refactor(main): extract api_client.rs — AnthropicRuntimeClient and API helpers
- Created src/api_client.rs (310 lines) with AnthropicRuntimeClient struct,
impl blocks, resolve_cli_auth_source, request_ends_with_tool_result
- Removed ~425 lines from main.rs, added mod api_client + re-export
- Removed redundant format_user_visible_api_error / format_context_window_blocked_error
(already live in render.rs as pub(crate))
- Removed debug println!("args:") that broke --compact output
- All 214 tests pass
This commit is contained in:
parent
bd4c618332
commit
c01ae72e30
|
|
@ -73,9 +73,9 @@
|
|||
|
||||
## 13. Create api_client.rs — extract AnthropicRuntimeClient
|
||||
|
||||
- [ ] 13.1 Create `src/api_client.rs` with AnthropicRuntimeClient struct, its impl, ApiClient trait impl, `resolve_cli_auth_source`, `resolve_cli_auth_source_for_cwd`, `request_ends_with_tool_result`, `format_user_visible_api_error`, `format_context_window_blocked_error` (main.rs lines 7497-7912)
|
||||
- [ ] 13.2 Add `mod api_client;` to `main.rs` and re-export
|
||||
- [ ] 13.3 Run `cargo build` and `cargo test` — must pass
|
||||
- [x] 13.1 Create `src/api_client.rs` with AnthropicRuntimeClient struct, its impl, ApiClient trait impl, `resolve_cli_auth_source`, `resolve_cli_auth_source_for_cwd`, `request_ends_with_tool_result`, `format_user_visible_api_error`, `format_context_window_blocked_error` (main.rs lines 7497-7912)
|
||||
- [x] 13.2 Add `mod api_client;` to `main.rs` and re-export
|
||||
- [x] 13.3 Run `cargo build` and `cargo test` — must pass
|
||||
|
||||
## 14. Create tool_executor.rs — extract CliToolExecutor
|
||||
|
||||
|
|
|
|||
|
|
@ -0,0 +1,309 @@
|
|||
use std::io::{self, Write};
|
||||
|
||||
use api::{
|
||||
detect_provider_kind, resolve_startup_auth_source, AnthropicClient, AuthSource,
|
||||
ContentBlockDelta, MessageRequest, PromptCache, ProviderClient as ApiProviderClient,
|
||||
ProviderKind, StreamEvent as ApiStreamEvent, ToolChoice,
|
||||
};
|
||||
use crate::render::{MarkdownStreamState, TerminalRenderer};
|
||||
use runtime::{ApiClient, ApiRequest, AssistantEvent, MessageRole, RuntimeError};
|
||||
use tools::GlobalToolRegistry;
|
||||
|
||||
use crate::constants::*;
|
||||
use crate::model::max_tokens_for_model;
|
||||
use crate::render::{
|
||||
format_tool_call_start, format_user_visible_api_error, push_output_block,
|
||||
render_thinking_block_summary,
|
||||
};
|
||||
use crate::InternalPromptProgressReporter;
|
||||
|
||||
// NOTE: Despite the historical name `AnthropicRuntimeClient`, this struct
|
||||
// now holds an `ApiProviderClient` which dispatches to Anthropic, xAI,
|
||||
// OpenAI, or DashScope at construction time based on
|
||||
// `detect_provider_kind(&model)`. The struct name is kept to avoid
|
||||
// churning `BuiltRuntime` and every Deref/DerefMut site that references
|
||||
// it. See ROADMAP #29 for the provider-dispatch routing fix.
|
||||
pub(crate) struct AnthropicRuntimeClient {
|
||||
runtime: tokio::runtime::Runtime,
|
||||
client: ApiProviderClient,
|
||||
session_id: String,
|
||||
model: String,
|
||||
enable_tools: bool,
|
||||
emit_output: bool,
|
||||
allowed_tools: Option<AllowedToolSet>,
|
||||
tool_registry: GlobalToolRegistry,
|
||||
progress_reporter: Option<InternalPromptProgressReporter>,
|
||||
reasoning_effort: Option<String>,
|
||||
}
|
||||
|
||||
impl AnthropicRuntimeClient {
|
||||
pub(crate) fn new(
|
||||
session_id: &str,
|
||||
model: String,
|
||||
enable_tools: bool,
|
||||
emit_output: bool,
|
||||
allowed_tools: Option<AllowedToolSet>,
|
||||
tool_registry: GlobalToolRegistry,
|
||||
progress_reporter: Option<InternalPromptProgressReporter>,
|
||||
) -> Result<Self, Box<dyn std::error::Error>> {
|
||||
let resolved_model = api::resolve_model_alias(&model);
|
||||
let client = match detect_provider_kind(&resolved_model) {
|
||||
ProviderKind::Anthropic => {
|
||||
let auth = resolve_cli_auth_source()?;
|
||||
let inner = AnthropicClient::from_auth(auth)
|
||||
.with_base_url(api::read_base_url())
|
||||
.with_prompt_cache(PromptCache::new(session_id));
|
||||
ApiProviderClient::Anthropic(inner)
|
||||
}
|
||||
ProviderKind::Xai | ProviderKind::OpenAi => {
|
||||
ApiProviderClient::from_model_with_anthropic_auth(&resolved_model, None)?
