: if state.call_id.is_empty() {
|
||||
build_generated_tool_call_id(index)
|
||||
} else {
|
||||
state.call_id.clone()
|
||||
},
|
||||
name: if state.name.is_empty() {
|
||||
"unknown".to_string()
|
||||
} else {
|
||||
state.name.clone()
|
||||
},
|
||||
},
|
||||
});
|
||||
state.started_emitted = true;
|
||||
{
|
||||
state.call_id = call_id.to_string();
|
||||
} else if state.call_id.is_empty() {
|
||||
state.call_id = value
|
||||
.get("item_id")
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or_default()
|
||||
.to_string();
|
||||
}
|
||||
state.arguments.push_str(delta);
|
||||
out.push(CanonicalStreamFrame {
|
||||
id,
|
||||
model,
|
||||
event: CanonicalStreamEvent::ToolCallArgumentsDelta {
|
||||
index,
|
||||
arguments: delta.to_string(),
|
||||
},
|
||||
});
|
||||
self.emit_ready_tool_call(report_context, &mut out, index);
|
||||
}
|
||||
"response.function_call_arguments.done" => {
|
||||
let arguments = value
|
||||
@@ -1029,9 +1038,6 @@ impl OpenAIResponsesProviderState {
|
||||
.and_then(Value::as_str)
|
||||
})
|
||||
.unwrap_or_default();
|
||||
if arguments.is_empty() {
|
||||
return Ok(out);
|
||||
}
|
||||
self.ensure_started(report_context, &mut out);
|
||||
let key = value
|
||||
.get("item_id")
|
||||
@@ -1052,11 +1058,9 @@ impl OpenAIResponsesProviderState {
|
||||
.and_then(Value::as_u64)
|
||||
.map(|value| value as usize);
|
||||
let index = self.tool_index_for_key(key, output_index);
|
||||
let (id, model) = self.identity(report_context);
|
||||
let state = self.tool_calls.entry(index).or_default();
|
||||
state.call_id = value
|
||||
.get("item_id")
|
||||
.or_else(|| value.get("call_id"))
|
||||
.get("call_id")
|
||||
.or_else(|| value.get("id"))
|
||||
.and_then(Value::as_str)
|
||||
.or_else(|| {
|
||||
@@ -1066,46 +1070,23 @@ impl OpenAIResponsesProviderState {
|
||||
.and_then(|item| item.get("call_id").or_else(|| item.get("id")))
|
||||
.and_then(Value::as_str)
|
||||
})
|
||||
.or_else(|| value.get("item_id").and_then(Value::as_str))
|
||||
.unwrap_or(state.call_id.as_str())
|
||||
.to_string();
|
||||
if !state.started_emitted {
|
||||
out.push(CanonicalStreamFrame {
|
||||
id: id.clone(),
|
||||
model: model.clone(),
|
||||
event: CanonicalStreamEvent::ToolCallStart {
|
||||
index,
|
||||
call_id: if state.call_id.is_empty() {
|
||||
build_generated_tool_call_id(index)
|
||||
} else {
|
||||
state.call_id.clone()
|
||||
},
|
||||
name: if state.name.is_empty() {
|
||||
"unknown".to_string()
|
||||
} else {
|
||||
state.name.clone()
|
||||
},
|
||||
},
|
||||
});
|
||||
state.started_emitted = true;
|
||||
}
|
||||
let missing = if arguments.starts_with(&state.arguments) {
|
||||
arguments[state.arguments.len()..].to_string()
|
||||
} else if state.arguments == arguments {
|
||||
String::new()
|
||||
} else {
|
||||
arguments.to_string()
|
||||
};
|
||||
if !missing.is_empty() {
|
||||
state.arguments.push_str(&missing);
|
||||
out.push(CanonicalStreamFrame {
|
||||
id,
|
||||
model,
|
||||
event: CanonicalStreamEvent::ToolCallArgumentsDelta {
|
||||
index,
|
||||
arguments: missing,
|
||||
},
|
||||
});
|
||||
}
|
||||
state.name = value
|
||||
.get("name")
|
||||
.and_then(Value::as_str)
|
||||
.or_else(|| {
|
||||
value
|
||||
.get("item")
|
||||
.and_then(Value::as_object)
|
||||
.and_then(|item| item.get("name"))
|
||||
.and_then(Value::as_str)
|
||||
})
|
||||
.unwrap_or(state.name.as_str())
|
||||
.to_string();
|
||||
Self::merge_tool_call_arguments(state, arguments);
|
||||
self.emit_ready_tool_call(report_context, &mut out, index);
|
||||
}
|
||||
"response.function_call_output.delta" | "response.function_call_output.done" => {
|
||||
let tool_use_id = value
|
||||
@@ -2661,6 +2642,7 @@ fn openai_tool_result_content_from_value(value: Option<&Value>) -> String {
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
use crate::formats::claude::messages::stream::ClaudeClientEmitter;
|
||||
|
||||
fn data_line(value: Value) -> Vec<u8> {
|
||||
format!("data: {}\n", value).into_bytes()
|
||||
@@ -3176,6 +3158,91 @@ mod tests {
|
||||
)));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn openai_responses_provider_state_delays_arguments_until_tool_name_is_known() {
|
||||
let mut state = OpenAIResponsesProviderState::default();
|
||||
let report_context = json!({});
|
||||
let arguments = r#"{"file_path":"D:/projects/UIAutoTest/docs/prd/msr.md","offset":0,"limit":2000,"pages":""}"#;
|
||||
let mut frames = Vec::new();
|
||||
|
||||
let delta_frames = state
|
||||
.push_line(
|
||||
&report_context,
|
||||
data_line(json!({
|
||||
"type": "response.function_call_arguments.delta",
|
||||
"response_id": "resp_read_123",
|
||||
"output_index": 0,
|
||||
"item_id": "fc_read_123",
|
||||
"delta": arguments,
|
||||
})),
|
||||
)
|
||||
.expect("arguments delta should parse");
|
||||
|
||||
assert!(matches!(
|
||||
delta_frames.first().map(|frame| &frame.event),
|
||||
Some(CanonicalStreamEvent::Start)
|
||||
));
|
||||
assert!(!delta_frames.iter().any(|frame| matches!(
|
||||
frame.event,
|
||||
CanonicalStreamEvent::ToolCallStart { .. }
|
||||
| CanonicalStreamEvent::ToolCallArgumentsDelta { .. }
|
||||
)));
|
||||
frames.extend(delta_frames);
|
||||
|
||||
frames.extend(
|
||||
state
|
||||
.push_line(
|
||||
&report_context,
|
||||
data_line(json!({
|
||||
"type": "response.function_call_arguments.done",
|
||||
"response_id": "resp_read_123",
|
||||
"output_index": 0,
|
||||
"item_id": "fc_read_123",
|
||||
"item": {
|
||||
"type": "function_call",
|
||||
"id": "fc_read_123",
|
||||
"call_id": "call_read_123",
|
||||
"name": "Read",
|
||||
"arguments": arguments,
|
||||
}
|
||||
})),
|
||||
)
|
||||
.expect("arguments done should parse"),
|
||||
);
|
||||
|
||||
assert!(frames.iter().any(|frame| matches!(
|
||||
frame.event,
|
||||
CanonicalStreamEvent::ToolCallStart {
|
||||
ref call_id,
|
||||
ref name,
|
||||
..
