Detailed changes
@@ -127,6 +127,16 @@ spec:
secretKeyRef:
name: google-ai
key: api_key
+ - name: QWEN2_7B_API_KEY
+ valueFrom:
+ secretKeyRef:
+ name: hugging-face
+ key: api_key
+ - name: QWEN2_7B_API_URL
+ valueFrom:
+ secretKeyRef:
+ name: hugging-face
+ key: qwen2_api_url
- name: BLOB_STORE_ACCESS_KEY
valueFrom:
secretKeyRef:
@@ -151,6 +151,8 @@ pub struct Config {
pub openai_api_key: Option<Arc<str>>,
pub google_ai_api_key: Option<Arc<str>>,
pub anthropic_api_key: Option<Arc<str>>,
+ pub qwen2_7b_api_key: Option<Arc<str>>,
+ pub qwen2_7b_api_url: Option<Arc<str>>,
pub zed_client_checksum_seed: Option<String>,
pub slack_panics_webhook: Option<String>,
pub auto_join_channel_id: Option<ChannelId>,
@@ -4706,6 +4706,30 @@ async fn stream_complete_with_language_model(
})?;
}
}
+ Some(proto::LanguageModelProvider::Zed) => {
+ let api_key = config
+ .qwen2_7b_api_key
+ .as_ref()
+ .context("no Qwen2-7B API key configured on the server")?;
+ let api_url = config
+ .qwen2_7b_api_url
+ .as_ref()
+ .context("no Qwen2-7B URL configured on the server")?;
+ let mut events = open_ai::stream_completion(
+ session.http_client.as_ref(),
+ &api_url,
+ api_key,
+ serde_json::from_str(&request.request)?,
+ None,
+ )
+ .await?;
+ while let Some(event) = events.next().await {
+ let event = event?;
+ response.send(proto::StreamCompleteWithLanguageModelResponse {
+ event: serde_json::to_string(&event)?,
+ })?;
+ }
+ }
None => return Err(anyhow!("unknown provider"))?,
}
@@ -672,6 +672,8 @@ impl TestServer {
stripe_api_key: None,
stripe_price_id: None,
supermaven_admin_api_key: None,
+ qwen2_7b_api_key: None,
+ qwen2_7b_api_url: None,
},
})
}
@@ -1,5 +1,6 @@
use schemars::JsonSchema;
use serde::{Deserialize, Serialize};
+use strum::EnumIter;
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize, JsonSchema)]
#[serde(tag = "provider", rename_all = "lowercase")]
@@ -7,6 +8,33 @@ pub enum CloudModel {
Anthropic(anthropic::Model),
OpenAi(open_ai::Model),
Google(google_ai::Model),
+ Zed(ZedModel),
+}
+
+#[derive(Clone, Debug, PartialEq, Serialize, Deserialize, JsonSchema, EnumIter)]
+pub enum ZedModel {
+ #[serde(rename = "qwen2-7b-instruct")]
+ Qwen2_7bInstruct,
+}
+
+impl ZedModel {
+ pub fn id(&self) -> &str {
+ match self {
+ ZedModel::Qwen2_7bInstruct => "qwen2-7b-instruct",
+ }
+ }
+
+ pub fn display_name(&self) -> &str {
+ match self {
+ ZedModel::Qwen2_7bInstruct => "Qwen2 7B Instruct",
+ }
+ }
+
+ pub fn max_token_count(&self) -> usize {
+ match self {
+ ZedModel::Qwen2_7bInstruct => 8192,
+ }
+ }
}
impl Default for CloudModel {
@@ -21,6 +49,7 @@ impl CloudModel {
CloudModel::Anthropic(model) => model.id(),
CloudModel::OpenAi(model) => model.id(),
CloudModel::Google(model) => model.id(),
+ CloudModel::Zed(model) => model.id(),
}
}
@@ -29,6 +58,7 @@ impl CloudModel {
CloudModel::Anthropic(model) => model.display_name(),
CloudModel::OpenAi(model) => model.display_name(),
CloudModel::Google(model) => model.display_name(),
+ CloudModel::Zed(model) => model.display_name(),
}
}
@@ -37,6 +67,7 @@ impl CloudModel {
CloudModel::Anthropic(model) => model.max_token_count(),
CloudModel::OpenAi(model) => model.max_token_count(),
CloudModel::Google(model) => model.max_token_count(),
+ CloudModel::Zed(model) => model.max_token_count(),
