Library & Maps
Case Library
Every case in the common synthetic suite, by project. Each shows the value written before the run, the value the code produced and whether the two are equal.
These cases test code behaviour on fixed synthetic inputs. They are not patient cases and not research outcomes.
All cases as JSON · Replay command
Discovery and data
5-1 · AI drug discovery (11/11 matched)
Relative binding free-energy format and error calculation · Synthetic benchmark sample and scaffold separation
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| ENERGY-ERRORS | "MAE 0.333 / RMSE 0.408 / RMSEc 0.408 kcal/mol" | "MAE 0.333 / RMSE 0.408 / RMSEc 0.408 kcal/mol" | Matched |
| ENERGY-CENTERING | "MAE 0.667 / RMSE 0.816 / RMSEc 0.471 kcal/mol" | "MAE 0.667 / RMSE 0.816 / RMSEc 0.471 kcal/mol" | Matched |
| ENERGY-UNIT | "unit-mismatch" | "unit-mismatch" | Matched |
| ENERGY-MISSING | "coverage-mismatch" | "coverage-mismatch" | Matched |
| ENERGY-DUPLICATE | "duplicate-or-missing-id" | "duplicate-or-missing-id" | Matched |
| ENERGY-REFERENCE | "reference-not-zero" | "reference-not-zero" | Matched |
| ENERGY-NONFINITE | "invalid-number" | "invalid-number" | Matched |
| ENERGY-SEM | "invalid-number" | "invalid-number" | Matched |
| FAMILY-DISJOINT | "disjoint-synthetic-split" | "disjoint-synthetic-split" | Matched |
| FAMILY-OVERLAP | "scaffold-overlap" | "scaffold-overlap" | Matched |
| SAMPLE-OVERLAP | "sample-overlap" | "sample-overlap" | Matched |
8-1 · Synthetic data (5/5 matched)
Synthetic record schema and variable consistency
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| SYNTHETIC-VALID | "schema-accepted" | "schema-accepted" | Matched |
| SYNTHETIC-EXTRA | "schema-rejected" | "schema-rejected" | Matched |
| SYNTHETIC-ID | "synthetic-id-required" | "synthetic-id-required" | Matched |
| SYNTHETIC-TYPE | "invalid-type-or-range" | "invalid-type-or-range" | Matched |
| SYNTHETIC-TIME | "inconsistent-observation" | "inconsistent-observation" | Matched |
19-1 · Rare disease diagnosis (4/4 matched)
Top-five ranks and tie handling
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| RANK-FIFTH | "top-five-hit" | "top-five-hit" | Matched |
| RANK-SIXTH | "top-five-miss" | "top-five-miss" | Matched |
| RANK-TIE | "cutoff-tie" | "cutoff-tie" | Matched |
| RANK-DUPLICATE | "duplicate-id" | "duplicate-id" | Matched |
Cognitive and neurological health
16-2 · Adaptive cognitive training (4/4 matched)
Simulated difficulty-control replay
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| ADAPT-UP | "settled" | "settled" | Matched |
| ADAPT-DOWN | "settled" | "settled" | Matched |
| ADAPT-FATIGUE | "paused-fatigue" | "paused-fatigue" | Matched |
| ADAPT-BOUNDARY | "outside-target" | "outside-target" | Matched |
16-3 · Dementia diagnosis (5/5 matched)
Disjoint sites, scanners and participants, with subject-level modality pairing
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| COHORT-DISJOINT | "eligible-on-metadata" | "eligible-on-metadata" | Matched |
| COHORT-DEVICE | "site-or-device-overlap" | "site-or-device-overlap" | Matched |
| COHORT-SUBJECT | "subject-overlap" | "subject-overlap" | Matched |
| COHORT-UNPAIRED | "unpaired-modalities" | "unpaired-modalities" | Matched |
| COHORT-DUPLICATE | "duplicate-id" | "duplicate-id" | Matched |
Patient safety
18-2 · Medication safety (18/18 matched)
Mass unit consistency · Boundary cases for units, combinations and identification of synthetic products
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| MASS-EQUIVALENT | "mass-matched" | "mass-matched" | Matched |
| MASS-THOUSAND | "mass-mismatch" | "mass-mismatch" | Matched |
| MASS-ZERO | "invalid-quantity" | "invalid-quantity" | Matched |
| MASS-UNKNOWN | "unknown-unit" | "unknown-unit" | Matched |
| MASS-NEGATIVE | "invalid-quantity" | "invalid-quantity" | Matched |
| MED-BOUNDARY | "no-rule-triggered" | "no-rule-triggered" | Matched |
| MED-AMOUNT | "flag:mass" | "flag:mass" | Matched |