|
||||
} if call_id == "call_read_123" && name == "Read"
|
||||
)));
|
||||
assert!(frames.iter().any(|frame| matches!(
|
||||
frame.event,
|
||||
CanonicalStreamEvent::ToolCallArgumentsDelta {
|
||||
ref arguments,
|
||||
..
|
||||
} if arguments.contains(r#""pages":"""#)
|
||||
)));
|
||||
|
||||
let mut emitter = ClaudeClientEmitter::default();
|
||||
let mut bytes = Vec::new();
|
||||
for frame in frames {
|
||||
bytes.extend(emitter.emit(frame).expect("claude frame should encode"));
|
||||
}
|
||||
bytes.extend(
|
||||
emitter
|
||||
.finish()
|
||||
.expect("claude stream finish should encode"),
|
||||
);
|
||||
let sse = String::from_utf8(bytes).expect("claude sse should be utf8");
|
||||
|
||||
assert!(sse.contains("\"name\":\"Read\""));
|
||||
assert!(sse.contains("\\\"limit\\\":2000"));
|
||||
assert!(!sse.contains("\\\"pages\\\":\\\"\\\""));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn openai_responses_provider_state_parses_function_call_output_as_tool_result() {
|
||||
let mut state = OpenAIResponsesProviderState::default();
|
||||
|
||||
@@ -29,7 +29,10 @@ pub fn canonicalize_tool_arguments(value: Option<Value>) -> String {
|
||||
}
|
||||
}
|
||||
|
||||
pub fn remove_empty_pages_from_tool_arguments(arguments: &str) -> String {
|
||||
pub fn remove_empty_pages_from_tool_arguments(tool_name: &str, arguments: &str) -> String {
|
||||
if tool_name != "Read" {
|
||||
return arguments.to_string();
|
||||
}
|
||||
let Ok(mut value) = serde_json::from_str::<Value>(arguments) else {
|
||||
return arguments.to_string();
|
||||
};
|
||||
@@ -43,6 +46,44 @@ pub fn remove_empty_pages_from_tool_arguments(arguments: &str) -> String {
|
||||
serde_json::to_string(&value).unwrap_or_else(|_| arguments.to_string())
|
||||
}
|
||||
|
||||
pub fn remove_empty_pages_from_tool_input_value(tool_name: &str, input: &Value) -> Value {
|
||||
if tool_name != "Read" || input.get("pages").and_then(Value::as_str) != Some("") {
|
||||
return input.clone();
|
||||
}
|
||||
let Some(object) = input.as_object() else {
|
||||
return input.clone();
|
||||
};
|
||||
let mut object = object.clone();
|
||||
object.remove("pages");
|
||||
Value::Object(object)
|
||||
}
|
||||
|
||||
pub fn sanitize_claude_read_tool_inputs(value: &mut Value) -> bool {
|
||||
let Some(content) = value.get_mut("content").and_then(Value::as_array_mut) else {
|
||||
return false;
|
||||
};
|
||||
let mut changed = false;
|
||||
for block in content {
|
||||
let Some(block_object) = block.as_object_mut() else {
|
||||
continue;
|
||||
};
|
||||
if block_object.get("type").and_then(Value::as_str) != Some("tool_use")
|
||||
|| block_object.get("name").and_then(Value::as_str) != Some("Read")
|
||||
{
|
||||
continue;
|
||||
}
|
||||
let Some(input) = block_object.get("input") else {
|
||||
continue;
|
||||
};
|
||||
let sanitized = remove_empty_pages_from_tool_input_value("Read", input);
|
||||
if sanitized != *input {
|
||||
block_object.insert("input".to_string(), sanitized);
|
||||
changed = true;
|
||||
}
|
||||
}
|
||||
changed
|
||||
}
|
||||
|
||||
pub fn prepare_local_success_response_parts(
|
||||
headers: &BTreeMap<String, String>,
|
||||
body_json: &Value,
|
||||
@@ -131,7 +172,8 @@ mod tests {
|
||||
build_generated_tool_call_id, build_local_success_background_report,
|
||||
build_local_success_conversion_background_report, canonicalize_tool_arguments,
|
||||
prepare_local_success_response_parts, prepare_local_success_response_parts_owned,
|
||||
remove_empty_pages_from_tool_arguments, LocalSyncReportParts,
|
||||
remove_empty_pages_from_tool_arguments, sanitize_claude_read_tool_inputs,
|
||||
LocalSyncReportParts,
|
||||
};
|
||||
use std::collections::BTreeMap;
|
||||
|
||||
@@ -153,20 +195,77 @@ mod tests {
|
||||
fn removes_empty_pages_from_tool_arguments() {
|
||||
assert_eq!(
|
||||
remove_empty_pages_from_tool_arguments(
|
||||
"Read",
|
||||
r#"{"file_path":"/tmp/a.txt","offset":1,"limit":20,"pages":""}"#
|
||||
),
|
||||
r#"{"file_path":"/tmp/a.txt","offset":1,"limit":20}"#
|
||||
);
|
||||
assert_eq!(
|
||||
remove_empty_pages_from_tool_arguments(r#"{"pages":"1-2"}"#),
|
||||
remove_empty_pages_from_tool_arguments("Search", r#"{"query":"","pages":""}"#),
|
||||
r#"{"query":"","pages":""}"#
|
||||
);
|
||||
assert_eq!(
|
||||
remove_empty_pages_from_tool_arguments("Read", r#"{"pages":"1-2"}"#),
|
||||
r#"{"pages":"1-2"}"#
|
||||
);
|
||||
assert_eq!(
|
||||
remove_empty_pages_from_tool_arguments(r#"{"pages":"#),
|
||||
remove_empty_pages_from_tool_arguments("Read", r#"{"pages":"#),
|
||||
r#"{"pages":"#
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn sanitizes_claude_read_tool_inputs_only() {
|
||||
let mut value = serde_json::json!({
|
||||
"content": [
|
||||
{
|
||||
"type": "tool_use",
|
||||
"name": "Read",
|
||||
"input": {
|
||||
"file_path": "/tmp/a.txt",
|
||||
"limit": 20,
|
||||
"pages": ""
|
||||
}
|
||||
},
|
||||
{
|
||||
"type": "tool_use",
|
||||
"name": "Search",
|
||||
"input": {
|
||||
"query": "",
|
||||
"pages": ""
|
||||
}
|
||||
},
|
||||
{
|
||||
"type": "tool_use",
|
||||
"name": "Read",
|
||||
"input": {
|
||||
"pages": "1-2"
|
||||
}
|
||||
}
|
||||
]
|
||||
});
|
||||
|
||||
assert!(sanitize_claude_read_tool_inputs(&mut value));
|
||||
assert_eq!(
|
||||
value["content"][0]["input"],
|
||||
serde_json::json!({
|
||||
"file_path": "/tmp/a.txt",
|
||||
"limit": 20,
|
||||
})
|
||||
);
|
||||
assert_eq!(
|
||||
value["content"][1]["input"],
|
||||
serde_json::json!({
|
||||
"query": "",
|
||||
"pages": "",
|
||||
})
|
||||
);
|
||||
assert_eq!(
|
||||
value["content"][2]["input"],
|
||||
serde_json::json!({"pages": "1-2"})
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn prepare_local_success_response_parts_normalizes_headers() {
|
||||
let headers = BTreeMap::from([
|
||||
|
||||
@@ -1,7 +1,13 @@
|
||||
use serde_json::Value;
|
||||
use std::collections::BTreeMap;
|
||||
|
||||
use serde_json::{json, Map, Value};
|
||||
|
||||
use crate::formats::openai::image::stream::{OpenAiImageChatStreamState, OpenAiImageStreamState};