|
||||
}
|
||||
};
|
||||
Ok(Self {
|
||||
runtime: tokio::runtime::Runtime::new()?,
|
||||
client,
|
||||
session_id: session_id.to_string(),
|
||||
model,
|
||||
enable_tools,
|
||||
emit_output,
|
||||
allowed_tools,
|
||||
tool_registry,
|
||||
progress_reporter,
|
||||
reasoning_effort: None,
|
||||
})
|
||||
}
|
||||
|
||||
pub(crate) fn set_reasoning_effort(&mut self, effort: Option<String>) {
|
||||
self.reasoning_effort = effort;
|
||||
}
|
||||
}
|
||||
|
||||
impl ApiClient for AnthropicRuntimeClient {
|
||||
#[allow(clippy::too_many_lines)]
|
||||
fn stream(&mut self, request: ApiRequest) -> Result<Vec<AssistantEvent>, RuntimeError> {
|
||||
if let Some(progress_reporter) = &self.progress_reporter {
|
||||
progress_reporter.mark_model_phase();
|
||||
}
|
||||
let is_post_tool = request_ends_with_tool_result(&request);
|
||||
let message_request = MessageRequest {
|
||||
model: self.model.clone(),
|
||||
max_tokens: max_tokens_for_model(&self.model),
|
||||
messages: crate::convert_messages(&request.messages),
|
||||
system: (!request.system_prompt.is_empty()).then(|| request.system_prompt.join("\n\n")),
|
||||
tools: self
|
||||
.enable_tools
|
||||
.then(|| crate::filter_tool_specs(&self.tool_registry, self.allowed_tools.as_ref())),
|
||||
tool_choice: self.enable_tools.then_some(ToolChoice::Auto),
|
||||
stream: true,
|
||||
reasoning_effort: self.reasoning_effort.clone(),
|
||||
..Default::default()
|
||||
};
|
||||
|
||||
self.runtime.block_on(async {
|
||||
let max_attempts: usize = if is_post_tool { 2 } else { 1 };
|
||||
|
||||
for attempt in 1..=max_attempts {
|
||||
let result = self
|
||||
.consume_stream(&message_request, is_post_tool && attempt == 1)
|
||||
.await;
|
||||
match result {
|
||||
Ok(events) => return Ok(events),
|
||||
Err(error)
|
||||
if error.to_string().contains("post-tool stall")
|
||||
&& attempt < max_attempts =>
|
||||
{
|
||||
}
|
||||
Err(error) => return Err(error),
|
||||
}
|
||||
}
|
||||
|
||||
Err(RuntimeError::new("post-tool continuation nudge exhausted"))
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
impl AnthropicRuntimeClient {
|
||||
#[allow(clippy::too_many_lines)]
|
||||
async fn consume_stream(
|
||||
&self,
|
||||
message_request: &MessageRequest,
|
||||
apply_stall_timeout: bool,
|
||||
) -> Result<Vec<AssistantEvent>, RuntimeError> {
|
||||
let mut stream = self
|
||||
.client
|
||||
.stream_message(message_request)
|
||||
.await
|
||||
.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?;
|
||||
let mut stdout = io::stdout();
|
||||
let mut sink = io::sink();
|
||||
let out: &mut dyn Write = if self.emit_output {
|
||||
&mut stdout
|
||||
} else {
|
||||
&mut sink
|
||||
};
|
||||
let renderer = TerminalRenderer::new();
|
||||
let mut markdown_stream = MarkdownStreamState::default();
|
||||
let mut events = Vec::new();
|
||||
let mut pending_tool: Option<(String, String, String)> = None;
|
||||
let mut block_has_thinking_summary = false;
|
||||
let mut saw_stop = false;
|
||||
let mut received_any_event = false;
|
||||
|
||||
loop {
|
||||
let next = if apply_stall_timeout && !received_any_event {
|
||||
match tokio::time::timeout(POST_TOOL_STALL_TIMEOUT, stream.next_event()).await {
|
||||
Ok(inner) => inner.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?,
|
||||
Err(_elapsed) => {
|
||||
return Err(RuntimeError::new(
|
||||
"post-tool stall: model did not respond within timeout",
|
||||
));
|
||||
}
|
||||
}
|
||||
} else {
|
||||
stream.next_event().await.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?
