IB-WP-0020-T04: example routing config + live routing smoke

examples/routing/trading-literature.yaml is the checked-in starting
config for a Lefevre-style run. It applies the IB-WP-0018 task-type
taxonomy: cheap candidates for summary + evaluation, smart candidates
for entity + relation extraction, and a separate baseline rule wiring
claude_code for a follow-on T05 ShadowingAdapter step. Workspace-
relative ledger_path keeps adaptive observations with the workspace.

tests/test_routing_config.py gains a regression test that asserts the
shipped example parses cleanly, every stage in stage_to_task_type maps
to a declared task type, and the baseline candidate uses the
claude_code provider — so the example will not bit-rot silently.

tests/test_openrouter_live.py gains test_provider_routing_one_chapter_live_smoke
gated on the same INFOSPACE_BENCH_ENABLE_LIVE_OPENROUTER + OPENROUTER_API_KEY
opt-in as the existing static smoke. It builds a one-candidate routing
config, runs a single chapter through --provider routing, and asserts
the per-stage adapter-choices report section names the routed model
and the routed artifacts carry adapter_id provenance.

docs/generic-source-generator.md gains a "Live runs with --provider
routing" subsection that walks through the one-command routed run,
explains the --quality-floor override, and points at the parallel
live smoke test.

174 tests pass, 2 skipped (both live smokes, correctly gated).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

docs/generic-source-generator.md

@@ -94,6 +94,42 @@ skipped unless both `OPENROUTER_API_KEY` and
 chapter through the same path and asserts the provider metadata
 plumb-through.
 
+### Live runs with `--provider routing`
+
+When the routing CLI is what you want to exercise live, swap
+`--provider openrouter --model ...` for the routing pair:
+
+```bash
+infospace-bench generate from-source ./LEFEVRE.epub \
+  --workspace ./infospaces \
+  --slug reminiscences-routed \
+  --name "Reminiscences (Routed)" \
+  --profile trading-literature \
+  --provider routing \
+  --routing-config ./examples/routing/trading-literature.yaml \
+  --chapter I \
+  --apply
+```
+
+`examples/routing/trading-literature.yaml` is a checked-in starting
+config: cheap candidates for summary/evaluation, smart candidates for
+entity/relation, a `claude_code` baseline rule for future shadow
+sampling, and a workspace-relative `output/routing/quality.jsonl`
+ledger so adaptive observations stay with the workspace.
+
+`--quality-floor <float>` on the same command overrides the config's
+`default_quality_floor` for a single invocation — useful for
+tightening the bar for a specific run without editing the file. The
+ledger fills up as the `AdaptiveRoutingPolicy` records each
+observation; later runs against the same workspace get the benefit
+without re-grading from scratch.
+
+The parallel live-smoke test
+(`test_provider_routing_one_chapter_live_smoke`) is also gated on
+`INFOSPACE_BENCH_ENABLE_LIVE_OPENROUTER=1` + `OPENROUTER_API_KEY` and
+asserts the per-stage adapter-choices report section names the routed
+model.
+
 ### Budget and usage registry
 