|
||||
use crate::formats::shared::model_directives::model_directive_display_model_from_report_context;
|
||||
use crate::formats::shared::response::{
|
||||
remove_empty_pages_from_tool_arguments, remove_empty_pages_from_tool_input_value,
|
||||
};
|
||||
use crate::formats::shared::sse::encode_json_sse;
|
||||
use crate::formats::shared::stream_core::StreamingStandardFormatMatrix;
|
||||
use crate::formats::shared::AiSurfaceFinalizeError;
|
||||
use crate::provider_compat::kiro_stream::KiroToClaudeCliStreamState;
|
||||
@@ -16,6 +22,7 @@ pub enum FinalizeStreamRewriteMode {
|
||||
ModelDirectiveDisplay,
|
||||
OpenAiImage,
|
||||
OpenAiImageToOpenAiChat,
|
||||
ClaudeReadToolSanitize,
|
||||
Standard,
|
||||
KiroToClaudeCli,
|
||||
KiroToClaudeCliThenStandard,
|
||||
@@ -84,6 +91,9 @@ pub fn resolve_finalize_stream_rewrite_mode(
|
||||
// Parsing→rebuilding only adds overhead and may lose information
|
||||
// (encrypted_content, original item IDs, etc.).
|
||||
if is_same_format_family(provider_api_format.as_str(), client_api_format.as_str()) {
|
||||
if provider_api_format == "claude:messages" && client_api_format == "claude:messages" {
|
||||
return Some(FinalizeStreamRewriteMode::ClaudeReadToolSanitize);
|
||||
}
|
||||
return model_directive_display_model_from_report_context(report_context)
|
||||
.map(|_| FinalizeStreamRewriteMode::ModelDirectiveDisplay);
|
||||
}
|
||||
@@ -108,9 +118,16 @@ pub fn resolve_finalize_stream_rewrite_mode(
|
||||
provider_api_format.as_str(),
|
||||
)
|
||||
{
|
||||
if provider_api_format == "claude:messages" {
|
||||
return Some(FinalizeStreamRewriteMode::ClaudeReadToolSanitize);
|
||||
}
|
||||
return Some(FinalizeStreamRewriteMode::ModelDirectiveDisplay);
|
||||
}
|
||||
|
||||
if provider_api_format == "claude:messages" && client_api_format == "claude:messages" {
|
||||
return Some(FinalizeStreamRewriteMode::ClaudeReadToolSanitize);
|
||||
}
|
||||
|
||||
(provider_api_format == client_api_format
|
||||
&& provider_adaptation_should_unwrap_stream_envelope(
|
||||
envelope_name.as_str(),
|
||||
@@ -144,6 +161,7 @@ enum AiSurfaceStreamRewriteState {
|
||||
ModelDirectiveDisplay,
|
||||
OpenAiImage(Box<OpenAiImageStreamState>),
|
||||
OpenAiImageToOpenAiChat(Box<OpenAiImageChatStreamState>),
|
||||
ClaudeReadToolSanitize(Box<ClaudeReadToolStreamSanitizer>),
|
||||
Standard(Box<StreamingStandardFormatMatrix>),
|
||||
KiroToClaudeCli(Box<KiroToClaudeCliStreamState>),
|
||||
KiroToClaudeCliThenStandard {
|
||||
@@ -175,6 +193,11 @@ pub fn maybe_build_ai_surface_stream_rewriter<'a>(
|
||||
Box::<OpenAiImageChatStreamState>::default(),
|
||||
)
|
||||
}
|
||||
FinalizeStreamRewriteMode::ClaudeReadToolSanitize => {
|
||||
AiSurfaceStreamRewriteState::ClaudeReadToolSanitize(
|
||||
Box::<ClaudeReadToolStreamSanitizer>::default(),
|
||||
)
|
||||
}
|
||||
FinalizeStreamRewriteMode::Standard => {
|
||||
AiSurfaceStreamRewriteState::Standard(Box::<StreamingStandardFormatMatrix>::default())
|
||||
}
|
||||
@@ -205,6 +228,9 @@ impl AiSurfaceStreamRewriter<'_> {
|
||||
AiSurfaceStreamRewriteState::OpenAiImageToOpenAiChat(state) => {
|
||||
state.push_chunk(self.report_context, chunk)
|
||||
}
|
||||
AiSurfaceStreamRewriteState::ClaudeReadToolSanitize(state) => {
|
||||
state.push_chunk(self.report_context, chunk)
|
||||
}
|
||||
AiSurfaceStreamRewriteState::KiroToClaudeCli(state) => {
|
||||
state.push_chunk(self.report_context, chunk)
|
||||
}
|
||||
@@ -232,6 +258,9 @@ impl AiSurfaceStreamRewriter<'_> {
|
||||
AiSurfaceStreamRewriteState::OpenAiImageToOpenAiChat(state) => {
|
||||
state.finish(self.report_context)
|
||||
}
|
||||
AiSurfaceStreamRewriteState::ClaudeReadToolSanitize(state) => {
|
||||
state.finish(self.report_context)
|
||||
}
|
||||
AiSurfaceStreamRewriteState::KiroToClaudeCli(state) => {
|
||||
state.finish(self.report_context)
|
||||
}
|
||||
@@ -278,12 +307,272 @@ impl AiSurfaceStreamRewriter<'_> {
|
||||
}
|
||||
AiSurfaceStreamRewriteState::OpenAiImage(_)
|
||||
| AiSurfaceStreamRewriteState::OpenAiImageToOpenAiChat(_)
|
||||
| AiSurfaceStreamRewriteState::ClaudeReadToolSanitize(_)
|
||||
| AiSurfaceStreamRewriteState::KiroToClaudeCli(_)
|
||||
| AiSurfaceStreamRewriteState::KiroToClaudeCliThenStandard { .. } => Ok(Vec::new()),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[derive(Default)]
|
||||
struct ClaudeReadToolBlockState {
|
||||
name: String,
|
||||
buffered_input_json: String,
|
||||
}
|
||||
|
||||
#[derive(Default)]
|
||||
struct ClaudeReadToolStreamSanitizer {
|
||||
buffered: Vec<u8>,
|
||||
blocks: BTreeMap<usize, ClaudeReadToolBlockState>,
|
||||
}
|
||||
|
||||
impl ClaudeReadToolStreamSanitizer {
|
||||
fn push_chunk(
|
||||
&mut self,
|
||||
report_context: &Value,
|
||||
chunk: &[u8],
|
||||
) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
self.buffered.extend_from_slice(chunk);
|
||||
let mut output = Vec::new();
|
||||
while let Some(record) = drain_next_sse_record(&mut self.buffered) {
|
||||
output.extend(self.transform_record(report_context, record)?);
|
||||
}
|
||||
Ok(output)
|
||||
}
|
||||
|
||||
fn finish(&mut self, report_context: &Value) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
if self.buffered.is_empty() {
|
||||
return Ok(Vec::new());
|
||||
}
|
||||
let record = std::mem::take(&mut self.buffered);
|
||||
self.transform_record(report_context, record)
|
||||
}
|
||||
|
||||
fn transform_record(
|
||||
&mut self,
|
||||
report_context: &Value,
|
||||
record: Vec<u8>,
|
||||
) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
let Some((event, mut payload)) = parse_sse_record_json(&record) else {
|
||||
return rewrite_model_directive_stream_record(report_context, record);
|
||||
};
|
||||
let event_type = payload
|
||||
.get("type")
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or(event.as_deref().unwrap_or_default())
|
||||
.to_string();
|
||||
let mut output = match event_type.as_str() {
|
||||
"content_block_start" => {
|
||||
self.transform_content_block_start(event.as_deref(), payload, record)?
