Add blocking to JSON specification as common in healthcare pathways

[TomMonks]

The specification needs to handle the common situation where a healthcare pathway is blocked.

Suggested modificatiom -> maybe to the transitions section e.g.

  "transitions": [
    {
      "from": "Acute Treatment",
      "to": "Transfer to Rehab",
      "probability": 1.0,
      "blocking": true
    },
    {
      "from": "Transfer to Rehab",
      "to": "Exit",
      "probability": 1.0
    }
  ]

In context of a full example:

{
  "name": "Acute stroke pathway",
  "description": "Acute stroke care to rehabilitation with bed blocking",
  "activities": [
    {
      "name": "Acute Treatment",
      "type": "activity",
      "resource": {"name": "AcuteBed", "capacity": 10},
      "service_distribution": {
        "type": "lognormal",
        "parameters": {"mean": 7.0, "stdev": 1.0}
      },
      "arrival_distribution": {
        "type": "exponential",
        "parameters": {"rate": 0.067}
      }
    },
    {
      "name": "Transfer to Rehab",
      "type": "activity",
      "resource": {"name": "RehabBed", "capacity": 5},
      "service_distribution": {
        "type": "lognormal",
        "parameters": {"mean": 30.0, "stdev": 5.0}
      }
    }
  ],
  "transitions": [
    {
      "from": "Acute Treatment",
      "to": "Transfer to Rehab",
      "probability": 1.0,
      "blocking": true
    },
    {
      "from": "Transfer to Rehab",
      "to": "Exit",
      "probability": 1.0
    }
  ]
}

[TomMonks]

Current pydantic class is:

class Transition(BaseModel):
    source: str = Field(..., alias="from")
    target: str = Field(..., alias="to")
    probability: float = Field(..., ge=0.0, le=1.0)

the proposed json would mean adding a new field blocking

from typing import Optional

class Transition(BaseModel):
    source: str = Field(..., alias="from")
    target: str = Field(..., alias="to")
    probability: float = Field(..., ge=0.0, le=1.0)
    blocking: Optional[bool] = Field(
        default=False, 
        description="If True, source activity holds its resource until target resource is available (Type I blocking)"
     )

[TomMonks]

I think ciw model would look this - need to verify that I get the same results as expected.

import ciw

params = {
    "arrival_distributions": [
        ciw.dists.Exponential(rate=0.067),   # Acute Treatment
        None,                                                    # Rehab no exeternal
    ],
    "service_distributions": [
       # TM note - I'd need to convert lognormal mean and sd to underlying normal for ciw!
        ciw.dists.Lognormal(mean=7.0, sd=1.0),
        ciw.dists.Lognormal(mean=30.0, sd=5.0),
    ],
    "number_of_servers": [10, 5],
    "routing": [
        [0.0, 1.0],   # Acute Treatment -> Transfer to Rehab
        [0.0, 0.0],   # Transfer to Rehab -> Exit
    ],
    "queue_capacities": [
        float("inf"), # no  cap on waiting at Acute
        0,            # only the 5 beds for rehab therefore bed blocking in place.
    ],
}

N = ciw.create_network(**params)

[TomMonks]

would need to modify multiple_replications to output blocking information: details here: https://ciw.readthedocs.io/en/latest/Guides/Queues/queue_capacities.html

[TomMonks]

currently working on v0.12 and that would modify this slightly as it allows for multiple classes of patients.