}
}
}
@@ -2,7 +2,7 @@ use super::open_ai::count_open_ai_tokens;
use crate::{
settings::AllLanguageModelSettings, CloudModel, LanguageModel, LanguageModelId,
LanguageModelName, LanguageModelProviderId, LanguageModelProviderName,
- LanguageModelProviderState, LanguageModelRequest, RateLimiter,
+ LanguageModelProviderState, LanguageModelRequest, RateLimiter, ZedModel,
};
use anyhow::{anyhow, Context as _, Result};
use client::{Client, UserStore};
@@ -146,6 +146,9 @@ impl LanguageModelProvider for CloudLanguageModelProvider {
models.insert(model.id().to_string(), CloudModel::Google(model));
}
}
+ for model in ZedModel::iter() {
+ models.insert(model.id().to_string(), CloudModel::Zed(model));
+ }
// Override with available models from settings
for model in &AllLanguageModelSettings::get_global(cx)
@@ -263,6 +266,9 @@ impl LanguageModel for CloudLanguageModel {
}
.boxed()
}
+ CloudModel::Zed(_) => {
+ count_open_ai_tokens(request, open_ai::Model::ThreePointFiveTurbo, cx)
+ }
}
}
@@ -323,6 +329,24 @@ impl LanguageModel for CloudLanguageModel {
});
async move { Ok(future.await?.boxed()) }.boxed()
}
+ CloudModel::Zed(model) => {
+ let client = self.client.clone();
+ let mut request = request.into_open_ai(model.id().into());
+ request.max_tokens = Some(4000);
+ let future = self.request_limiter.stream(async move {
+ let request = serde_json::to_string(&request)?;
+ let stream = client
+ .request_stream(proto::StreamCompleteWithLanguageModel {
+ provider: proto::LanguageModelProvider::Zed as i32,
+ request,
+ })
+ .await?;
+ Ok(open_ai::extract_text_from_events(
+ stream.map(|item| Ok(serde_json::from_str(&item?.event)?)),
+ ))
+ });
+ async move { Ok(future.await?.boxed()) }.boxed()
+ }
}
}
@@ -382,6 +406,9 @@ impl LanguageModel for CloudLanguageModel {
CloudModel::Google(_) => {
future::ready(Err(anyhow!("tool use not implemented for Google AI"))).boxed()
}
+ CloudModel::Zed(_) => {
+ future::ready(Err(anyhow!("tool use not implemented for Zed models"))).boxed()
+ }
}
}
}
@@ -37,6 +37,7 @@ impl LanguageModelRequest {
stream: true,
stop: self.stop,
temperature: self.temperature,
+ max_tokens: None,
tools: Vec::new(),
tool_choice: None,
}
@@ -116,6 +116,8 @@ pub struct Request {
pub model: String,
pub messages: Vec<RequestMessage>,
pub stream: bool,
+ #[serde(default, skip_serializing_if = "Option::is_none")]
+ pub max_tokens: Option<usize>,
pub stop: Vec<String>,
pub temperature: f32,
#[serde(default, skip_serializing_if = "Option::is_none")]
@@ -216,6 +218,13 @@ pub struct ChoiceDelta {
pub finish_reason: Option<String>,
}
+#[derive(Serialize, Deserialize, Debug)]
+#[serde(untagged)]
+pub enum ResponseStreamResult {
+ Ok(ResponseStreamEvent),
+ Err { error: String },
+}
+
#[derive(Serialize, Deserialize, Debug)]
pub struct ResponseStreamEvent {
pub created: u32,
@@ -256,7 +265,10 @@ pub async fn stream_completion(
None
} else {
match serde_json::from_str(line) {
- Ok(response) => Some(Ok(response)),
+ Ok(ResponseStreamResult::Ok(response)) => Some(Ok(response)),
+ Ok(ResponseStreamResult::Err { error }) => {
+ Some(Err(anyhow!(error)))
+ }
Err(error) => Some(Err(anyhow!(error))),
}
}
@@ -2099,6 +2099,7 @@ enum LanguageModelProvider {
Anthropic = 0;
OpenAI = 1;
Google = 2;
+ Zed = 3;
}
message GetCachedEmbeddings {