|
||||
}
|
||||
"content_block_delta" => self.transform_content_block_delta(payload, record)?,
|
||||
"content_block_stop" => self.transform_content_block_stop(payload, record)?,
|
||||
_ => {
|
||||
if !rewrite_stream_payload_model_from_context(report_context, &mut payload) {
|
||||
return Ok(record);
|
||||
}
|
||||
encode_json_sse(event.as_deref(), &payload)?
|
||||
}
|
||||
};
|
||||
if model_directive_display_model_from_report_context(report_context).is_some()
|
||||
&& !matches!(event_type.as_str(), "message_start" | "message_delta")
|
||||
{
|
||||
output = rewrite_model_directive_stream_record(report_context, output)?;
|
||||
}
|
||||
Ok(output)
|
||||
}
|
||||
|
||||
fn transform_content_block_start(
|
||||
&mut self,
|
||||
event: Option<&str>,
|
||||
mut payload: Value,
|
||||
original_record: Vec<u8>,
|
||||
) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
let index = payload
|
||||
.get("index")
|
||||
.and_then(Value::as_u64)
|
||||
.map(|value| value as usize)
|
||||
.unwrap_or(0);
|
||||
let Some(block) = payload
|
||||
.get_mut("content_block")
|
||||
.and_then(Value::as_object_mut)
|
||||
else {
|
||||
return Ok(original_record);
|
||||
};
|
||||
let block_type = block
|
||||
.get("type")
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or_default();
|
||||
if block_type != "tool_use" {
|
||||
return Ok(original_record);
|
||||
}
|
||||
let name = block
|
||||
.get("name")
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or_default()
|
||||
.to_string();
|
||||
self.blocks.insert(
|
||||
index,
|
||||
ClaudeReadToolBlockState {
|
||||
name: name.clone(),
|
||||
buffered_input_json: String::new(),
|
||||
},
|
||||
);
|
||||
if sanitize_claude_tool_input_object(block, &name) {
|
||||
encode_json_sse(event, &payload)
|
||||
} else {
|
||||
Ok(original_record)
|
||||
}
|
||||
}
|
||||
|
||||
fn transform_content_block_delta(
|
||||
&mut self,
|
||||
payload: Value,
|
||||
original_record: Vec<u8>,
|
||||
) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
let index = payload
|
||||
.get("index")
|
||||
.and_then(Value::as_u64)
|
||||
.map(|value| value as usize)
|
||||
.unwrap_or(0);
|
||||
let delta_type = payload
|
||||
.get("delta")
|
||||
.and_then(Value::as_object)
|
||||
.and_then(|delta| delta.get("type"))
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or_default();
|
||||
let partial_json = payload
|
||||
.get("delta")
|
||||
.and_then(Value::as_object)
|
||||
.and_then(|delta| delta.get("partial_json"))
|
||||
.and_then(Value::as_str);
|
||||
if delta_type != "input_json_delta" {
|
||||
return Ok(original_record);
|
||||
}
|
||||
let Some(state) = self.blocks.get_mut(&index) else {
|
||||
return Ok(original_record);
|
||||
};
|
||||
if state.name != "Read" {
|
||||
return Ok(original_record);
|
||||
}
|
||||
if let Some(partial_json) = partial_json {
|
||||
state.buffered_input_json.push_str(partial_json);
|
||||
}
|
||||
Ok(Vec::new())
|
||||
}
|
||||
|
||||
fn transform_content_block_stop(
|
||||
&mut self,
|
||||
payload: Value,
|
||||
original_record: Vec<u8>,
|
||||
) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
let index = payload
|
||||
.get("index")
|
||||
.and_then(Value::as_u64)
|
||||
.map(|value| value as usize)
|
||||
.unwrap_or(0);
|
||||
let Some(state) = self.blocks.remove(&index) else {
|
||||
return Ok(original_record);
|
||||
};
|
||||
let mut output = Vec::new();
|
||||
if state.name == "Read" && !state.buffered_input_json.is_empty() {
|
||||
let partial_json =
|
||||
remove_empty_pages_from_tool_arguments("Read", &state.buffered_input_json);
|
||||
if !partial_json.is_empty() {
|
||||
output.extend(encode_json_sse(
|
||||
Some("content_block_delta"),
|
||||
&json!({
|
||||
"type": "content_block_delta",
|
||||
"index": index,
|
||||
"delta": {
|
||||
"type": "input_json_delta",
|
||||
"partial_json": partial_json,
|
||||
}
|
||||
}),
|
||||
)?);
|
||||
}
|
||||
}
|
||||
if output.is_empty() {
|
||||
output = original_record;
|
||||
} else {
|
||||
output.extend(original_record);
|
||||
}
|
||||
Ok(output)
|
||||
}
|
||||
}
|
||||
|
||||
fn sanitize_claude_tool_input_object(block: &mut Map<String, Value>, name: &str) -> bool {
|
||||
let Some(input) = block.get("input") else {
|
||||
return false;
|
||||
};
|
||||
let sanitized = remove_empty_pages_from_tool_input_value(name, input);
|
||||
if sanitized == *input {
|
||||
return false;
|
||||
}
|
||||
block.insert("input".to_string(), sanitized);
|
||||
true
|
||||
}
|
||||
|
||||
fn drain_next_sse_record(buffer: &mut Vec<u8>) -> Option<Vec<u8>> {
|
||||
let mut line_start = 0usize;
|
||||
let mut index = 0usize;
|
||||
while index < buffer.len() {
|
||||
if buffer[index] != b'\n' {
|
||||
index += 1;
|
||||
continue;
|
||||
}
|
||||
let line_end = index + 1;
|
||||
let line = &buffer[line_start..line_end];
|
||||
let line_without_newline = line
|
||||
.strip_suffix(b"\n")