|
||||
};
|
||||
|
||||
let Some(event) = next else {
|
||||
break;
|
||||
};
|
||||
received_any_event = true;
|
||||
|
||||
match event {
|
||||
ApiStreamEvent::MessageStart(start) => {
|
||||
for block in start.message.content {
|
||||
push_output_block(
|
||||
block,
|
||||
out,
|
||||
&mut events,
|
||||
&mut pending_tool,
|
||||
true,
|
||||
&mut block_has_thinking_summary,
|
||||
)?;
|
||||
}
|
||||
}
|
||||
ApiStreamEvent::ContentBlockStart(start) => {
|
||||
push_output_block(
|
||||
start.content_block,
|
||||
out,
|
||||
&mut events,
|
||||
&mut pending_tool,
|
||||
true,
|
||||
&mut block_has_thinking_summary,
|
||||
)?;
|
||||
}
|
||||
ApiStreamEvent::ContentBlockDelta(delta) => match delta.delta {
|
||||
ContentBlockDelta::TextDelta { text } => {
|
||||
if !text.is_empty() {
|
||||
if let Some(progress_reporter) = &self.progress_reporter {
|
||||
progress_reporter.mark_text_phase(&text);
|
||||
}
|
||||
if let Some(rendered) = markdown_stream.push(&renderer, &text) {
|
||||
write!(out, "{rendered}")
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
}
|
||||
events.push(AssistantEvent::TextDelta(text));
|
||||
}
|
||||
}
|
||||
ContentBlockDelta::InputJsonDelta { partial_json } => {
|
||||
if let Some((_, _, input)) = &mut pending_tool {
|
||||
input.push_str(&partial_json);
|
||||
}
|
||||
}
|
||||
ContentBlockDelta::ThinkingDelta { thinking } => {
|
||||
if !block_has_thinking_summary {
|
||||
render_thinking_block_summary(out, None, false)?;
|
||||
block_has_thinking_summary = true;
|
||||
}
|
||||
events.push(AssistantEvent::ThinkingDelta(thinking));
|
||||
}
|
||||
ContentBlockDelta::SignatureDelta { signature } => {
|
||||
events.push(AssistantEvent::SignatureDelta(signature));
|
||||
}
|
||||
},
|
||||
ApiStreamEvent::ContentBlockStop(_) => {
|
||||
block_has_thinking_summary = false;
|
||||
if let Some(rendered) = markdown_stream.flush(&renderer) {
|
||||
write!(out, "{rendered}")
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
}
|
||||
if let Some((id, name, input)) = pending_tool.take() {
|
||||
if let Some(progress_reporter) = &self.progress_reporter {
|
||||
progress_reporter.mark_tool_phase(&name, &input);
|
||||
}
|
||||
writeln!(out, "\n{}", format_tool_call_start(&name, &input))
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
events.push(AssistantEvent::ToolUse { id, name, input });
|
||||
}
|
||||
}
|
||||
ApiStreamEvent::MessageDelta(delta) => {
|
||||
events.push(AssistantEvent::Usage(delta.usage.token_usage()));
|
||||
}
|
||||
ApiStreamEvent::MessageStop(_) => {
|
||||
saw_stop = true;
|
||||
if let Some(rendered) = markdown_stream.flush(&renderer) {
|
||||
write!(out, "{rendered}")
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
}
|
||||
events.push(AssistantEvent::MessageStop);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
crate::push_prompt_cache_record(&self.client, &mut events);
|
||||
|
||||
if !saw_stop
|
||||
&& events.iter().any(|event| {
|
||||
matches!(event, AssistantEvent::TextDelta(text) if !text.is_empty())
|
||||
|| matches!(event, AssistantEvent::ToolUse { .. })
|
||||
})
|
||||
{
|
||||
events.push(AssistantEvent::MessageStop);
|
||||
}
|
||||
|
||||
if events
|
||||
.iter()
|
||||
.any(|event| matches!(event, AssistantEvent::MessageStop))
|
||||
{
|
||||
return Ok(events);
|
||||
}
|
||||
|
||||
let response = self
|
||||
.client
|
||||
.send_message(&MessageRequest {
|
||||
stream: false,
|
||||
..message_request.clone()
|
||||
})
|
||||
.await
|
||||
.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?;
|
||||
let mut events = crate::response_to_events(response, out)?;
|
||||
crate::push_prompt_cache_record(&self.client, &mut events);
|
||||
Ok(events)
|
||||
}
|
||||
}
|
||||
|
||||
pub(crate) fn resolve_cli_auth_source() -> Result<AuthSource, Box<dyn std::error::Error>> {
|
||||
Ok(resolve_cli_auth_source_for_cwd()?)
