New distributions code moved down into emod-api from emodpy (Fix #50, was issue 43 in emodpy previously) by ckirkman-IDM · Pull Request #50 · EMOD-Hub/emod-api

ckirkman-IDM · 2025-11-19T21:52:21Z

There were a lot of things using the old distributions. Changes in here include, but my memory may (momentarily) be forgetting more:

migrated distribution classes from emodpy into emod-api
migrated emodpy Demographics#set_X_distribution() methods into emod-api DemographicsBase
removed old distribution classes and related conditional code to allow them & new distributions (e.g. AgeDistributionOld is now gone, code won't even try to check for it anymore)
restored demographics loading from_file() to support current migration usage.
removing Demographics.from_params() and associated migration call into it.
Restored many, many tests. Deleted many, many duplicative, outmoded tests of old distributions in emod-api
Extended Node tests to ensure above deletions of tests don't leave any gaps. (nodes cover virtually all of Demog data writing)
Removed DemographicsTemplates
Implicit functions now exist in their own dedicated space: emod_api/demographics/implicit_functions.py
Removed many, many old outmoded code from DemographicsBase supporting "raw" editing of default node, users should use the set_X_distribution() methods now.
Reworked DemographicsTemplates, which now look like they are identical in design to Demographics. It'd be great to get some thumbs up/down after trying things out from Kurt to delete DemgographicsTemplates (or leave an extremely minor stub).
Removed other dead, crusty code like demog_send_over_pipes to reduce maintenance and eliminate future questions over what it is and keeping their tests working for no useful reason. And I admit, also because I wanted to.
Loading demographics via Demographics.from_file() now sets implicit functions on its return for config consistency.

…structor call (Fix: emod-api-old/685)

…file() demog, updated tests

…moving migration.from_params().

emod_api/demographics/demographics.py

YeChen-IDM · 2025-11-19T23:06:06Z

emod_api/utils/distributions/base_distribution.py

+from emod_api import schema_to_class as s2c
+
+
+class BaseDistribution(ABC):


Is there a specific reason to put the distribution classes into separate .py file? I would prefer to merge them into a single file so user can import them more simply:

from emod_api.utils.distributions import BimodalDistribution

Not:
from emod_api.utils.distributions.bimodal_distribution import BimodalDistribution

One class per file aids in readability and more easily understanding where things are by glancing at file lists. I find this to be highly important. For us (me? :D) and for users, especially since these distributions will be user-facing.

Most users will end up using a small subset at any one time, so their imports will not look so detailed.

Does this break what we did in emodpy-hiv? This seems opposite from our 2.0 pattern for interventions and targeting config.

This will require a small update in emodpy-hiv , yes.

will it break user code? We really don't want to break anything in emodpy-hiv.

Since emodpy needs to be updated, too, I propose that the emod-api distribution classes be imported into init.py in emodpy.utils.distributions. This will avoid breakage in code.

In fact, this is currently what I have in emodpy for PR'ing.

…ject for config consistencty

Merging from mercury, after merging from main, to obtain s2c updates from main.

Bridenbecker

Looks good. Most of my comments are about adding comments and creating tickets for changes to emodpy-malaria and emodpy-hiv.

emod_api/demographics/demographics.py

Bridenbecker · 2025-11-20T23:50:39Z

emod_api/demographics/demographics.py

            default_node: (Node) Represents default values for all nodes, unless overridden on a per-node basis.
+            metadata: (Dict) set the demographics metadata to the supplied dictionary. Default yields default
+                metadata values.
+            set_defaults: (bool) Whether to set default node attributes on the default node. Defaults to True.


Shouldn't the documentation for the init() method go above it since it is part of creating an object of the class?

Uh, I have never put such documentation there. I can put it wherever. It would be good to talk about a team-standard and decide (moving forward). As for here, after taking a look at the below, I'll do whatever.

PEP #257 apparently indicates init() is the "proper" place for this. It is the "most liked" solution in this relevant stackoverflow discussion, but by not means the only solution proposed/discussed.

https://stackoverflow.com/questions/37019744/is-there-a-consensus-on-what-should-be-documented-in-the-class-and-init-docs

emod_api/demographics/demographics.py

emod_api/demographics/demographics_base.py

tests/test_demographics.py

tests/unittests/test_demographics_imports.py

…t constructor. Changing default idref for demographic files to default_id_reference .

… hiv or malaria sims

… hiv sims

…ine printing, numbers indicate nesting depth in space count.

… Appears to work either way.

… of PropertiesAndAttributes

Bridenbecker

Thank you for making all of those changes. the comments about what is used in what disease are really going to be helpful later.

emod_api/demographics/base_input_file.py

emod_api/demographics/demographics.py

Bridenbecker · 2025-12-03T01:28:35Z

emod_api/demographics/properties_and_attributes.py

+                 mortality_distribution_female: MortalityDistribution = None,
                 innate_immune_distribution_flag: int = None,
                 innate_immune_distribution1: int = None,
                 innate_immune_distribution2: int = None


My reason for doing the subclassing is that when we support multiple diseases, it helps to the user and developer to know what is supported in what disease.

emod_api/demographics/properties_and_attributes.py

Bridenbecker · 2025-12-03T01:31:56Z

tests/test_config_demog.py

+
+        # Ensure we have the right config items set now.
+        self.assertEqual(self.config.parameters.Age_Initialization_Distribution_Type, "DISTRIBUTION_SIMPLE")
+        self.assertEqual(self.config.parameters.Susceptibility_Initialization_Distribution_Type, "DISTRIBUTION_COMPLEX")


I guess it depends on whether implicit function for Susceptibility_Initialization_Distribution_Type cascades up.

