Initial implementation

docs/evaluation-and-inspection.md

@@ -0,0 +1,42 @@
+# Evaluation And Inspection
+
+`infospace-bench` now has a deterministic baseline for evaluation and
+inspection. It is intentionally small: the repo can produce structured quality
+objects and relationship summaries before any LLM or engine integration is
+introduced.
+
+## Evaluation Objects
+
+- `ScoreEntry`
+- `EntityEvaluation`
+- `MetricValue`
+- `EvaluationSnapshot`
+- `SnapshotDiff`
+
+Snapshots are serializable through `to_dict()` / `from_dict()` and can be
+compared with `diff_snapshots()`.
+
+## Collection Checks
+
+`run_collection_checks()` produces five baseline metrics:
+
+- `redundancy_ratio`
+- `coverage_ratio`
+- `coherence_components`
+- `consistency_cycles`
+- `granularity_entropy`
+
+These metrics are deliberately deterministic and file-backed. Later work can
+replace or extend their internals with embeddings, richer graph analysis, or
+agent-assisted evaluation without changing the result contract.
+
+## Viability
+
+`evaluate_viability()` compares metric values against declared
+`ViabilityThreshold` values. Missing metrics fail visibly.
+
+## Relationship Inspection
+
+`relationship_summary()` extracts nodes, edges, and relationship type counts
+from artifact manifests. `export_mermaid()` provides the first graph-friendly
+representation.

docs/infospace-layout.md

@@ -0,0 +1,55 @@
+# Infospace Layout
+
+An infospace is a file-backed project rooted at:
+
+```text
+infospaces/<slug>/
+```
+
+## Required Files
+
+```text
+infospace.yaml
+artifacts/index.yaml
+```
+
+`infospace.yaml` declares identity, topic, schema references, workflow
+references, discipline bindings, and viability thresholds. `artifacts/index.yaml`
+is the deterministic manifest of artifacts that have been added to the
+infospace.
+
+## Required Directories
+
+```text
+artifacts/sources/
+artifacts/generated/
+output/evaluations/
+output/metrics/
+reports/
+exports/
+```
+
+## Artifact Manifest
+
+Artifacts are represented by stable IDs such as `source/chapter-01.md`.
+
+Each manifest entry records:
+
+- `id`
+- `path`
+- `kind`
+- `title`
+- `provenance`
+- `relationships`
+
+The manifest is intentionally plain YAML so it can be inspected, diffed, and
+regenerated by tools or agents.
+
+## Commands
+
+```bash
+python3 -m infospace_bench create . pilot --name "Pilot Infospace"
+python3 -m infospace_bench add-artifact infospaces/pilot ./source.md --kind source
+python3 -m infospace_bench inspect infospaces/pilot
+python3 -m infospace_bench export infospaces/pilot
+```

docs/reference-pilot-decision.md

@@ -0,0 +1,24 @@
+# Reference Pilot Decision
+
+Date: 2026-05-14
+
+## Decision
+
+Use a small purpose-built corpus as the first maintained reference infospace.
+
+## Rationale
+
+`markitect-main/examples/infospace-with-history/` remains the primary migration
+candidate for a larger pilot, but it contains a large public-domain book corpus
+and substantial generated output. Pulling it in before the lifecycle and
+evaluation baseline exists would make the new repo noisy before it is useful.
+
+The bootstrap pilot uses this repo's own PRD/FRS intent as a compact corpus. It
+proves the expected file layout, artifact manifest, relationship inspection,
+collection metrics, and viability thresholding with minimal bulk.
+
+## Follow-up
+
+After the baseline is stable, migrate a pruned Wealth of Nations/VSM fixture or
+a similarly representative slice from `markitect-main` under a separate
+workplan.