feat: add populate_knowledge batch binary with configurable timeout

Adds a standalone binary that walks a directory and runs the agentic
insight loop over every image/video, skipping files already processed.
Supports --path, --model, --max-iterations, --timeout-secs, --num-ctx,
and --reprocess flags for flexible overnight/VPS batch runs.

Also adds OllamaClient::with_request_timeout() builder method so slow
large models are not cut off by the default 120s limit.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/ai/ollama.rs

@@ -73,6 +73,17 @@ impl OllamaClient {
         self.num_ctx = num_ctx;
     }
 
+    /// Replace the HTTP client with one using a custom request timeout.
+    /// Useful for slow models where the default 120s may be insufficient.
+    pub fn with_request_timeout(mut self, secs: u64) -> Self {
+        self.client = Client::builder()
+            .connect_timeout(Duration::from_secs(5))
+            .timeout(Duration::from_secs(secs))
+            .build()
+            .unwrap_or_else(|_| Client::new());
+        self
+    }
+
     /// List available models on an Ollama server (cached for 15 minutes)
     pub async fn list_models(url: &str) -> Result<Vec<String>> {
         // Check cache first