Bridenbecker · 2025-12-03T01:32:34Z

tests/test_demographics.py

+        input_filename = os.path.join(manifest.demo_folder, "Namawala_four_node_demographics_for_Thomas.json")
+        output_filename = os.path.join(self.out_folder, "Namawala_four_node_demographics_for_Thomas_comparison.json")
+
+        self.pass_through_test(input_filename, output_filename)


tests/test_demographics.py

…ither (age, susceptibility) will now throw an exception to prevent ambiguity when a demographics file is being created.

ckirkman-IDM · 2025-12-04T21:21:43Z

@Bridenbecker
BTW, I added demographics.to_dict() checking of dual complex/simple distribution setting for age and susceptibility. An exception will be thrown if both exist for either (ambiguous). Thus, if someone reads in a demographics.json and then tries to use it for a new simulation, by calling demographics.to_file()/generate_file()/to_dict() it will fail with an appropriate message.

Updating after merging main into mercury, closing PR 43 after this if tests pass.

ckirkman-IDM added 5 commits November 13, 2025 14:32

fixing up distribution tests from emodpy

204d0be

Fixing up IndividualProperties to properly handle a None (no IPs) con…

9428b2d

…structor call (Fix: emod-api-old/685)

Big cleanup, removed demog templates, old dist objs, reenabling from_…

119abee

…file() demog, updated tests

restoring migration tests after having restored demog,from_file(). Re…

0ace70b

…moving migration.from_params().

doc fixes and linting

a2cb1c8

ckirkman-IDM requested review from Bridenbecker, YeChen-IDM, kfrey-idm and stitova-idm November 19, 2025 21:59

kfrey-idm reviewed Nov 19, 2025

View reviewed changes

emod_api/demographics/demographics.py Outdated Show resolved Hide resolved

YeChen-IDM reviewed Nov 19, 2025

View reviewed changes

Demographics.from_file() now sets implicit functions on its return ob…

d080adf

…ject for config consistencty

ckirkman-IDM changed the title ~~New distributions code moved down into emod-api from emodpy (Fix #43)~~ New distributions code moved down into emod-api from emodpy (Fix #50, was issue 43 in emodpy previously) Nov 20, 2025

ckirkman-IDM and others added 7 commits November 20, 2025 12:43

merging from mercury

f30e770

fixing minor merge error, param rename

7125f4e

Merge branch 'mercury' into 43

7b4948a

Merging from mercury, after merging from main, to obtain s2c updates from main.

linting

6eb5668

Simplify schema processing of interventions; add tests (EMOD-Hub#49)

d680a0f

Bump version: 2.0.36 → 2.0.37

ae4a61e

added google tag manager for analytics tracking (EMOD-Hub#53)

a93ee9e

Bridenbecker requested changes Nov 24, 2025

View reviewed changes

ckirkman-IDM and others added 9 commits November 25, 2025 12:53

removing ability to pass in unverified metadata to demographics objec…

fcfa59a

…t constructor. Changing default idref for demographic files to default_id_reference .

minor file_name -> filename usage homogenization

3c582dd

Change tracking_id to property in mkdocs.yml (EMOD-Hub#54)

dbebc43

adding note that susceptibility distributions are not compatible with…

43703f8

… hiv or malaria sims

adding note that susceptibility distributions are not compatible with…

381c0ab

… hiv sims

removing e.g. usage

186f242

minor doc update

c5c159b

distribution explanatory note

b68ae0f

minor comment fix

8ab7f4e

ckirkman-IDM and others added 9 commits November 25, 2025 16:16

cleanup of cruft and a touch of named parameters for clarity

71f93ac

Clean up docs (EMOD-Hub#56)

5837553

Bump version: 2.0.37 → 2.0.38

167b65c

demographics.to_file() indent is now optional, where None means one-l…

568cb5c

…ine printing, numbers indicate nesting depth in space count.

adding Path to demographics.to_file() type hinting

4d420d1

review updates

ab49bbe

changing Demgoraphics.from_file() to use 'r' reading instead of 'rb'.…

eded5d2

… Appears to work either way.

further extending Demographics.from_file() warning

9028fd0

Updating review comments related to hiv/malaria only comment sections…

d9b7252

… of PropertiesAndAttributes

ckirkman-IDM requested a review from Bridenbecker December 2, 2025 23:17

Bridenbecker approved these changes Dec 3, 2025

View reviewed changes

ckirkman-IDM added 3 commits December 4, 2025 12:37

Setting both a simple and complex distribution for those that allow e…

e17b450

…ither (age, susceptibility) will now throw an exception to prevent ambiguity when a demographics file is being created.

final cleanup from review comments

5b4cc8f

linting

de9668c

ckirkman-IDM added 2 commits December 4, 2025 13:29

Merge branch 'main' into mercury

2e97a01

Merge branch 'mercury' into 43

dbe35d3

Updating after merging main into mercury, closing PR 43 after this if tests pass.

ckirkman-IDM merged commit d544df9 into EMOD-Hub:mercury Dec 4, 2025
2 checks passed

ckirkman-IDM deleted the 43 branch December 4, 2025 21:39

This was referenced Nov 20, 2025

DemographicsBase: remove deprecated functions. EMOD-Hub/issues-and-discussions#95

Closed

Remove DemographicsTemplates EMOD-Hub/issues-and-discussions#98

Closed

		from emod_api import schema_to_class as s2c


		class BaseDistribution(ABC):

Comments

Conversation

ckirkman-IDM commented Nov 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ckirkman-IDM Nov 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Bridenbecker left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Bridenbecker left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ckirkman-IDM commented Dec 4, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

ckirkman-IDM commented Nov 19, 2025 •

edited

Loading

ckirkman-IDM Nov 25, 2025 •

edited

Loading