|
||||
.unwrap_or(line)
|
||||
.strip_suffix(b"\r")
|
||||
.unwrap_or_else(|| line.strip_suffix(b"\n").unwrap_or(line));
|
||||
if line_without_newline.is_empty() {
|
||||
return Some(buffer.drain(..line_end).collect());
|
||||
}
|
||||
line_start = line_end;
|
||||
index = line_end;
|
||||
}
|
||||
None
|
||||
}
|
||||
|
||||
fn parse_sse_record_json(record: &[u8]) -> Option<(Option<String>, Value)> {
|
||||
let text = std::str::from_utf8(record).ok()?;
|
||||
let mut event = None;
|
||||
let mut data = String::new();
|
||||
for line in text.lines() {
|
||||
let line = line.strip_suffix('\r').unwrap_or(line);
|
||||
if let Some(value) = line.strip_prefix("event:") {
|
||||
event = Some(value.trim().to_string());
|
||||
} else if let Some(value) = line.strip_prefix("data:") {
|
||||
if !data.is_empty() {
|
||||
data.push('\n');
|
||||
}
|
||||
data.push_str(value.trim_start());
|
||||
}
|
||||
}
|
||||
if data.trim().is_empty() || data.trim() == "[DONE]" {
|
||||
return None;
|
||||
}
|
||||
let value = serde_json::from_str::<Value>(data.trim()).ok()?;
|
||||
Some((event, value))
|
||||
}
|
||||
|
||||
fn rewrite_model_directive_stream_record(
|
||||
report_context: &Value,
|
||||
record: Vec<u8>,
|
||||
) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
let mut output = Vec::new();
|
||||
for line in record.split_inclusive(|byte| *byte == b'\n') {
|
||||
output.extend(rewrite_model_directive_stream_line(
|
||||
report_context,
|
||||
line.to_vec(),
|
||||
)?);
|
||||
}
|
||||
Ok(output)
|
||||
}
|
||||
|
||||
fn rewrite_model_directive_stream_line(
|
||||
report_context: &Value,
|
||||
line: Vec<u8>,
|
||||
@@ -341,6 +630,14 @@ fn rewrite_stream_payload_model(value: &mut Value, display_model: &str) -> bool
|
||||
changed
|
||||
}
|
||||
|
||||
fn rewrite_stream_payload_model_from_context(report_context: &Value, value: &mut Value) -> bool {
|
||||
let Some(display_model) = model_directive_display_model_from_report_context(report_context)
|
||||
else {
|
||||
return false;
|
||||
};
|
||||
rewrite_stream_payload_model(value, &display_model)
|
||||
}
|
||||
|
||||
fn transform_standard_bytes(
|
||||
standard: &mut StreamingStandardFormatMatrix,
|
||||
report_context: &Value,
|
||||
@@ -718,14 +1015,106 @@ data: {\"type\":\"content_block_delta\",\"index\":1,\"delta\":{\"type\":\"thinki
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn same_format_claude_without_display_model_passes_through_verbatim() {
|
||||
// Claude→Claude without display model: no rewriter needed at all.
|
||||
fn same_format_claude_uses_read_tool_sanitizer_without_display_model() {
|
||||
// Claude→Claude needs a narrow sanitizer for Claude Code Read input.
|
||||
let report_context = json!({
|
||||
"provider_api_format": "claude:messages",
|
||||
"client_api_format": "claude:messages",
|
||||
"needs_conversion": true,
|
||||
});
|
||||
assert!(maybe_build_ai_surface_stream_rewriter(Some(&report_context)).is_none());
|
||||
assert_eq!(
|
||||
resolve_finalize_stream_rewrite_mode(&report_context),
|
||||
Some(FinalizeStreamRewriteMode::ClaudeReadToolSanitize)
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn same_format_claude_stream_sanitizes_read_start_input() {
|
||||
let report_context = json!({
|
||||
"provider_api_format": "claude:messages",
|
||||
"client_api_format": "claude:messages",
|
||||
"needs_conversion": false,
|
||||
});
|
||||
let mut rewriter = maybe_build_ai_surface_stream_rewriter(Some(&report_context))
|
||||
.expect("same-format claude sanitizer should exist");
|
||||
let output = rewriter
|
||||
.push_chunk(
|
||||
b"event: content_block_start\n\
|
||||
data: {\"type\":\"content_block_start\",\"index\":0,\"content_block\":{\"type\":\"tool_use\",\"id\":\"call_read_1\",\"name\":\"Read\",\"input\":{\"file_path\":\"/tmp/a.txt\",\"limit\":20,\"pages\":\"\"}}}\n\n",
|
||||
)
|
||||
.expect("rewrite should succeed");
|
||||
let output = String::from_utf8(output).expect("output should be utf8");
|
||||
|
||||
assert!(output.contains("\"name\":\"Read\""));
|
||||
assert!(output.contains("\"file_path\":\"/tmp/a.txt\""));
|
||||
assert!(!output.contains("\"pages\":\"\""));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn same_format_claude_stream_sanitizes_read_input_json_delta() {
|
||||
let report_context = json!({
|
||||
"provider_api_format": "claude:messages",
|
||||
"client_api_format": "claude:messages",
|
||||
"needs_conversion": false,
|
||||
});
|
||||
let mut rewriter = maybe_build_ai_surface_stream_rewriter(Some(&report_context))
|
||||
.expect("same-format claude sanitizer should exist");
|
||||
let mut output = rewriter
|
||||
.push_chunk(
|
||||
b"event: content_block_start\n\
|
||||
data: {\"type\":\"content_block_start\",\"index\":0,\"content_block\":{\"type\":\"tool_use\",\"id\":\"call_read_1\",\"name\":\"Read\",\"input\":{}}}\n\n",
|
||||
)
|
||||
.expect("start should rewrite");
