fix(evidence): make result summaries safe

harness/tht/cli/preprocess_cmd.py

@@ -172,9 +172,11 @@ def evidence_cmd(
     if json_output:
         typer.echo(json.dumps(payload, ensure_ascii=False, sort_keys=True))
     else:
+        counts = payload["counts"]
         typer.echo(
-            f"OK: run={payload['run_id']} generation={payload['generation']} changed={len(payload['changed'])} "
-            f"unchanged={len(payload['unchanged'])} removed={len(payload['removed'])}"
+            f"OK: run={payload['run_id']} generation={payload['generation']} "
+            f"changed={counts['changed']} unchanged={counts['unchanged']} "
+            f"removed={counts['removed']}"
         )
 
 

harness/tht/corpus/pipeline.py

@@ -7,7 +7,7 @@ import json
 import re
 import uuid
 from collections.abc import Mapping, Sequence
-from dataclasses import asdict, dataclass
+from dataclasses import asdict, dataclass, field
 from datetime import UTC
 from pathlib import Path
 
@@ -45,10 +45,22 @@ class PipelineResult:
     changed: tuple[str, ...]
     unchanged: tuple[str, ...]
     removed: tuple[str, ...]
-    manifest: CorpusManifest
+    manifest: CorpusManifest = field(repr=False)
     run_id: str | None = None
     resumed_from: str | None = None
 
+    def __repr__(self) -> str:
+        counts = {
+            "changed": len(self.changed),
+            "unchanged": len(self.unchanged),
+            "removed": len(self.removed),
+        }
+        return (
+            f"PipelineResult(status={self.status!r}, generation={self.generation!r}, "
+            f"published={self.published!r}, counts={counts!r}, "
+            f"run_id={self.run_id!r}, resumed_from={self.resumed_from!r})"
+        )
+
     def model_dump(self, mode=None):
         def bounded(values: tuple[str, ...]) -> list[str]:
             return [value[:200] for value in values[:100]]