fix: Langfuse SDK v4 migration, tracing improvements, and LLM config

- Langfuse SDK v4: fix prompt-to-trace linking (as_type=generation)
- tracing: compile_prompt with Langfuse managed prompt fallback
- journey: remove journey CLI subcommand (keep only interactive)
- LLM: add service-specific llm modules for batch-agent and chat
- gitignore: exclude eval private test data
- config: add LANGFUSE settings to shared config

services/batch-agent/app/tracing.py

@@ -232,6 +232,38 @@ def compile_prompt(
         return fallback.format(**variables)
 
 
+def get_prompt_object(
+    name: str,
+    *,
+    version: int | None = None,
+    label: str | None = None,
+    cache_ttl_seconds: int = 300,
+) -> Any | None:
+    """Fetch the raw Langfuse prompt *object* (not the compiled string).
+
+    Returns ``None`` when Langfuse is disabled or the prompt is not found.
+    Use this when you need to pass the prompt to ``start_observation(prompt=...)``
+    for linking the prompt to a trace in the Langfuse UI.
+    """
+    lf = _get_client()
+    if lf is None:
+        return None
+
+    try:
+        kwargs: dict[str, Any] = {
+            "name": name,
+            "cache_ttl_seconds": cache_ttl_seconds,
+        }
+        if version is not None:
+            kwargs["version"] = version
+        if label is not None:
+            kwargs["label"] = label
+        return lf.get_prompt(**kwargs)
+    except Exception as exc:
+        logger.warning("tracing: get_prompt_object(%s) failed: %s", name, exc)
+        return None
+
+
 def link_prompt_to_trace(
     span: Any,
     prompt_name: str,
@@ -239,19 +271,19 @@ def link_prompt_to_trace(
     version: int | None = None,
     label: str | None = None,
 ) -> None:
-    """Attach prompt metadata to a span/trace."""
+    """Link a Langfuse managed prompt to a span/observation.
+
+    Uses the SDK v4 ``prompt=`` parameter so that the prompt version
+    appears linked in the Langfuse UI with metrics tracking.
+    """
     lf = _get_client()
     if lf is None or isinstance(span, _NullSpan):
         return
 
     try:
-        kwargs: dict[str, Any] = {"name": prompt_name}
-        if version is not None:
-            kwargs["version"] = version
-        if label is not None:
-            kwargs["label"] = label
-        prompt = lf.get_prompt(**kwargs)
-        span.update(metadata={"prompt": {"name": prompt_name, "version": prompt.version}})
+        prompt = get_prompt_object(prompt_name, version=version, label=label)
+        if prompt is not None:
+            span.update(prompt=prompt)
     except Exception as exc:
         logger.warning("tracing: link_prompt_to_trace(%s) failed: %s", prompt_name, exc)