|
||||
}
|
||||
|
||||
pub(crate) fn resolve_cli_auth_source_for_cwd() -> Result<AuthSource, api::ApiError> {
|
||||
resolve_startup_auth_source(|| Ok(None))
|
||||
}
|
||||
|
||||
pub(crate) fn request_ends_with_tool_result(request: &ApiRequest) -> bool {
|
||||
request
|
||||
.messages
|
||||
.last()
|
||||
.is_some_and(|message| message.role == MessageRole::Tool)
|
||||
}
|
||||
|
|
@ -21,10 +21,12 @@ mod input;
|
|||
mod model;
|
||||
mod models;
|
||||
mod permission;
|
||||
mod api_client;
|
||||
mod progress;
|
||||
mod repl;
|
||||
mod render;
|
||||
|
||||
pub(crate) use api_client::*;
|
||||
pub(crate) use args::*;
|
||||
pub(crate) use constants::*;
|
||||
pub(crate) use doctor::*;
|
||||
|
|
@ -229,8 +231,6 @@ fn merge_prompt_with_stdin(prompt: &str, stdin_content: Option<&str>) -> String
|
|||
|
||||
fn run() -> Result<(), Box<dyn std::error::Error>> {
|
||||
let args: Vec<String> = env::args().skip(1).collect();
|
||||
// 添加打印args 的日志
|
||||
println!("args: {:?}", args);
|
||||
|
||||
match parse_args(&args)? {
|
||||
// 导出所有工具/插件的 manifest 清单文件到指定目录
|
||||
|
|
@ -3092,427 +3092,7 @@ fn build_runtime_with_plugin_state(
|
|||
Ok(BuiltRuntime::new(runtime, plugin_registry, mcp_state))
|
||||
}
|
||||
|
||||
// NOTE: Despite the historical name `AnthropicRuntimeClient`, this struct
|
||||
// now holds an `ApiProviderClient` which dispatches to Anthropic, xAI,
|
||||
// OpenAI, or DashScope at construction time based on
|
||||
// `detect_provider_kind(&model)`. The struct name is kept to avoid
|
||||
// churning `BuiltRuntime` and every Deref/DerefMut site that references
|
||||
// it. See ROADMAP #29 for the provider-dispatch routing fix.
|
||||
struct AnthropicRuntimeClient {
|
||||
runtime: tokio::runtime::Runtime,
|
||||
client: ApiProviderClient,
|
||||
session_id: String,
|
||||
model: String,
|
||||
enable_tools: bool,
|
||||
emit_output: bool,
|
||||
allowed_tools: Option<AllowedToolSet>,
|
||||
tool_registry: GlobalToolRegistry,
|
||||
progress_reporter: Option<InternalPromptProgressReporter>,
|
||||
reasoning_effort: Option<String>,
|
||||
}
|
||||
|
||||
impl AnthropicRuntimeClient {
|
||||
fn new(
|
||||
session_id: &str,
|
||||
model: String,
|
||||
enable_tools: bool,
|
||||
emit_output: bool,
|
||||
allowed_tools: Option<AllowedToolSet>,
|
||||
tool_registry: GlobalToolRegistry,
|
||||
progress_reporter: Option<InternalPromptProgressReporter>,
|
||||
) -> Result<Self, Box<dyn std::error::Error>> {
|
||||
// Dispatch to the correct provider at construction time.
|
||||
// `ApiProviderClient` (exposed by the api crate as
|
||||
// `ProviderClient`) is an enum over Anthropic / xAI / OpenAI
|
||||
// variants, where xAI and OpenAI both use the OpenAI-compat
|
||||
// wire format under the hood. We consult
|
||||
// `detect_provider_kind(&resolved_model)` so model-name prefix
|
||||
// routing (`openai/`, `gpt-`, `grok`, `qwen/`) wins over
|
||||
// env-var presence.
