feat(preprocess): add resumable DWH jobs

harness/tht/cli/preprocess_cmd.py

@@ -15,6 +15,39 @@ from tht.cli.config_cmd import CONFIG_OPT
 preprocess_app = typer.Typer(help="Materialize versioned preprocessing artifacts")
 
 
+def run_dwh_from_config(
+    config: Path, *, steps: tuple[str, ...], resume: str | None = None,
+):
+    from tht.cli.lsh_cmd import build_lsh_artifacts
+    from tht.cli.schema_cmd import _load_config_or_exit, refresh_catalog
+    from tht.jobs.dwh_pipeline import DwhPreprocessPipeline, fingerprint
+
+    cfg = _load_config_or_exit(config)
+    workspace_root = cfg.paths.artifacts.parent
+    pipeline = DwhPreprocessPipeline(
+        workspace_id=config.stem.lower().replace(".", "-").replace("_", "-"),
+        workspace_root=workspace_root,
+        config_fingerprint=fingerprint(cfg.model_dump_json()),
+        input_fingerprint=fingerprint(config.resolve().as_posix()),
+        introspect=lambda: refresh_catalog(cfg),
+        build_lsh=lambda: build_lsh_artifacts(cfg),
+    )
+    return pipeline.run(steps, resume_run_id=resume)
+
+
+def _parse_dwh_steps(value: str) -> tuple[str, ...]:
+    allowed = ("introspect", "lsh")
+    steps = tuple(part.strip() for part in value.split(",") if part.strip())
+    if (
+        not steps
+        or len(steps) != len(set(steps))
+        or any(step not in allowed for step in steps)
+        or tuple(sorted(steps, key=allowed.index)) != steps
+    ):
+        raise ValueError("steps must be a unique ordered subset of introspect,lsh")
+    return steps
+
+
 def run_from_config(config: Path, *, dry_run: bool = False, resume: str | None = None):
     from tht.adapters.factory import build_evidence_sources, build_vector_store
     from tht.cli.schema_cmd import _load_config_or_exit
@@ -122,3 +155,46 @@ def evidence_cmd(
             f"OK: run={payload['run_id']} generation={payload['generation']} changed={len(payload['changed'])} "
             f"unchanged={len(payload['unchanged'])} removed={len(payload['removed'])}"
         )
+
+
+@preprocess_app.command("dwh")
+def dwh_cmd(
+    config: Path = CONFIG_OPT,
+    steps: str = typer.Option("introspect,lsh", "--steps"),
+    resume: str | None = typer.Option(None, "--resume"),
+    json_output: bool = typer.Option(False, "--json"),
+) -> None:
+    try:
+        selected = _parse_dwh_steps(steps)
+    except ValueError:
+        payload = {"status": "failed", "error": "invalid DWH preprocessing steps"}
+        if json_output:
+            typer.echo(json.dumps(payload, sort_keys=True))
+        else:
+            typer.secho("ERRORE: invalid DWH preprocessing steps", fg=typer.colors.RED, err=True)
+        raise typer.Exit(code=2) from None
+    if resume is not None and re.fullmatch(r"[0-9a-f]{32}", resume) is None:
+        payload = {"status": "failed", "error": "resume requires a preprocessing run id"}
+        if json_output:
+            typer.echo(json.dumps(payload, sort_keys=True))
+        else:
+            typer.secho("ERRORE: resume requires a preprocessing run id", fg=typer.colors.RED, err=True)
+        raise typer.Exit(code=2)
+    try:
+        result = run_dwh_from_config(config, steps=selected, resume=resume)
+    except Exception:
+        payload = {"status": "failed", "error": "DWH preprocessing failed"}
+        if json_output:
+            typer.echo(json.dumps(payload, sort_keys=True))
+        else:
+            typer.secho("ERRORE: DWH preprocessing failed", fg=typer.colors.RED, err=True)
+        raise typer.Exit(code=1) from None
+    payload = result.model_dump(mode="json")
+    if json_output:
+        typer.echo(json.dumps(payload, ensure_ascii=False, sort_keys=True))
+    elif result.status == "succeeded":
+        typer.echo(f"OK: run={result.run_id} stages={','.join(selected)}")
+    else:
+        typer.secho(f"ERRORE: run={result.run_id} DWH preprocessing failed", fg=typer.colors.RED, err=True)
+    if result.status != "succeeded":
+        raise typer.Exit(code=1)