|
||||
output.extend(
|
||||
rewriter
|
||||
.push_chunk(
|
||||
b"event: content_block_delta\n\
|
||||
data: {\"type\":\"content_block_delta\",\"index\":0,\"delta\":{\"type\":\"input_json_delta\",\"partial_json\":\"{\\\"file_path\\\":\\\"/tmp/a.txt\\\",\"}}\n\n\
|
||||
event: content_block_delta\n\
|
||||
data: {\"type\":\"content_block_delta\",\"index\":0,\"delta\":{\"type\":\"input_json_delta\",\"partial_json\":\"\\\"limit\\\":20,\\\"pages\\\":\\\"\\\"}\"}}\n\n",
|
||||
)
|
||||
.expect("deltas should buffer"),
|
||||
);
|
||||
let buffered_output = String::from_utf8(output.clone()).expect("output should be utf8");
|
||||
assert!(!buffered_output.contains("input_json_delta"));
|
||||
|
||||
output.extend(
|
||||
rewriter
|
||||
.push_chunk(
|
||||
b"event: content_block_stop\n\
|
||||
data: {\"type\":\"content_block_stop\",\"index\":0}\n\n",
|
||||
)
|
||||
.expect("stop should flush sanitized delta"),
|
||||
);
|
||||
let output = String::from_utf8(output).expect("output should be utf8");
|
||||
|
||||
assert!(output.contains("event: content_block_delta"));
|
||||
assert!(output.contains("\\\"limit\\\":20"));
|
||||
assert!(!output.contains("\\\"pages\\\":\\\"\\\""));
|
||||
assert!(output.contains("event: content_block_stop"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn same_format_claude_stream_preserves_other_tool_empty_pages() {
|
||||
let report_context = json!({
|
||||
"provider_api_format": "claude:messages",
|
||||
"client_api_format": "claude:messages",
|
||||
"needs_conversion": false,
|
||||
});
|
||||
let mut rewriter = maybe_build_ai_surface_stream_rewriter(Some(&report_context))
|
||||
.expect("same-format claude sanitizer should exist");
|
||||
let output = rewriter
|
||||
.push_chunk(
|
||||
b"event: content_block_start\n\
|
||||
data: {\"type\":\"content_block_start\",\"index\":0,\"content_block\":{\"type\":\"tool_use\",\"id\":\"call_search_1\",\"name\":\"Search\",\"input\":{}}}\n\n\
|
||||
event: content_block_delta\n\
|
||||
data: {\"type\":\"content_block_delta\",\"index\":0,\"delta\":{\"type\":\"input_json_delta\",\"partial_json\":\"{\\\"query\\\":\\\"\\\",\\\"pages\\\":\\\"\\\"}\"}}\n\n",
|
||||
)
|
||||
.expect("rewrite should succeed");
|
||||
let output = String::from_utf8(output).expect("output should be utf8");
|
||||
|
||||
assert!(output.contains("\"name\":\"Search\""));
|
||||
assert!(output.contains("\\\"pages\\\":\\\"\\\""));
|
||||
}
|
||||
|
||||
#[test]
|
||||
|
||||
@@ -21,7 +21,10 @@ use serde_json::{json, Map, Value};
|
||||
use super::AiSurfaceFinalizeError;
|
||||
use crate::formats::gemini::generate_content::stream::GeminiProviderState;
|
||||
use crate::formats::shared::model_directives::model_directive_display_model_from_report_context;
|
||||
use crate::formats::shared::response::remove_empty_pages_from_tool_arguments;
|
||||
use crate::formats::shared::response::{
|
||||
remove_empty_pages_from_tool_arguments, remove_empty_pages_from_tool_input_value,
|
||||
sanitize_claude_read_tool_inputs,
|
||||
};
|
||||
use crate::formats::shared::stream_core::common::{
|
||||
content_part_from_openai_image_generation_item, map_openai_finish_reason_to_gemini,
|
||||
parse_json_arguments_value, CanonicalContentPart, CanonicalStreamEvent, CanonicalUsage,
|
||||
@@ -480,8 +483,13 @@ fn maybe_build_standard_same_format_sync_body(
|
||||
return None;
|
||||
}
|
||||
|
||||
let mut body_json = body_json.clone();
|
||||
if expected_api_format == "claude:messages" {
|
||||
sanitize_claude_read_tool_inputs(&mut body_json);
|
||||
}
|
||||
|
||||
Some(client_body_with_report_context_model(
|
||||
body_json.clone(),
|
||||
body_json,
|
||||
report_context,
|
||||
&client_api_format,
|
||||
))
|
||||
@@ -2530,8 +2538,20 @@ pub fn aggregate_claude_stream_sync_response(body: &[u8]) -> Option<Value> {
|
||||
}
|
||||
}
|
||||
"tool_use" => {
|
||||
let tool_name = block
|
||||
.get("name")
|
||||
.and_then(Value::as_str)
|
||||
.unwrap_or_default()
|
||||
.to_string();
|
||||
if let Some(input) = block.get("input") {
|
||||
let sanitized = remove_empty_pages_from_tool_input_value(&tool_name, input);
|
||||
if sanitized != *input {
|
||||
block.insert("input".to_string(), sanitized);
|
||||
}
|
||||
}
|
||||
if !state.partial_json.is_empty() {
|
||||
let arguments = remove_empty_pages_from_tool_arguments(&state.partial_json);
|
||||
let arguments =
|
||||
remove_empty_pages_from_tool_arguments(&tool_name, &state.partial_json);
|
||||
let input = serde_json::from_str::<Value>(&arguments)
|
||||
.unwrap_or(Value::String(arguments));
|
||||
block.insert("input".to_string(), input);
|
||||
@@ -3114,6 +3134,59 @@ mod tests {
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn aggregates_claude_stream_removes_empty_pages_from_start_tool_input() {
|
||||
let body = concat!(
|
||||
"event: message_start\n",
|
||||