|
||||
//
|
||||
// For Anthropic we build the client directly instead of going
|
||||
// through `ApiProviderClient::from_model_with_anthropic_auth`
|
||||
// so we can explicitly apply `api::read_base_url()` — that
|
||||
// reads `ANTHROPIC_BASE_URL` and is required for the local
|
||||
// mock-server test harness
|
||||
// (`crates/rusty-claude-cli/tests/compact_output.rs`) to point
|
||||
// claw at its fake Anthropic endpoint. We also attach a
|
||||
// session-scoped prompt cache on the Anthropic path; the
|
||||
// prompt cache is Anthropic-only so non-Anthropic variants
|
||||
// skip it.
|
||||
let resolved_model = api::resolve_model_alias(&model);
|
||||
let client = match detect_provider_kind(&resolved_model) {
|
||||
ProviderKind::Anthropic => {
|
||||
let auth = resolve_cli_auth_source()?;
|
||||
let inner = AnthropicClient::from_auth(auth)
|
||||
.with_base_url(api::read_base_url())
|
||||
.with_prompt_cache(PromptCache::new(session_id));
|
||||
ApiProviderClient::Anthropic(inner)
|
||||
}
|
||||
ProviderKind::Xai | ProviderKind::OpenAi => {
|
||||
// The api crate's `ProviderClient::from_model_with_anthropic_auth`
|
||||
// with `None` for the anthropic auth routes via
|
||||
// `detect_provider_kind` and builds an
|
||||
// `OpenAiCompatClient::from_env` with the matching
|
||||
// `OpenAiCompatConfig` (openai / xai / dashscope).
|
||||
// That reads the correct API-key env var and BASE_URL
|
||||
// override internally, so this one call covers OpenAI,
|
||||
// OpenRouter, xAI, DashScope, Ollama, and any other
|
||||
// OpenAI-compat endpoint users configure via
|
||||
// `OPENAI_BASE_URL` / `XAI_BASE_URL` / `DASHSCOPE_BASE_URL`.
|
||||
ApiProviderClient::from_model_with_anthropic_auth(&resolved_model, None)?
|
||||
}
|
||||
};
|
||||
Ok(Self {
|
||||
runtime: tokio::runtime::Runtime::new()?,
|
||||
client,
|
||||
session_id: session_id.to_string(),
|
||||
model,
|
||||
enable_tools,
|
||||
emit_output,
|
||||
allowed_tools,
|
||||
tool_registry,
|
||||
progress_reporter,
|
||||
reasoning_effort: None,
|
||||
})
|
||||
}
|
||||
|
||||
fn set_reasoning_effort(&mut self, effort: Option<String>) {
|
||||
self.reasoning_effort = effort;
|
||||
}
|
||||
}
|
||||
|
||||
fn resolve_cli_auth_source() -> Result<AuthSource, Box<dyn std::error::Error>> {
|
||||
Ok(resolve_cli_auth_source_for_cwd()?)
|
||||
}
|
||||
|
||||
fn resolve_cli_auth_source_for_cwd() -> Result<AuthSource, api::ApiError> {
|
||||
resolve_startup_auth_source(|| Ok(None))
|
||||
}
|
||||
|
||||
impl ApiClient for AnthropicRuntimeClient {
|
||||
#[allow(clippy::too_many_lines)]
|
||||
fn stream(&mut self, request: ApiRequest) -> Result<Vec<AssistantEvent>, RuntimeError> {
|
||||
if let Some(progress_reporter) = &self.progress_reporter {
|
||||
progress_reporter.mark_model_phase();
|
||||
}
|
||||
let is_post_tool = request_ends_with_tool_result(&request);
|
||||
let message_request = MessageRequest {
|
||||
model: self.model.clone(),
|
||||
max_tokens: max_tokens_for_model(&self.model),
|
||||
messages: convert_messages(&request.messages),
|
||||
system: (!request.system_prompt.is_empty()).then(|| request.system_prompt.join("\n\n")),
|
||||
tools: self
|
||||
.enable_tools
|
||||
.then(|| filter_tool_specs(&self.tool_registry, self.allowed_tools.as_ref())),
|
||||
tool_choice: self.enable_tools.then_some(ToolChoice::Auto),
|
||||
stream: true,
|
||||
reasoning_effort: self.reasoning_effort.clone(),
|
||||
..Default::default()
|
||||
};
|
||||
|
||||
self.runtime.block_on(async {
|
||||
// When resuming after tool execution, apply a stall timeout on the
|
||||
// first stream event. If the model does not respond within the
|
||||
// deadline we drop the stalled connection and re-send the request as
|
||||
// a continuation nudge (one retry only).