 Every `generate plan` invocation appends a compact snapshot to

examples/routing/trading-literature.yaml

@@ -0,0 +1,81 @@
+# Example routing config for a trading-literature Lefevre-style run.
+#
+# Captures the IB-WP-0018 task-type taxonomy from docs/routing-task-types.md:
+#   summarize-source  → cheap model (volume-heavy, recoverable downstream)
+#   extract-entities  → smart model (durable output; be strict)
+#   extract-relations → smart model (depends on entities)
+#   evaluate-entity   → judge model (different family from extraction)
+#   synthesize-report → smart model (volume-of-one, quality matters, cheap)
+#
+# Quality floors are the recommended starting points from
+# docs/routing-task-types.md. With a ledger configured, AdaptiveRoutingPolicy
+# will pick the cheapest *qualifying* adapter per task type as observations
+# accumulate; until then it falls back to the static prefer/fallback order.
+#
+# Refresh the model rates in src/infospace_bench/model_rates.yaml before any
+# full-book run — list prices drift, and the rough USD estimate in the budget
+# log depends on them.
+
+schema_version: 1
+
+# Workspace-relative ledger so QualityLedger observations from this workspace
+# stay with this workspace. Drop this line to run pure static routing.
+ledger_path: output/routing/quality.jsonl
+
+# Floors apply when --quality-floor is not passed at the call site. The CLI
+# flag wins, then the per-task quality_floor below, then this default.
+default_quality_floor: 0.80
+
+stage_to_task_type:
+  summarize-source: cheap
+  extract-entities: smart
+  extract-relations: smart
+  evaluate-entity: judge
+  synthesize-report: smart
+
+task_types:
+
+  cheap:
+    quality_floor: 0.70
+    candidates:
+      - id: openrouter:gpt-4o-mini
+        provider: openrouter
+        model: openai/gpt-4o-mini
+        api_key_env: OPENROUTER_API_KEY
+        max_cost_per_1k: 0.001
+      - id: openrouter:claude-3.5-haiku
+        provider: openrouter
+        model: anthropic/claude-3.5-haiku
+        api_key_env: OPENROUTER_API_KEY
+        max_cost_per_1k: 0.003
+
+  smart:
+    quality_floor: 0.85
+    candidates:
+      - id: openrouter:claude-3.5-haiku
+        provider: openrouter
+        model: anthropic/claude-3.5-haiku
+        api_key_env: OPENROUTER_API_KEY
+      - id: openrouter:claude-3.5-sonnet
+        provider: openrouter
+        model: anthropic/claude-3.5-sonnet
+        api_key_env: OPENROUTER_API_KEY
+
+  judge:
+    quality_floor: 0.80
+    candidates:
+      # Evaluation goes through a different family than extraction to limit
+      # self-preference bias.
+      - id: openrouter:gpt-4o-mini
+        provider: openrouter
+        model: openai/gpt-4o-mini
+        api_key_env: OPENROUTER_API_KEY
+
+  # Baseline is wired here so a follow-up T05 ShadowingAdapter step can
+  # reference `claude-code` as the grading oracle without editing the
+  # task_types stanza.
+  baseline:
+    candidates:
+      - id: claude-code
+        provider: claude_code
+        model: claude-opus-4-7