"data: {\"type\":\"message_start\",\"message\":{\"id\":\"msg_123\",\"type\":\"message\",\"role\":\"assistant\",\"model\":\"claude-sonnet-4-5\",\"content\":[],\"stop_reason\":null,\"stop_sequence\":null}}\n\n",
|
||||
"event: content_block_start\n",
|
||||
"data: {\"type\":\"content_block_start\",\"index\":0,\"content_block\":{\"type\":\"tool_use\",\"id\":\"toolu_read\",\"name\":\"Read\",\"input\":{\"file_path\":\"/tmp/a.txt\",\"limit\":20,\"pages\":\"\"}}}\n\n",
|
||||
"event: content_block_stop\n",
|
||||
"data: {\"type\":\"content_block_stop\",\"index\":0}\n\n",
|
||||
"event: message_stop\n",
|
||||
"data: {\"type\":\"message_stop\"}\n\n",
|
||||
);
|
||||
|
||||
let aggregated =
|
||||
aggregate_claude_stream_sync_response(body.as_bytes()).expect("body should aggregate");
|
||||
|
||||
assert_eq!(
|
||||
aggregated["content"][0]["input"],
|
||||
json!({
|
||||
"file_path": "/tmp/a.txt",
|
||||
"limit": 20,
|
||||
})
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn aggregates_claude_stream_preserves_empty_pages_for_non_read_tool_input() {
|
||||
let body = concat!(
|
||||
"event: message_start\n",
|
||||
"data: {\"type\":\"message_start\",\"message\":{\"id\":\"msg_123\",\"type\":\"message\",\"role\":\"assistant\",\"model\":\"claude-sonnet-4-5\",\"content\":[],\"stop_reason\":null,\"stop_sequence\":null}}\n\n",
|
||||
"event: content_block_start\n",
|
||||
"data: {\"type\":\"content_block_start\",\"index\":0,\"content_block\":{\"type\":\"tool_use\",\"id\":\"toolu_search\",\"name\":\"Search\",\"input\":{}}}\n\n",
|
||||
"event: content_block_delta\n",
|
||||
"data: {\"type\":\"content_block_delta\",\"index\":0,\"delta\":{\"type\":\"input_json_delta\",\"partial_json\":\"{\\\"query\\\":\\\"\\\",\\\"pages\\\":\\\"\\\"}\"}}\n\n",
|
||||
"event: content_block_stop\n",
|
||||
"data: {\"type\":\"content_block_stop\",\"index\":0}\n\n",
|
||||
"event: message_stop\n",
|
||||
"data: {\"type\":\"message_stop\"}\n\n",
|
||||
);
|
||||
|
||||
let aggregated =
|
||||
aggregate_claude_stream_sync_response(body.as_bytes()).expect("body should aggregate");
|
||||
|
||||
assert_eq!(aggregated["content"][0]["type"], "tool_use");
|
||||
assert_eq!(
|
||||
aggregated["content"][0]["input"],
|
||||
json!({
|
||||
"query": "",
|
||||
"pages": "",
|
||||
})
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn aggregates_gemini_stream_deltas_media_and_signatures_into_sync_body() {
|
||||
let body = concat!(
|
||||
@@ -3385,6 +3458,67 @@ mod tests {
|
||||
assert_eq!(body_json, provider_body_json);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn same_format_claude_sync_body_sanitizes_read_tool_input() {
|
||||
let report_context = json!({
|
||||
"provider_api_format": "claude:messages",
|
||||
"client_api_format": "claude:messages",
|
||||
"needs_conversion": false,
|
||||
});
|
||||
let provider_body_json = json!({
|
||||
"id": "msg_read",
|
||||
"type": "message",
|
||||
"role": "assistant",
|
||||
"model": "claude-sonnet-4-6",
|
||||
"content": [
|
||||
{
|
||||
"type": "tool_use",
|
||||
"id": "toolu_read",
|
||||
"name": "Read",
|
||||
"input": {
|
||||
"file_path": "/tmp/a.txt",
|
||||
"limit": 20,
|
||||
"pages": ""
|
||||
}
|
||||
},
|
||||
{
|
||||
"type": "tool_use",
|
||||
"id": "toolu_search",
|
||||
"name": "Search",
|
||||
"input": {
|
||||
"query": "",
|
||||
"pages": ""
|
||||
}
|
||||
}
|
||||
]
|
||||
});
|
||||
|
||||
let body_json = maybe_build_standard_same_format_sync_body_from_normalized_payload(
|
||||
"claude_chat_sync_finalize",
|
||||
200,
|
||||
Some(&report_context),
|
||||
Some(&provider_body_json),
|
||||
None,
|
||||
)
|
||||
.expect("same-format sync body should succeed")
|
||||
.expect("body should exist");
|
||||
|
||||
assert_eq!(
|
||||
body_json["content"][0]["input"],
|
||||
json!({
|
||||
"file_path": "/tmp/a.txt",
|
||||
"limit": 20,
|
||||
})
|
||||
);
|
||||
assert_eq!(
|
||||
body_json["content"][1]["input"],
|
||||
json!({
|
||||
"query": "",
|
||||
"pages": "",
|
||||
})
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn same_format_sync_response_restores_model_directive_display_model() {
|
||||
let report_context = json!({
|
||||
|
||||
@@ -20,6 +20,7 @@ use crate::formats::shared::stream_core::common::{
|
||||
use crate::formats::shared::stream_core::{
|
||||
CanonicalStreamFrame, StreamingStandardFormatMatrix, StreamingStandardTerminalObserver,
|
||||
};
|
||||
use crate::formats::shared::stream_rewrite::maybe_build_ai_surface_stream_rewriter;
|
||||
use crate::formats::shared::AiSurfaceFinalizeError;
|
||||
|
||||
pub struct SyncToStreamBridgeOutcome {
|
||||
@@ -668,7 +669,11 @@ fn maybe_bridge_aether_sse_response_capture_to_stream(
|
||||
client_api_format,
|
||||
);
|
||||
let sse_body = if captured_api_format == client_api_format {
|
||||
body_text.as_bytes().to_vec()
|
||||
if captured_api_format == "claude:messages" {
|
||||
sanitize_same_format_claude_sse_body(body_text.as_bytes(), report_context)?
