feat(inputs): add data dictionary (adapted from old but JSON-style)

Author: amyheather

inputs/data_dictionary.md

@@ -0,0 +1,59 @@
+# Data Dictionary: `parameters` JSON
+
+## Top-level key: `simulation_parameters`
+
+Type: `object`
+
+Description:  Maps parameter names (str) to a specification describing how to sample from a statistical distribution for this metric in the simulation.
+
+## Structure summary
+
+Each item under `simulation_parameters` is itself an object with:
+
+* `class_name`: The name of the distribution class to use.
+* `params`: An object containing parameters required by that distribution.
+
+## Parameter specification table
+
+| Field | Data type | Description | Example/Allowed values |
+| - | - | - | - |
+| Parameter name | str (object key) | Description name for the parameter: `<unit>_<metric>_<type>` | `asu_arrival_stroke`, `rehab_los_other`, `asu_routing_tia` |
+| `class_name` | str | Statistical distribution for the parameter | `Exponential`, `Lognormal`, `DiscreteEmpricial` |
+| `params ` | Object | Dictionary of parameters required to instantiate the distribution | See subsequent rows per distribution type |
+
+## Distribution-specific `params` field
+
+| `class_name` | Parameter key(s) | Data type | Description | Example values |
+| - | - | - | - | - |
+| `Exponential` | mean | float | Mean of exponential distribution | `1.2`, `9.3` |
+| `Lognormal` | mean, stdev | float | Mean and standard deviation of lognormal distribution | `mean: 7.4`, `stdev: 8.61` |
+| `DiscreteEmpirical` | values, freq | list (str/float) | Possible discrete values; corresponding probabilities or frequencies | `values: ["rehab", "esd", "other"]`, `freq: [0.24, 0.13, 0.63]` |
+
+## Glossary
+
+### Unit
+
+* `asu`: Acute Stroke Unit
+* `rehab`: Rehabilitation Unit
+
+### Metric
+
+* `arrival`: Interarrival time (days between admissions)
+* `los`: Length of stay (days in unit)
+* `routing`: Probabilities of routing/discharge
+
+### Type
+
+For `iat`/`los`:
+
+* `stroke`: Stroke patients
+* `tia`: Transient ischaemic attack patients
+* `neuro`: Complex neurological patients
+* `other`: Other patient types
+* `stroke_noesd`, `stroke_esd`: Stroke patients split by whether they were transferred to early supported discharge (ESD).
+
+For `routing`:
+
+* `[diagnosis]_rehab`: Probability of transferring to rehabilitation unit
+* `[diagnosis]_esd`: Probability of early supported discharge (ESD).
+* `[diagnosis]_other`: Probability of other discharge pathways