Update demographic parameters to TxSxJ shape

.github/workflows/build_and_test.yml

@@ -26,7 +26,7 @@ jobs:
     strategy:
       matrix:
         os: [ubuntu-latest, macos-latest, windows-latest]
-        python-version: ["3.11", "3.12"]
+        python-version: ["3.12", "3.13"]
 
     steps:
       - name: Checkout

CHANGELOG.md

@@ -5,26 +5,13 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [Unreleased]
+## [0.2.0] - 2026-08-31 9:00:00
 
 ### Changed
 
-- Rewrote the installation instructions. The README now documents two
-  install paths — the OG family universal installer
-  (`bash install.sh --repo og-bra`, from OG-Core's `scripts/`) and a
-  manual uv install — with complete, copy-pasteable per-platform blocks
-  for macOS, Linux, and Windows, and a "what happens" note for the
-  example run (runtime, output locations, and the UN Data Portal token
-  prompt fallback). Removed the PyPI install section and badges: the
-  `ogbra` package has never been published to PyPI, so `pip install
-  ogbra` fails for everyone who tries it.
-- Updated the contributor guide to the uv workflow: `uv sync --extra
-  dev` and `uv run` replace the conda environment steps, which have
-  been broken since `environment.yml` was removed in 0.1.0. Fixed the
-  test command to this repo's real marker set (`pytest -m "not
-  local"`; the old text cited OG-USA's `needs_puf`/`regression`
-  markers and 24-hour suite) and replaced stale `master`-branch
-  references with `main`.
+- Calibration now works with OG-Core > 0.18.0, which has demographic parameters varying by income goup.
+- Rewrote the installation instructions. The README now documents two install paths — the OG family universal installer (`bash install.sh --repo og-bra`, from OG-Core's `scripts/`) and a manual uv install — with complete, copy-pasteable per-platform blocks for macOS, Linux, and Windows, and a "what happens" note for the  example run (runtime, output locations, and the UN Data Portal token prompt fallback). Removed the PyPI install section and badges: the  `ogbra` package has never been published to PyPI, so `pip install ogbra` fails for everyone who tries it.
+- Updated the contributor guide to the uv workflow: `uv sync --extra dev` and `uv run` replace the conda environment steps, which have  been broken since `environment.yml` was removed in 0.1.0. Fixed the test command to this repo's real marker set (`pytest -m "not local"`; the old text cited OG-USA's `needs_puf`/`regression` markers and 24-hour suite) and replaced stale `master`-branch references with `main`.
 
 ## [0.1.0] - 2026-06-25 12:00:00
 
@@ -45,3 +32,5 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - This version is a pre-release alpha. The example run script OG-BRA/examples/run_og_bra.py runs, but the model is not currently calibrated to represent the Brazilan economy and population.
 
 
+[0.2.0]: https://github.com/PSLmodels/OG-BRA/compare/v0.1.0...v0.2.0
+[0.1.0]: https://github.com/PSLmodels/OG-BRA/compare/v0.0.0...v0.1.0

ogbra/__init__.py

@@ -8,4 +8,4 @@
 from ogbra.macro_params import *  # noqa: F403
 from ogbra.utils import *  # noqa: F403
 
-__version__ = "0.1.0"
+__version__ = "0.2.0"

ogbra/calibrate.py

@@ -90,6 +90,7 @@ def __init__(
                 country_id=UN_COUNTRY_CODE,
                 initial_data_year=p.start_year - 1,
                 final_data_year=p.start_year + 1,
+                income_percentiles=p.lambdas.flatten(),
                 GraphDiag=False,
                 download_path=demographic_data_path,
             )
@@ -104,6 +105,7 @@ def __init__(
                 country_id=UN_COUNTRY_CODE,
                 initial_data_year=p.start_year - 1,
                 final_data_year=p.start_year + 1,
+                income_percentiles=p.lambdas.flatten(),
                 GraphDiag=False,
             )
 

ogbra/income.py

@@ -115,14 +115,7 @@ def f(
     )
     a = x.root
     e_new = usa_params.e[0, :, :] * np.exp(a * usa_params.e[0, :, :])
-    emat_new_scaled = (
-        e_new
-        / (
-            e_new
-            * usa_params.omega_SS.reshape(usa_params.S, 1)
-            * usa_params.lambdas.reshape(1, usa_params.J)
-        ).sum()
-    )
+    emat_new_scaled = e_new / (e_new * usa_params.omega_SS).sum()
     # Now interpolate for the cases where S and/or J not the same in the
     # country parameterization as in the default USA parameterization
     if (
@@ -146,8 +139,8 @@ def f(
         emat_j_midp = np.zeros(M)
         pct_lb = 0.0
         for m in range(M):
-            emat_j_midp[m] = pct_lb + 0.5 * usa_params.lambdas[m]
-            pct_lb += usa_params.lambdas[m]
+            emat_j_midp[m] = pct_lb + 0.5 * float(usa_params.lambdas[m])
+            pct_lb += float(usa_params.lambdas[m])
 
         # Make sure that values in abil_midp are within interpolating
         # bounds
@@ -181,10 +174,7 @@ def f(
             (new_s_mesh, new_j_mesh),
             method="linear",
         )
-        emat_new_scaled = (
-            emat_new
-            / (emat_new * age_wgts.reshape(S, 1) * lambdas.reshape(1, J)).sum()
-        )
+        emat_new_scaled = emat_new / (emat_new * age_wgts).sum()
 
         if plot:
             kwargs = {"filesuffix": "_intrp_scaled"}