|
||||
} else {
|
||||
body_text.as_bytes().to_vec()
|
||||
}
|
||||
} else {
|
||||
rewrite_sse_body_between_formats(
|
||||
body_text.as_bytes(),
|
||||
@@ -689,6 +694,34 @@ fn maybe_bridge_aether_sse_response_capture_to_stream(
|
||||
}))
|
||||
}
|
||||
|
||||
fn sanitize_same_format_claude_sse_body(
|
||||
body: &[u8],
|
||||
report_context: Option<&Value>,
|
||||
) -> Result<Vec<u8>, AiSurfaceFinalizeError> {
|
||||
let mut context = report_context
|
||||
.cloned()
|
||||
.filter(Value::is_object)
|
||||
.unwrap_or_else(|| json!({}));
|
||||
let object = context
|
||||
.as_object_mut()
|
||||
.expect("same-format Claude context should stay object");
|
||||
object.insert(
|
||||
"provider_api_format".to_string(),
|
||||
Value::String("claude:messages".to_string()),
|
||||
);
|
||||
object.insert(
|
||||
"client_api_format".to_string(),
|
||||
Value::String("claude:messages".to_string()),
|
||||
);
|
||||
|
||||
let Some(mut rewriter) = maybe_build_ai_surface_stream_rewriter(Some(&context)) else {
|
||||
return Ok(body.to_vec());
|
||||
};
|
||||
let mut out = rewriter.push_chunk(body)?;
|
||||
out.extend(rewriter.finish()?);
|
||||
Ok(out)
|
||||
}
|
||||
|
||||
fn response_capture_header<'a>(headers: &'a Map<String, Value>, name: &str) -> Option<&'a str> {
|
||||
headers
|
||||
.iter()
|
||||
@@ -1349,6 +1382,49 @@ mod tests {
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn rewrites_same_format_claude_capture_to_sanitize_read_tool_input() {
|
||||
let captured_body = concat!(
|
||||
"event: message_start\n",
|
||||
"data: {\"type\":\"message_start\",\"message\":{\"id\":\"msg_read_1\",\"type\":\"message\",\"role\":\"assistant\",\"model\":\"gpt-5.5\",\"content\":[],\"stop_reason\":null,\"stop_sequence\":null,\"usage\":{\"input_tokens\":0,\"output_tokens\":0}}}\n\n",
|
||||
"event: content_block_start\n",
|
||||
"data: {\"type\":\"content_block_start\",\"index\":0,\"content_block\":{\"type\":\"tool_use\",\"id\":\"call_read_1\",\"name\":\"Read\",\"input\":{\"file_path\":\"D:/projects/UIAutoTest/docs/prd/msr.md\",\"offset\":0,\"limit\":2000,\"pages\":\"\"}}}\n\n",
|
||||
"event: content_block_stop\n",
|
||||
"data: {\"type\":\"content_block_stop\",\"index\":0}\n\n",
|
||||
"event: content_block_start\n",
|
||||
"data: {\"type\":\"content_block_start\",\"index\":1,\"content_block\":{\"type\":\"server_tool_use\",\"id\":\"srv_1\",\"name\":\"web_search\",\"input\":{\"query\":\"rust\"}}}\n\n",
|
||||
"event: content_block_stop\n",
|
||||
"data: {\"type\":\"content_block_stop\",\"index\":1}\n\n",
|
||||
"event: message_delta\n",
|
||||
"data: {\"type\":\"message_delta\",\"delta\":{\"stop_reason\":\"tool_use\"},\"usage\":{\"input_tokens\":1,\"output_tokens\":2}}\n\n",
|
||||
"event: message_stop\n",
|
||||
"data: {\"type\":\"message_stop\"}\n\n",
|
||||
);
|
||||
let outcome = maybe_bridge_standard_sync_json_to_stream(
|
||||
&json!({
|
||||
"status_code": 200,
|
||||
"headers": {
|
||||
"content-type": "text/event-stream",
|
||||
"x-aether-control-endpoint-signature": "claude:messages"
|
||||
},
|
||||
"body": captured_body
|
||||
}),
|
||||
"openai:responses",
|
||||
"claude:messages",
|
||||
None,
|
||||
)
|
||||
.expect("bridge should succeed")
|
||||
.expect("capture should bridge");
|
||||
|
||||
let output = utf8(outcome.sse_body);
|
||||
assert!(output.contains("\"name\":\"Read\""));
|
||||
assert!(output.contains("\"limit\":2000"));
|
||||
assert!(output.contains("\"type\":\"server_tool_use\""));
|
||||
assert!(output.contains("\"name\":\"web_search\""));
|
||||
assert!(!output.contains("\"pages\":\"\""));
|
||||
assert!(!output.contains("\\\"pages\\\":\\\"\\\""));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn rewrites_aether_sse_response_capture_to_requested_client_stream() {
|
||||
let captured_body = concat!(
|
||||
|
||||
@@ -4,6 +4,7 @@ use serde::{Deserialize, Serialize};
|
||||
use serde_json::{json, Map, Value};
|
||||
|
||||
use crate::formats::openai::shared::map_thinking_budget_to_openai_reasoning_effort;
|
||||
use crate::formats::shared::response::remove_empty_pages_from_tool_input_value;
|
||||
|
||||
pub use crate::protocol::stream::{CanonicalStreamEvent, CanonicalStreamFrame};
|
||||
|
||||
@@ -3825,6 +3826,7 @@ pub(crate) fn canonical_block_to_claude(
|
||||
input,
|
||||
extensions,
|
||||
} => {
|
||||
let input = remove_empty_pages_from_tool_input_value(name, input);
|
||||
let mut out = Map::new();
|
||||
out.insert("type".to_string(), Value::String("tool_use".to_string()));
|
||||
out.insert(
|
||||
@@ -3832,7 +3834,7 @@ pub(crate) fn canonical_block_to_claude(
|
||||
Value::String(claude_compatible_tool_use_id(id)),
|
||||
);
|
||||
out.insert("name".to_string(), Value::String(name.clone()));
|
||||
out.insert("input".to_string(), input.clone());
|
||||
out.insert("input".to_string(), input);
|
||||
out.extend(namespace_extension_object(extensions, "claude", &out));
|
||||
Some(Some(Value::Object(out)))
|
||||
}
|
||||
@@ -5596,6 +5598,74 @@ mod tests {
|
||||
assert_eq!(rebuilt["service_tier"], "flex");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn openai_responses_to_claude_response_drops_empty_pages_only_for_read_tool() {
|
||||
let response = json!({
|
||||
"id": "resp_read_pages",
|
||||
"object": "response",
|
||||
"status": "completed",
|
||||
"model": "gpt-5.5",
|
||||
"output": [
|
||||
{
|
||||
"type": "function_call",
|
||||
"id": "call_read",
|
||||
"call_id": "call_read",
|
||||
"name": "Read",
|
||||
"arguments": "{\"file_path\":\"/tmp/a.txt\",\"offset\":0,\"limit\":20,\"pages\":\"\"}"
|
||||
},
|
||||
{
|
||||
"type": "function_call",
|
||||
"id": "call_search",
|
||||
"call_id": "call_search",
|
||||
"name": "Search",
|
||||
"arguments": "{\"query\":\"\",\"pages\":\"\"}"
|
||||
}
|
||||
],
|
||||
"usage": {
|
||||
"input_tokens": 1,
|
||||
"output_tokens": 1,
|
||||
"total_tokens": 2
|
||||
}
|
||||
});
|
||||
|
||||
let canonical =
|
||||
from_openai_responses_to_canonical_response(&response).expect("canonical response");
|
||||
let claude = canonical_to_claude_response(&canonical);
|
||||
|
||||
assert_eq!(
|
||||
claude["content"][0]["input"],
|
||||
json!({
|
||||
"file_path": "/tmp/a.txt",
|
||||
"offset": 0,
|
||||
"limit": 20,
|
||||
})
|
||||
);
|
||||
assert_eq!(
|
||||
claude["content"][1]["input"],
|
||||
json!({
|
||||
"query": "",
|
||||
"pages": "",
|
||||
})
|
||||
);
|
||||
|
||||
let rebuilt_responses = canonical_to_openai_responses_response(&canonical, &json!({}));
|
||||
let read_arguments = serde_json::from_str::<Value>(
|
||||
rebuilt_responses["output"][0]["arguments"]
|
||||
.as_str()
|
||||
.expect("arguments should be a string"),
|
||||
)
|
||||
.expect("arguments should be json");
|
||||
assert_eq!(
|
||||
read_arguments,
|
||||
json!({
|
||||
"file_path": "/tmp/a.txt",
|
||||
"offset": 0,
|
||||
"limit": 20,
|
||||
"pages": "",
|
||||
})
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn openai_responses_image_generation_call_becomes_canonical_image_block() {
|
||||
let response = json!({
|
||||
|
||||
Reference in New Issue
Block a user