Remove rpy2 dependency

.github/workflows/build_and_test.yml

@@ -27,18 +27,14 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [ubuntu-latest, macos-latest]
+        os: [ubuntu-latest, macos-latest, windows-latest]
         python-version: ["3.10", "3.11"]
 
     steps:
       - name: Checkout
         uses: actions/checkout@v4
         with:
           persist-credentials: false
-      - name: Set up R ${{ matrix.r-version }}
-        uses: r-lib/actions/setup-r@v2
-        with:
-          r-version: release
       - name: Setup Miniconda using Python ${{ matrix.python-version }}
         uses: conda-incubator/setup-miniconda@v3
         with:

CHANGELOG.md

@@ -5,6 +5,14 @@ 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).
 
+
+## [0.1.10] - 2024-06-10 12:00:00
+
+### Added
+
+- Removes the `rpy2` dependency from the `environment.yml` and `setup.py` files, and modifies use of PSID data to avoid needing this package in OG-USA.
+
+
 ## [0.1.9] - 2024-06-07 12:00:00
 
 ### Added
@@ -111,6 +119,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 
 
+[0.1.10]: https://github.com/PSLmodels/OG-USA/compare/v0.1.9...v0.1.10
 [0.1.9]: https://github.com/PSLmodels/OG-USA/compare/v0.1.8...v0.1.9
 [0.1.8]: https://github.com/PSLmodels/OG-USA/compare/v0.1.7...v0.1.8
 [0.1.7]: https://github.com/PSLmodels/OG-USA/compare/v0.1.6...v0.1.7

cs-config/cs_config/functions.py

@@ -1,3 +1,4 @@
+import ogusa
 from ogusa.calibrate import Calibration
 from ogcore.parameters import Specifications
 from ogusa.constants import (
@@ -82,7 +83,7 @@ class MetaParams(paramtools.Parameters):
 
 
 def get_version():
-    return "0.1.9"
+    return ogusa.__version__
 
 
 def get_inputs(meta_param_dict):

data/PSID/psid_download.R

@@ -176,4 +176,5 @@ for (var in names(ind_var_names)){
 print('Beginning to build panel')
 # Build PSID panel
 psid_df <- build.panel(datadir=mydir, fam.vars=famvars, ind.vars=indvars, sample="SRC", design='all')
-save(psid_df,file=file.path(script.dir, 'psid1968to2017.RData'))
+
+write.csv(psid_df, file=gzfile("psid1968to2017.csv.gz"))

environment.yml

@@ -30,7 +30,6 @@ dependencies:
 - black>=24.1.1
 - pip
 - pip:
-  - rpy2>=3.5.12
   - openpyxl>=3.1.2
   - cs-kit
   - cs2tc

ogusa/__init__.py

@@ -0,0 +1,14 @@
+"""
+Specify what is available to import from the ogusa package.
+"""
+
+from ogusa.calibrate import *
+from ogusa.bequest_transmission import *
+from ogusa.get_micro_data import *
+from ogusa.income import *
+from ogusa.macro_params import *
+from ogusa.transfer_distribution import *
+from ogusa.utils import *
+from ogusa.wealth import *
+
+__version__ = "0.1.10"

ogusa/psid_data_setup.py

@@ -5,17 +5,9 @@
 from pandas_datareader import data as web
 import datetime
 from linearmodels import PanelOLS
-from rpy2.robjects import r
-from rpy2.robjects import conversion
-from rpy2.robjects import default_converter
-from rpy2.robjects import pandas2ri
-from rpy2.robjects.packages import importr
 from ogusa.constants import PSID_NOMINAL_VARS, PSID_CONSTANT_VARS
 
 
-pandas2ri.activate()
-pd.options.mode.chained_assignment = "raise"
-
 try:
     # This is the case when a separate script is calling this function in
     # this module
@@ -29,7 +21,9 @@
     os.makedirs(output_dir)
 
 
-def prep_data(data="psid1968to2015.RData"):
+def prep_data(
+    data=os.path.join(CURDIR, "..", "data", "PSID", "psid1968to2015.csv.gz")
+):
     """
     This script takes PSID data created from psid_download.R and:
     1) Creates variables at the "tax filing unit" (equal to family
@@ -47,10 +41,7 @@ def prep_data(data="psid1968to2015.RData"):
             income groups defined
     """
     # Read data from R into pandas dataframe
-    r["load"](os.path.join(CURDIR, "..", "data", "PSID", data))
-    raw_r_df = r("psid_df")
-    with (default_converter + pandas2ri.converter).context():
-        raw_df = conversion.get_conversion().rpy2py(raw_r_df)
+    raw_df = pd.read_csv(data, compression="gzip")
 
     # Create unique identifier for each household
     # note that will define a new household if head or spouse changes

setup.py

@@ -5,7 +5,7 @@
 
 setuptools.setup(
     name="ogusa",
-    version="0.1.9",
+    version="0.1.10",
     author="Jason DeBacker and Richard W. Evans",
     license="CC0 1.0 Universal (CC0 1.0) Public Domain Dedication",
     description="USA calibration for OG-Core",
@@ -37,7 +37,6 @@
         "paramtools>=0.15.0",
         "taxcalc>=3.0.0",
         "requests",
-        "rpy2>=3.5.12",
         "pandas-datareader",
         "xlwt",
         "openpyxl>=3.1.2",

tests/test_utils.py

@@ -1,12 +1,12 @@
-from ogusa import utils
+from ogusa.utils import read_cbo_forecast
 import numpy as np
 
 
 def test_read_cbo_forecast():
     """
     Test that CBO data read as expected.
     """
-    test_df = utils.read_cbo_forecast()
+    test_df = read_cbo_forecast()
 
     assert np.allclose(
         test_df.loc[test_df["year"] == 2017, "Y"].values[0], 20344