IB-WP-0018-T01+T02+T05: routing bridge to llm-connect

T01 — task-type taxonomy. docs/routing-task-types.md names the five
generation stages as the default identity-mapped task types
(summarize-source, extract-entities, extract-relations,
evaluate-entity, synthesize-report) and records the recommended quality
floors per stage. The taxonomy explicitly does not decide which adapter
ships per task type, where the ledger lives, or what a quality score
means — those stay with the caller per the LLM-WP-0004 scope guardrail.

T02 — RoutingAssistedGenerationAdapter bridge in
src/infospace_bench/routing.py. Wraps any llm-connect RoutingPolicy or
AdaptiveRoutingPolicy as an infospace-bench AssistedGenerationAdapter:
maps stage_id -> task_type (overridable), resolves an LLMAdapter,
delegates execute_prompt with a configurable RunConfig, and surfaces
the resolved adapter id, task type, model, usage, and finish_reason
back on AssistedGenerationResult.metadata. Provider tag stays
back-compatible with the strings already used in run records and the
budget rollup (openrouter / claude_code / openai / gemini / mock /
routing).

T05 — eight tests in tests/test_routing_adapter.py cover: static-policy
per-stage resolution, stage_to_task_type overrides, default-mapping
completeness, fall-through for unmapped stage ids, the adaptive path
selecting the cheaper qualifying adapter when a quality_floor is set,
adaptive policy falling back to static when no floor is set, response
metadata round-trip with provider tagging, and estimated_cost_per_1k
pass-through.

Adds llm-connect as a path dependency on pyproject.toml and to the
pytest pythonpath. Static OpenRouter and fixture paths are unchanged;
this commit only adds the option of routing.

139 tests pass, 1 skipped (the OpenRouter live smoke, gated as before).

T03 (shadow-mode integration) and T04 (CLI + per-stage chosen-adapter
in the generation report) follow next.

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

docs/routing-task-types.md

@@ -0,0 +1,78 @@
+# Task-Type Taxonomy for Routing
+
+Workplan: IB-WP-0018 (T01)
+Depends on: llm-connect LLM-WP-0004 (RoutingPolicy, AdaptiveRoutingPolicy)
+
+This file names the task types that infospace-bench emits when it routes
+each generation stage through llm-connect. The names are the consumer
+side of LLM-WP-0004's scope guardrail: llm-connect ships the routing
+primitives, infospace-bench owns the taxonomy.
+
+## Default identity mapping
+
+`RoutingAssistedGenerationAdapter` (see `src/infospace_bench/routing.py`)
+maps stage ids to task types using the identity mapping below by
+default. Callers override individual entries via
+`RoutingAssistedGenerationAdapter(..., stage_to_task_type={...})`.
+
+| Stage id | Task type | Notes |
+|---|---|---|
+| `summarize-source` | `summarize-source` | One call per source chunk. Cheap, high-volume; small models usually clear the bar. |
+| `extract-entities` | `extract-entities` | One call per source chunk. Quality matters most here — bad extractions cascade. |
+| `extract-relations` | `extract-relations` | One call per source chunk. Quality close to extraction; relations rely on entity titles being stable. |
+| `evaluate-entity` | `evaluate-entity` | One call per generated entity. Cheap, often a different model than extraction to avoid self-grading. |
+| `synthesize-report` | `synthesize-report` | One call at the end. Volume-of-one; quality matters; cost negligible. |
+
+## Quality expectations
+
+`AdaptiveRoutingPolicy.resolve(task_type, quality_floor=...)` picks the
+cheapest adapter whose ledger-observed mean quality clears the floor.
+The recommended starting floors:
+
+| Task type | Quality floor | Rationale |
+|---|---|---|
+| `summarize-source` | 0.70 | Summaries are intermediate. Slight quality loss is recoverable downstream. |
+| `extract-entities` | 0.85 | Entities are the durable output. Be strict. |
+| `extract-relations` | 0.80 | Relations depend on entities; slightly looser is OK as long as evidence is intact. |
+| `evaluate-entity` | 0.80 | Judge-level reliability. Self-grading bias is more of a concern than absolute score. |
+| `synthesize-report` | 0.70 | The report is a review surface; tolerate looser language for cheaper models. |
+
+These are starting points. Bind them at the calling site
+(`RoutingAssistedGenerationAdapter(..., quality_floor=0.85)` for
+extraction stages) — they are not enforced by this taxonomy.
+
+## Common overrides
+
+Callers may want to **collapse** task types to share observations across
+related stages, or **split** a task type to pin a specific model to a
+narrow workload. Two illustrative overrides:
+
+```python
+# Collapse extraction stages so a single ledger drives both
+stage_to_task_type = {
+    "extract-entities": "extraction",
+    "extract-relations": "extraction",
+}
+```
+
+```python
+# Split entity evaluation by category — useful when a profile has very
+# different quality bars for different entity categories (e.g.
+# trading-literature's `evidence_bearing_claim` is harder to judge than
+# `instrument`).
+stage_to_task_type = {
+    "evaluate-entity": "judge",
+}
+```
+
+Anything not in the override map falls through to the identity mapping.
+
+## What this taxonomy does NOT decide
+
+- **Which adapter ships per task type.** That belongs to the caller's
+  `RoutingPolicy` rule list.
+- **Where the quality ledger lives.** Caller-supplied path on the
+  `AdaptiveRoutingPolicy`.
+- **When to refresh observations.** Caller decides via the ledger's TTL
+  helpers in llm-connect.
+- **What a quality score means.** Each judge defines its own.