|
||||
let max_attempts: usize = if is_post_tool { 2 } else { 1 };
|
||||
|
||||
for attempt in 1..=max_attempts {
|
||||
let result = self
|
||||
.consume_stream(&message_request, is_post_tool && attempt == 1)
|
||||
.await;
|
||||
match result {
|
||||
Ok(events) => return Ok(events),
|
||||
Err(error)
|
||||
if error.to_string().contains("post-tool stall")
|
||||
&& attempt < max_attempts =>
|
||||
{
|
||||
// Stalled after tool completion — nudge the model by
|
||||
// re-sending the same request.
|
||||
}
|
||||
Err(error) => return Err(error),
|
||||
}
|
||||
}
|
||||
|
||||
Err(RuntimeError::new("post-tool continuation nudge exhausted"))
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
impl AnthropicRuntimeClient {
|
||||
/// Consume a single streaming response, optionally applying a stall
|
||||
/// timeout on the first event for post-tool continuations.
|
||||
#[allow(clippy::too_many_lines)]
|
||||
async fn consume_stream(
|
||||
&self,
|
||||
message_request: &MessageRequest,
|
||||
apply_stall_timeout: bool,
|
||||
) -> Result<Vec<AssistantEvent>, RuntimeError> {
|
||||
let mut stream = self
|
||||
.client
|
||||
.stream_message(message_request)
|
||||
.await
|
||||
.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?;
|
||||
let mut stdout = io::stdout();
|
||||
let mut sink = io::sink();
|
||||
let out: &mut dyn Write = if self.emit_output {
|
||||
&mut stdout
|
||||
} else {
|
||||
&mut sink
|
||||
};
|
||||
let renderer = TerminalRenderer::new();
|
||||
let mut markdown_stream = MarkdownStreamState::default();
|
||||
let mut events = Vec::new();
|
||||
let mut pending_tool: Option<(String, String, String)> = None;
|
||||
let mut block_has_thinking_summary = false;
|
||||
let mut saw_stop = false;
|
||||
let mut received_any_event = false;
|
||||
|
||||
loop {
|
||||
let next = if apply_stall_timeout && !received_any_event {
|
||||
match tokio::time::timeout(POST_TOOL_STALL_TIMEOUT, stream.next_event()).await {
|
||||
Ok(inner) => inner.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?,
|
||||
Err(_elapsed) => {
|
||||
return Err(RuntimeError::new(
|
||||
"post-tool stall: model did not respond within timeout",
|
||||
));
|
||||
}
|
||||
}
|
||||
} else {
|
||||
stream.next_event().await.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?
|
||||
};
|
||||
|
||||
let Some(event) = next else {
|
||||
break;
|
||||
};
|
||||
received_any_event = true;
|
||||
|
||||
match event {
|
||||
ApiStreamEvent::MessageStart(start) => {
|
||||
for block in start.message.content {
|
||||
push_output_block(
|
||||
block,
|
||||
out,
|
||||
&mut events,
|
||||
&mut pending_tool,
|
||||
true,
|
||||
&mut block_has_thinking_summary,
|
||||
)?;
|
||||
}
|
||||
}
|
||||
ApiStreamEvent::ContentBlockStart(start) => {
|
||||
push_output_block(
|
||||
start.content_block,
|
||||
out,
|
||||
&mut events,
|
||||
&mut pending_tool,
|
||||
true,
|
||||
&mut block_has_thinking_summary,
|
||||
)?;
|
||||
}
|
||||
ApiStreamEvent::ContentBlockDelta(delta) => match delta.delta {
|
||||
ContentBlockDelta::TextDelta { text } => {
|
||||
if !text.is_empty() {
|
||||
if let Some(progress_reporter) = &self.progress_reporter {
|
||||
progress_reporter.mark_text_phase(&text);
|
||||
}
|
||||
if let Some(rendered) = markdown_stream.push(&renderer, &text) {
|
||||
write!(out, "{rendered}")
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
}
|
||||
events.push(AssistantEvent::TextDelta(text));
|
||||
}
|
||||
}
|
||||
ContentBlockDelta::InputJsonDelta { partial_json } => {
|
||||
if let Some((_, _, input)) = &mut pending_tool {
|
||||
input.push_str(&partial_json);
|
||||
}
|
||||
}
|
||||
ContentBlockDelta::ThinkingDelta { thinking } => {
|
||||
if !block_has_thinking_summary {
|
||||
render_thinking_block_summary(out, None, false)?;