tests/test_openrouter_live.py

@@ -208,3 +208,87 @@ def test_openrouter_one_chapter_smoke(tmp_path: Path) -> None:
         and item.get("provenance", {}).get("provider_metadata", {}).get("request_id")
     ]
     assert generated_with_metadata, "generated artifacts should carry provider_metadata.request_id"
+
+
+_LIVE_ROUTING_REASON = (
+    "set INFOSPACE_BENCH_ENABLE_LIVE_OPENROUTER=1 and OPENROUTER_API_KEY to run "
+    "the optional one-chapter routing smoke against OpenRouter"
+)
+
+
+@pytest.mark.skipif(not (_LIVE_OPT_IN and _LIVE_API_KEY), reason=_LIVE_ROUTING_REASON)
+def test_provider_routing_one_chapter_live_smoke(tmp_path: Path) -> None:
+    """Live smoke: one chapter through --provider routing against OpenRouter.
+
+    Uses a minimal one-candidate-per-task-type routing config so the test
+    spends roughly the same as the static OpenRouter smoke. Asserts the run
+    completes, the routing bridge recorded adapter_id / task_type on
+    provider_metadata, and the per-stage adapter-choices report section
+    reflects routed choices.
+    """
+    book = _build_fixture_epub(tmp_path / "lefevre.epub")
+    model = os.environ.get("INFOSPACE_BENCH_LIVE_MODEL", "openai/gpt-4o-mini")
+
+    routing_config = tmp_path / "routing.yaml"
+    routing_config.write_text(
+        yaml.safe_dump(
+            {
+                "schema_version": 1,
+                "stage_to_task_type": {
+                    "summarize-source": "cheap",
+                    "extract-entities": "cheap",
+                    "extract-relations": "cheap",
+                    "evaluate-entity": "cheap",
+                    "synthesize-report": "cheap",
+                },
+                "task_types": {
+                    "cheap": {
+                        "candidates": [
+                            {
+                                "id": f"openrouter:{model}",
+                                "provider": "openrouter",
+                                "model": model,
+                                "api_key_env": "OPENROUTER_API_KEY",
+                            },
+                        ],
+                    },
+                },
+            },
+            sort_keys=False,
+        ),
+        encoding="utf-8",
+    )
+
+    infospace = init_generation_infospace(
+        tmp_path,
+        book,
+        "lefevre-live-routing",
+        name="Lefevre Live Routing",
+        profile="trading-literature",
+        chapter_filter=["I"],
+    )
+    plan_generation(infospace.root, cost_per_1k_tokens=0.5)
+    result = run_generation(
+        infospace.root,
+        provider="routing",
+        routing_config=routing_config,
+    )
+    status = status_generation(infospace.root)
+
+    assert result.status == "completed"
+    assert status["source_chunk_count"] == 1
+    assert status["entity_count"] >= 1
+
+    report = (infospace.root / "reports" / "generation-summary.md").read_text(encoding="utf-8")
+    assert "## Per-stage adapter choices" in report
+    assert model in report, "report should name the routed model"
+
+    # The routing bridge writes adapter_id + task_type onto provider_metadata.
+    index = yaml.safe_load((infospace.root / "artifacts" / "index.yaml").read_text(encoding="utf-8"))
+    routed_artifacts = [
+        item
+        for item in index["artifacts"]
+        if item["kind"] in {"entity", "relation", "generated"}
+        and (item.get("provenance") or {}).get("provider_metadata", {}).get("adapter_id")
+    ]
+    assert routed_artifacts, "routed artifacts must carry adapter_id provenance"

tests/test_routing_config.py

@@ -412,6 +412,25 @@ def test_build_routing_policy_claude_code_needs_no_api_key() -> None:
     assert isinstance(policy.rules[0].prefer, ClaudeCodeAdapter)
 
 
+def test_example_trading_literature_config_parses() -> None:
+    """Regression: the shipped example config must parse cleanly."""
+    from infospace_bench.routing_config import load_routing_config
+
+    example_path = Path(__file__).resolve().parent.parent / "examples" / "routing" / "trading-literature.yaml"
+
+    config = load_routing_config(example_path)
+
+    task_type_names = {task.task_type for task in config.task_types}
+    assert {"cheap", "smart", "judge", "baseline"} <= task_type_names
+    assert config.default_quality_floor == 0.80
+    # Each shipped stage maps to a task type the config actually declares.
+    for stage, task_type in config.stage_to_task_type.items():
+        assert task_type in task_type_names, f"stage {stage!r} maps to undeclared task type {task_type!r}"
+    # baseline is included so a T05 ShadowingAdapter wiring can reference it.
+    baseline = next(t for t in config.task_types if t.task_type == "baseline")
+    assert baseline.candidates[0].provider == "claude_code"
+
+
 def test_build_routing_policy_honours_custom_api_key_env() -> None:
     from infospace_bench.routing_config import build_routing_policy_from_config
     from llm_connect.openrouter import OpenRouterAdapter

workplans/IB-WP-0020-provider-routing-cli.md

@@ -138,7 +138,7 @@ state_hub_task_id: "fe5888e0-da33-413a-b026-71ed811b8c73"
 
 ```task
 id: IB-WP-0020-T04
-status: todo
+status: done
 priority: medium
 state_hub_task_id: "69288131-f265-4db5-a4b0-b0c8a6f55dd8"
 ```