Document procedural-typed-decisions configurations

README.md

@@ -16,11 +16,11 @@ configs:
 - config_name: arithmetic
   data_files:
   - split: train
-    path: arithmetic/train-*
+    path: arithmetic/train-*.parquet
   - split: validation
-    path: arithmetic/validation-*
+    path: arithmetic/validation-*.parquet
   - split: test
-    path: arithmetic/test-*
+    path: arithmetic/test-*.parquet
 - config_name: entity_belief_tracking
   data_files:
   - split: train
@@ -101,187 +101,13 @@ configs:
     path: table_lookup/validation-*.parquet
   - split: test
     path: table_lookup/test-*.parquet
-dataset_info:
-  config_name: arithmetic
-  features:
-  - name: id
-    dtype: string
-  - name: level
-    dtype: int32
-  - name: state
-    dtype: string
-  - name: questions
-    dtype: string
-  - name: answers
-    dtype: string
-  - name: starts_before_noon
-    dtype:
-      class_label:
-        names:
-          '0': '0'
-          '1': '1'
-          '2': '2'
-          '3': '3'
-          '4': '4'
-          '5': '5'
-          '6': '6'
-          '7': '7'
-          '8': '8'
-          '9': '9'
-          '10': '10'
-  - name: done_by_deadline
-    dtype:
-      class_label:
-        names:
-          '0': 'false'
-          '1': 'true'
-  - name: random_is_long
-    dtype: float32
-  - name: longest_task
-    dtype:
-      class_label:
-        names:
-          '0': backup check
-          '1': design sync
-          '2': inventory check
-          '3': report writing
-          '4': client call
-          '5': email triage
-          '6': planning
-          '7': standup
-          '8': interviews
-          '9': code review
-  - name: largest_line
-    dtype:
-      class_label:
-        names:
-          '0': kettle
-          '1': lamp
-          '2': mug
-          '3': clock
-          '4': scarf
-          '5': drill
-          '6': tent
-          '7': chair
-          '8': vase
-          '9': rope
-  - name: lines_above
-    dtype:
-      class_label:
-        names:
-          '0': '0'
-          '1': '1'
-          '2': '2'
-          '3': '3'
-          '4': '4'
-          '5': '5'
-          '6': '6'
-          '7': '7'
-          '8': '8'
-          '9': '9'
-          '10': '10'
-  - name: random_line_bulk
-    dtype: float32
-  - name: within_budget
-    dtype:
-      class_label:
-        names:
-          '0': 'false'
-          '1': 'true'
-  - name: amount_due
-    dtype: string
-  - name: budget_use
-    dtype:
-      class_label:
-        names:
-          '0': At most half of the budget.
-          '1': More than half of the budget, but within it.
-          '2': Over the budget.
-  - name: final_balance
-    dtype: string
-  - name: went_negative
-    dtype:
-      class_label:
-        names:
-          '0': 'false'
-          '1': 'true'
-  - name: withdrawal_count
-    dtype:
-      class_label:
-        names:
-          '0': '0'
-          '1': '1'
-          '2': '2'
-          '3': '3'
-          '4': '4'
-          '5': '5'
-          '6': '6'
-          '7': '7'
-          '8': '8'
-          '9': '9'
-          '10': '10'
-  - name: net_change
-    dtype:
-      class_label:
-        names:
-          '0': Fell by more than 50.
-          '1': Changed by 50 or less.
-          '2': Rose by more than 50.
-  - name: random_is_deposit
-    dtype: float32
-  - name: finish_time
-    dtype: string
-  - name: lowest_day
-    dtype:
-      class_label:
-        names:
-          '0': day 2
-          '1': day 12
-          '2': day 15
-          '3': day 18
-          '4': day 19
-          '5': day 20
-          '6': day 23
-          '7': day 25
-          '8': day 26
-          '9': day 27
-          '10': day 1
-          '11': day 6
-          '12': day 7
-          '13': day 9
-          '14': day 3
-          '15': day 5
-          '16': day 8
-          '17': day 24
-          '18': day 28
-          '19': day 4
-          '20': day 14
-          '21': day 17
-          '22': day 10
-          '23': day 13
-          '24': day 22
-          '25': day 11
-          '26': day 21
-          '27': day 16
-  splits:
-  - name: train
-    num_bytes: 30877562
-    num_examples: 20000
-  - name: validation
-    num_bytes: 1541040
-    num_examples: 1000
-  - name: test
-    num_bytes: 1547761
-    num_examples: 1000
-  download_size: 6078317
-  dataset_size: 33966363
 ---
 
 # procedural-typed-decisions
 
 Procedurally generated decision problems. Each row is one structured state
-(JSON, or a table, CSV, key=value lines, or prose for the retrieval and
-aggregation configs) with **several typed questions over that same state**, following the
+(JSON, or a table, CSV, key=value lines, or prose for the arithmetic,
+retrieval, and aggregation configs) with **several typed questions over that same state**, following the
 Jev / System One request shape: `choice` (pick one criterion), `noul` (a
 number in [0, 1]; a probability or a yes/no), and `score` (an ordered rubric).
 Every answer is computed exactly from the state by rules that the state
@@ -294,6 +120,7 @@ is not produced by or affiliated with TypeSafe or OpenJev.
 
 | config | questions |
 |---|---|
+| `arithmetic` | An order with a discount/shipping rule, an account ledger, or a schedule; each state asks 2–5 of: `amount_due` / `final_balance` / `finish_time` (choice among the result and typical slips), `within_budget`, `went_negative`, `done_by_deadline` (noul), `random_line_bulk`, `random_is_deposit`, `random_is_long` (noul, exact probability k/n), `budget_use`, `net_change` (score, descriptive levels), `lines_above`, `withdrawal_count`, `starts_before_noon` (score), `largest_line`, `lowest_day`, `longest_task` (choice) |
 | `entity_belief_tracking` | `world_location` (choice), `agent_belief_location` (choice), `belief_matches_world` (noul) |
 | `event_state_reconstruction` | `current_owner` (choice), `is_open` (noul), `current_severity` (score) |
 | `evidence_sufficiency` | `claim_supported` (noul), `has_conflict` (noul), `strongest_support_origin` (choice) |
@@ -314,15 +141,16 @@ is not produced by or affiliated with TypeSafe or OpenJev.
 | `state` | The state: a JSON string, or rendered text for the retrieval and aggregation configs. |
 | `questions` | JSON object of named System One questions (`type`, `instructions`, `criteria`). |
 | `answers` | JSON object of reference answers, in the System One `answers` shape. |
-| one column per question | Flat label, for browsing and filtering: a `ClassLabel` for choice, score, and yes/no noul questions; a float for graded noul (`incident_real`). |
+| one column per question | Flat label, for browsing and filtering: a `ClassLabel` for choice, score, and yes/no noul questions; a float for graded noul (`incident_real`, `random_*`); the option text for open numeric choices (`amount_due`, `final_balance`, `finish_time`). Null when the state does not ask that question (`arithmetic` only). |
 
 States are unique within a split, and validation/test states never occur in
 train.
 
 ## Use
 
-As a multi-question Jev request, send `{"state": json.loads(row["state"]),
-"questions": json.loads(row["questions"])}` and compare with `row["answers"]`.
+As a multi-question Jev request, send `{"state": row["state"], "questions":
+json.loads(row["questions"])}` (parsing the state first when it is JSON) and
+compare with `row["answers"]`.
 The same rows are included, grouped by state, in
 [`tasksource/tasksource-jev-typed-decisions`](https://huggingface.co/datasets/tasksource/tasksource-jev-typed-decisions).