|
||||
block_has_thinking_summary = true;
|
||||
}
|
||||
events.push(AssistantEvent::ThinkingDelta(thinking));
|
||||
}
|
||||
ContentBlockDelta::SignatureDelta { signature } => {
|
||||
events.push(AssistantEvent::SignatureDelta(signature));
|
||||
}
|
||||
},
|
||||
ApiStreamEvent::ContentBlockStop(_) => {
|
||||
block_has_thinking_summary = false;
|
||||
if let Some(rendered) = markdown_stream.flush(&renderer) {
|
||||
write!(out, "{rendered}")
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
}
|
||||
if let Some((id, name, input)) = pending_tool.take() {
|
||||
if let Some(progress_reporter) = &self.progress_reporter {
|
||||
progress_reporter.mark_tool_phase(&name, &input);
|
||||
}
|
||||
// Display tool call now that input is fully accumulated
|
||||
writeln!(out, "\n{}", format_tool_call_start(&name, &input))
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
events.push(AssistantEvent::ToolUse { id, name, input });
|
||||
}
|
||||
}
|
||||
ApiStreamEvent::MessageDelta(delta) => {
|
||||
events.push(AssistantEvent::Usage(delta.usage.token_usage()));
|
||||
}
|
||||
ApiStreamEvent::MessageStop(_) => {
|
||||
saw_stop = true;
|
||||
if let Some(rendered) = markdown_stream.flush(&renderer) {
|
||||
write!(out, "{rendered}")
|
||||
.and_then(|()| out.flush())
|
||||
.map_err(|error| RuntimeError::new(error.to_string()))?;
|
||||
}
|
||||
events.push(AssistantEvent::MessageStop);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
push_prompt_cache_record(&self.client, &mut events);
|
||||
|
||||
if !saw_stop
|
||||
&& events.iter().any(|event| {
|
||||
matches!(event, AssistantEvent::TextDelta(text) if !text.is_empty())
|
||||
|| matches!(event, AssistantEvent::ToolUse { .. })
|
||||
})
|
||||
{
|
||||
events.push(AssistantEvent::MessageStop);
|
||||
}
|
||||
|
||||
if events
|
||||
.iter()
|
||||
.any(|event| matches!(event, AssistantEvent::MessageStop))
|
||||
{
|
||||
return Ok(events);
|
||||
}
|
||||
|
||||
let response = self
|
||||
.client
|
||||
.send_message(&MessageRequest {
|
||||
stream: false,
|
||||
..message_request.clone()
|
||||
})
|
||||
.await
|
||||
.map_err(|error| {
|
||||
RuntimeError::new(format_user_visible_api_error(&self.session_id, &error))
|
||||
})?;
|
||||
let mut events = response_to_events(response, out)?;
|
||||
push_prompt_cache_record(&self.client, &mut events);
|
||||
Ok(events)
|
||||
}
|
||||
}
|
||||
|
||||
/// Returns `true` when the conversation ends with a tool-result message,
|
||||
/// meaning the model is expected to continue after tool execution.
|
||||
fn request_ends_with_tool_result(request: &ApiRequest) -> bool {
|
||||
request
|
||||
.messages
|
||||
.last()
|
||||
.is_some_and(|message| message.role == MessageRole::Tool)
|
||||
}
|
||||
|
||||
fn format_user_visible_api_error(session_id: &str, error: &api::ApiError) -> String {
|
||||
if error.is_context_window_failure() {
|
||||
format_context_window_blocked_error(session_id, error)
|
||||
} else if error.is_generic_fatal_wrapper() {
|
||||
let mut qualifiers = vec![format!("session {session_id}")];
|
||||
if let Some(request_id) = error.request_id() {
|
||||
qualifiers.push(format!("trace {request_id}"));
|
||||
}
|
||||
format!(
|
||||
"{} ({}): {}",
|
||||
error.safe_failure_class(),
|
||||
qualifiers.join(", "),
|
||||
error
|
||||
)
|
||||
} else {
|
||||
error.to_string()
|
||||
}
|
||||
}
|
||||
|
||||
fn format_context_window_blocked_error(session_id: &str, error: &api::ApiError) -> String {
|
||||
let mut lines = vec![
|
||||
"Context window blocked".to_string(),
|
||||
" Failure class context_window_blocked".to_string(),
|
||||
format!(" Session {session_id}"),
|
||||
];
|
||||
|
||||
if let Some(request_id) = error.request_id() {
|
||||
lines.push(format!(" Trace {request_id}"));
|
||||
}
|
||||
