feat: add cli backend response previews

crates/localllm-cli/src/main.rs

@@ -747,23 +747,41 @@ fn backend_request(cli: &Cli) -> BackendRequest {
     }
 }
 
-fn backend_response_preview(cli: &Cli, _status_code: u16, _body: &str) -> BackendResponsePreview {
+fn backend_response_preview(cli: &Cli, _status_code: u16, body: &str) -> BackendResponsePreview {
     match &cli.command {
-        Commands::Status => BackendResponsePreview::SummaryOnly {
+        Commands::Status => serde_json::from_str::<CliHealthResponse>(body)
+            .map(|health| BackendResponsePreview::Health {
+                status: health.status,
+                model_loaded: health.model_loaded,
+                model_id: health.model_id,
+                uptime_secs: health.uptime_secs,
+            })
+            .unwrap_or_else(|_| BackendResponsePreview::SummaryOnly {
                 command: "response-unimplemented".to_owned(),
-        },
+            }),
         Commands::Models(models) => match &models.command {
-            ModelsSubcommand::List => BackendResponsePreview::SummaryOnly {
+            ModelsSubcommand::List => serde_json::from_str::<CliModelList>(body)
+                .map(|models| BackendResponsePreview::Models {
+                    count: models.data.len(),
+                    ids: models.data.into_iter().map(|model| model.id).collect(),
+                })
+                .unwrap_or_else(|_| BackendResponsePreview::SummaryOnly {
                     command: "response-unimplemented".to_owned(),
-            },
+                }),
             _ => BackendResponsePreview::SummaryOnly {
                 command: command_summary(&cli.command),
             },
         },
         Commands::Bench(bench) => match &bench.command {
-            BenchSubcommand::Throughput(_) => BackendResponsePreview::SummaryOnly {
+            BenchSubcommand::Throughput(args) => serde_json::from_str::<CliBenchResponse>(body)
+                .map(|response| BackendResponsePreview::Throughput {
+                    model: args.model.clone(),
+                    total_tokens: response.usage.total_tokens,
+                    tokens_per_second: response.usage.tokens_per_second,
+                })
+                .unwrap_or_else(|_| BackendResponsePreview::SummaryOnly {
                     command: "response-unimplemented".to_owned(),
-            },
+                }),
             _ => BackendResponsePreview::SummaryOnly {
                 command: command_summary(&cli.command),
             },