fix(evidence): bound reports and fail failed jobs

harness/tht/cli/preprocess_cmd.py

@@ -159,6 +159,13 @@ def evidence_cmd(
             typer.secho("ERRORE: preprocessing failed", fg=typer.colors.RED, err=True)
         raise typer.Exit(code=1) from None
     payload = result.model_dump(mode="json")
+    if payload.get("status") != "succeeded":
+        payload["error"] = "preprocessing job failed"
+        if json_output:
+            typer.echo(json.dumps(payload, ensure_ascii=False, sort_keys=True))
+        else:
+            typer.secho("ERRORE: preprocessing job failed", fg=typer.colors.RED, err=True)
+        raise typer.Exit(code=1)
     if json_output:
         typer.echo(json.dumps(payload, ensure_ascii=False, sort_keys=True))
     else:

harness/tht/corpus/pipeline.py

@@ -50,14 +50,24 @@ class PipelineResult:
     resumed_from: str | None = None
 
     def model_dump(self, mode=None):
+        def bounded(values: tuple[str, ...]) -> list[str]:
+            return [value[:200] for value in values[:100]]
+
         return {
             "status": self.status,
             "generation": self.generation,
             "published": self.published,
-            "changed": list(self.changed),
-            "unchanged": list(self.unchanged),
-            "removed": list(self.removed),
-            "manifest": self.manifest.model_dump(mode="json"),
+            "changed": bounded(self.changed),
+            "unchanged": bounded(self.unchanged),
+            "removed": bounded(self.removed),
+            "counts": {
+                "changed": len(self.changed),
+                "unchanged": len(self.unchanged),
+                "removed": len(self.removed),
+                "documents": len(self.manifest.documents),
+                "chunks": len(self.manifest.chunks),
+            },
+            "manifest_id": self.manifest.manifest_id,
             "run_id": self.run_id,
             "resumed_from": self.resumed_from,
         }