| MED-UNKNOWN-UNIT | "review:unknown-unit" | "review:unknown-unit" | Matched |
| MED-RENAL | "flag:renal" | "flag:renal" | Matched |
| MED-UNKNOWN-RENAL | "review:unknown-renal" | "review:unknown-renal" | Matched |
| MED-PAIR | "flag:combination" | "flag:combination" | Matched |
| MED-UNKNOWN-PAIR | "review:unknown-combination" | "review:unknown-combination" | Matched |
| MED-SIMILAR | "flag:scan" | "flag:scan" | Matched |
| MED-TOLERANCE | "flag:scan" | "flag:scan" | Matched |
| MED-UNKNOWN-PRODUCT | "review:unknown-product" | "review:unknown-product" | Matched |
| MED-UNKNOWN-SCAN | "review:incomplete-scan" | "review:incomplete-scan" | Matched |
| MED-LEAK | "review:input-fields" | "review:input-fields" | Matched |
| MED-MULTIPLE | "flag:mass,renal,combination,scan" | "flag:mass,renal,combination,scan" | Matched |
18-1 · Fall early warning (6/6 matched)
Information cutoff and warning lead time
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| ALERT-3S | "early-alert" | "early-alert" | Matched |
| ALERT-BELOW-3S | "late-alert" | "late-alert" | Matched |
| ALERT-LEAK | "feature-not-available" | "feature-not-available" | Matched |
| ALERT-NO-EVENT | "no-event" | "no-event" | Matched |
| ALERT-FALSE | "false-alert" | "false-alert" | Matched |
| ALERT-MISSED | "missed-event" | "missed-event" | Matched |
18-3 · Surgical safety (5/5 matched)
Occlusion, arrival and removal tracking
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| COUNT-BALANCED | "inventory-balanced" | "inventory-balanced" | Matched |
| COUNT-UNBALANCED | "inventory-unbalanced" | "inventory-unbalanced" | Matched |
| COUNT-OBSCURED | "review-obscuration" | "review-obscuration" | Matched |
| COUNT-DUPLICATE | "invalid-sequence" | "invalid-sequence" | Matched |
| COUNT-NO-CLOSE | "missing-close" | "missing-close" | Matched |
Crisis prevention
12-1 · Acute crisis signals (5/5 matched)
Evidence spans available at the time of the decision
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| NOTE-MATCH | "span-matched" | "span-matched" | Matched |
| NOTE-ALTERED | "quote-mismatch" | "quote-mismatch" | Matched |
| NOTE-FUTURE | "future-source" | "future-source" | Matched |
| NOTE-WRONG-SOURCE | "unknown-source" | "unknown-source" | Matched |
| NOTE-RANGE | "invalid-span" | "invalid-span" | Matched |
12-2 · Safety-plan transitions (4/4 matched)
Constrained response transitions in synthetic event streams
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| PROTOCOL-ORDINARY | "protocol|protocol" | "protocol|protocol" | Matched |
| PROTOCOL-URGENT | "protocol|handoff|handoff" | "protocol|handoff|handoff" | Matched |
| PROTOCOL-BLOCKED | "handoff|handoff" | "handoff|handoff" | Matched |
| PROTOCOL-UNKNOWN | "review|review|handoff" | "review|review|handoff" | Matched |
12-3 · Social isolation (5/5 matched)
Distinguishing missing aggregate observations from changing records
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| OBS-DEVIATION | "synthetic-deviation" | "synthetic-deviation" | Matched |
| OBS-NORMAL | "no-deviation" | "no-deviation" | Matched |
| OBS-OUTAGE | "insufficient-observation" | "insufficient-observation" | Matched |
| OBS-SHORT | "insufficient-observation" | "insufficient-observation" | Matched |
| OBS-RAW | "schema-rejected" | "schema-rejected" | Matched |
12-4 · Student crisis handoffs (6/6 matched)
Immediate synthetic crisis handoff and termination of unrestricted replies
| Case | Expected | Observed | Outcome |
|---|---|---|---|
| HANDOFF-SAME-TICK | "handoff-complete" | "handoff-complete" | Matched |
| HANDOFF-DELAYED | "handoff-not-immediate" | "handoff-not-immediate" | Matched |
| HANDOFF-MISSING | "missing-handoff" | "missing-handoff" | Matched |
| HANDOFF-FREE-TEXT | "free-reply-after-handoff" | "free-reply-after-handoff" | Matched |
| HANDOFF-OUT-OF-ORDER | "invalid-sequence" | "invalid-sequence" | Matched |
| HANDOFF-UNPROMPTED | "unexpected-handoff" | "unexpected-handoff" | Matched |