Send up Zed version with edit prediction and completion requests (#30136)

This PR makes it so we send up an `x-zed-version` header with the
client's version when making a request to llm.zed.dev for edit
predictions and completions.

Release Notes:

- N/A

crates/language_models/Cargo.toml

@@ -41,6 +41,7 @@ open_ai = { workspace = true, features = ["schemars"] }
 partial-json-fixer.workspace = true
 project.workspace = true
 proto.workspace = true
+release_channel.workspace = true
 schemars.workspace = true
 serde.workspace = true
 serde_json.workspace = true

crates/language_models/src/provider/cloud.rs

@@ -6,7 +6,9 @@ use feature_flags::{FeatureFlagAppExt, LlmClosedBetaFeatureFlag, ZedProFeatureFl
 use futures::{
     AsyncBufReadExt, FutureExt, Stream, StreamExt, future::BoxFuture, stream::BoxStream,
 };
-use gpui::{AnyElement, AnyView, App, AsyncApp, Context, Entity, Subscription, Task};
+use gpui::{
+    AnyElement, AnyView, App, AsyncApp, Context, Entity, SemanticVersion, Subscription, Task,
+};
 use http_client::{AsyncBody, HttpClient, Method, Response, StatusCode};
 use language_model::{
     AuthenticateError, CloudModel, LanguageModel, LanguageModelCacheConfiguration,
@@ -20,6 +22,7 @@ use language_model::{
     MaxMonthlySpendReachedError, PaymentRequiredError, RefreshLlmTokenListener,
 };
 use proto::Plan;
+use release_channel::AppVersion;
 use schemars::JsonSchema;
 use serde::{Deserialize, Serialize, de::DeserializeOwned};
 use settings::{Settings, SettingsStore};
@@ -39,7 +42,7 @@ use zed_llm_client::{
     CompletionRequestStatus, CountTokensBody, CountTokensResponse, EXPIRED_LLM_TOKEN_HEADER_NAME,
     MAX_LLM_MONTHLY_SPEND_REACHED_HEADER_NAME, MODEL_REQUESTS_RESOURCE_HEADER_VALUE,
     SERVER_SUPPORTS_STATUS_MESSAGES_HEADER_NAME, SUBSCRIPTION_LIMIT_RESOURCE_HEADER_NAME,
-    TOOL_USE_LIMIT_REACHED_HEADER_NAME,
+    TOOL_USE_LIMIT_REACHED_HEADER_NAME, ZED_VERSION_HEADER_NAME,
 };
 
 use crate::AllLanguageModelSettings;
@@ -526,6 +529,7 @@ impl CloudLanguageModel {
     async fn perform_llm_completion(
         client: Arc<Client>,
         llm_api_token: LlmApiToken,
+        app_version: Option<SemanticVersion>,
         body: CompletionBody,
     ) -> Result<PerformLlmCompletionResponse> {
         let http_client = &client.http_client();
@@ -542,6 +546,12 @@ impl CloudLanguageModel {
             } else {
                 request_builder.uri(http_client.build_zed_llm_url("/completions", &[])?.as_ref())
             };
+            let request_builder = if let Some(app_version) = app_version {
+                request_builder.header(ZED_VERSION_HEADER_NAME, app_version.to_string())
+            } else {
+                request_builder
+            };
+
             let request = request_builder
                 .header("Content-Type", "application/json")
                 .header("Authorization", format!("Bearer {token}"))
@@ -774,7 +784,7 @@ impl LanguageModel for CloudLanguageModel {
     fn stream_completion(
         &self,
         request: LanguageModelRequest,
-        _cx: &AsyncApp,
+        cx: &AsyncApp,
     ) -> BoxFuture<
         'static,
         Result<
@@ -784,6 +794,7 @@ impl LanguageModel for CloudLanguageModel {
         let thread_id = request.thread_id.clone();
         let prompt_id = request.prompt_id.clone();
         let mode = request.mode;
+        let app_version = cx.update(|cx| AppVersion::global(cx)).ok();
         match &self.model {
             CloudModel::Anthropic(model) => {
                 let request = into_anthropic(
@@ -804,6 +815,7 @@ impl LanguageModel for CloudLanguageModel {
                     } = Self::perform_llm_completion(
                         client.clone(),
                         llm_api_token,
+                        app_version,
                         CompletionBody {
                             thread_id,
                             prompt_id,
@@ -855,6 +867,7 @@ impl LanguageModel for CloudLanguageModel {
                     } = Self::perform_llm_completion(
                         client.clone(),
                         llm_api_token,
+                        app_version,
                         CompletionBody {
                             thread_id,
                             prompt_id,
@@ -891,6 +904,7 @@ impl LanguageModel for CloudLanguageModel {
                     } = Self::perform_llm_completion(
                         client.clone(),
                         llm_api_token,
+                        app_version,
                         CompletionBody {
                             thread_id,
                             prompt_id,