|
||||
match error {
|
||||
api::ApiError::ContextWindowExceeded {
|
||||
model,
|
||||
estimated_input_tokens,
|
||||
requested_output_tokens,
|
||||
estimated_total_tokens,
|
||||
context_window_tokens,
|
||||
} => {
|
||||
lines.push(format!(" Model {model}"));
|
||||
lines.push(format!(
|
||||
" Input estimate ~{estimated_input_tokens} tokens (heuristic)"
|
||||
));
|
||||
lines.push(format!(
|
||||
" Requested output {requested_output_tokens} tokens"
|
||||
));
|
||||
lines.push(format!(
|
||||
" Total estimate ~{estimated_total_tokens} tokens (heuristic)"
|
||||
));
|
||||
lines.push(format!(" Context window {context_window_tokens} tokens"));
|
||||
}
|
||||
api::ApiError::Api { message, body, .. } => {
|
||||
let detail = message.as_deref().unwrap_or(body).trim();
|
||||
if !detail.is_empty() {
|
||||
lines.push(format!(
|
||||
" Detail {}",
|
||||
truncate_for_summary(detail, 120)
|
||||
));
|
||||
}
|
||||
}
|
||||
api::ApiError::RetriesExhausted { last_error, .. } => {
|
||||
let detail = match last_error.as_ref() {
|
||||
api::ApiError::Api { message, body, .. } => message.as_deref().unwrap_or(body),
|
||||
other => return format_context_window_blocked_error(session_id, other),
|
||||
}
|
||||
.trim();
|
||||
if !detail.is_empty() {
|
||||
lines.push(format!(
|
||||
" Detail {}",
|
||||
truncate_for_summary(detail, 120)
|
||||
));
|
||||
}
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
|
||||
lines.push(String::new());
|
||||
lines.push("Recovery".to_string());
|
||||
lines.push(" Compact /compact".to_string());
|
||||
lines.push(format!(
|
||||
" Resume compact claw --resume {session_id} /compact"
|
||||
));
|
||||
lines.push(" Fresh session /clear --confirm".to_string());
|
||||
lines.push(
|
||||
" Reduce scope remove large pasted context/files or ask for a smaller slice"
|
||||
.to_string(),
|
||||
);
|
||||
lines.push(" Retry rerun after compacting or reducing the request".to_string());
|
||||
|
||||
lines.join("\n")
|
||||
}
|
||||
|
||||
fn final_assistant_text(summary: &runtime::TurnSummary) -> String {
|
||||
summary
|
||||
|
|
@ -4576,12 +4156,12 @@ mod tests {
|
|||
format_connected_line, format_cost_report, format_history_timestamp,
|
||||
format_internal_prompt_progress_line, format_issue_report, format_model_report,
|
||||
format_model_switch_report, format_permissions_report, format_permissions_switch_report,
|
||||
format_pr_report, format_resume_report, format_status_report, format_tool_call_start,
|
||||
format_pr_report, format_resume_report, format_status_report,
|
||||
format_tool_result, format_ultraplan_report, format_unknown_slash_command,
|
||||
format_unknown_slash_command_message, format_user_visible_api_error,
|
||||
format_unknown_slash_command_message,
|
||||
merge_prompt_with_stdin, normalize_permission_mode, parse_args, parse_export_args,
|
||||
parse_git_status_branch, parse_git_status_metadata_for, parse_git_workspace_summary,
|
||||
parse_history_count, permission_policy, print_help_to, push_output_block,
|
||||
parse_history_count, permission_policy, print_help_to,
|
||||
render_config_report, render_diff_report, render_diff_report_for, render_help_topic,
|
||||
render_memory_report, render_prompt_history_report, render_repl_help, render_resume_usage,
|
||||
render_session_markdown, resolve_model_alias, resolve_model_alias_with_config,
|
||||
|
|
@ -4595,6 +4175,9 @@ mod tests {
|
|||
STUB_COMMANDS,
|
||||
};
|
||||
use api::{ApiError, MessageResponse, OutputContentBlock, Usage};
|
||||
use crate::render::{
|
||||
format_tool_call_start, format_user_visible_api_error, push_output_block,
|
||||
};
|
||||
use plugins::{
|
||||
PluginManager, PluginManagerConfig, PluginTool, PluginToolDefinition, PluginToolPermission,
|
||||
};
|
||||
|
|
|
|||
Loading…
Reference in New Issue