Add configuration and update Changelog

CHANGELOG.md

@@ -4,6 +4,9 @@
 
 ### New features
 
+* Add configuration facilities to Bambi (#745) 
+* Interpet submodule now outputs informative messages when computing default values (#745) 
+
 ### Maintenance and fixes
 
 ### Documentation

bambi/__init__.py

@@ -5,6 +5,7 @@
 from pymc import math
 
 from .backend import PyMCModel
+from .config import config
 from .data import clear_data_home, load_data
 from .families import Family, Likelihood, Link
 from .formula import Formula
@@ -23,6 +24,7 @@
     "PyMCModel",
     "Formula",
     "clear_data_home",
+    "config",
     "load_data",
     "math",
 ]

bambi/config.py

@@ -0,0 +1,51 @@
+class Config:
+    """Configuration variables for Bambi
+
+    It works with a pre-specified set of configuration variables and options for those.
+    When a user tries to set a configuration variable to a non-supported value, it raises an error.
+    """
+
+    __FIELDS = {"INTERPRET_VERBOSE": (True, False)}
+
+    def __init__(self, config_dict: dict = None):
+        config_dict = {} if config_dict is None else config_dict
+        # When an option is not specified at instantiation time, it uses the first value specified
+        # in __FIELDS.
+        for field, choices in Config.__FIELDS.items():
+            if field in config_dict:
+                value = config_dict[field]
+            else:
+                value = choices[0]
+            self[field] = value
+
+    def __setitem__(self, key, value):
+        setattr(self, key, value)
+
+    def __setattr__(self, key, value):
+        if key in Config.__FIELDS:
+            if value in Config.__FIELDS[key]:
+                super().__setattr__(key, value)
+            else:
+                raise ValueError(
+                    f"{value} is not a valid value for '{key}'"
+                    f"Valid options are: {Config.__FIELDS[key]}"
+                )
+        else:
+            raise KeyError(f"'{key}' is not a valid configuration option")
+
+    def __getitem__(self, key):
+        return getattr(self, key)
+
+    def __str__(self):  # pragma: no cover
+        lines = []
+        for field, choices in Config.__FIELDS.items():
+            lines.append(f"{field}: {self[field]} (available: {list(choices)})")
+        header = ["Bambi configuration"]
+        header.append("-" * len(header[0]))
+        return "\n".join(header + lines)
+
+    def __repr__(self):  # pragma: no cover
+        return str(self)
+
+
+config = Config()

bambi/interpret/__init__.py

@@ -1,4 +1,5 @@
-from bambi.interpret.config import logger
+import logging
+
 from bambi.interpret.effects import comparisons, predictions, slopes
 from bambi.interpret.plotting import plot_comparisons, plot_predictions, plot_slopes
 
@@ -11,3 +12,10 @@
     "plot_predictions",
     "plot_slopes",
 ]
+
+logger = logging.getLogger("__bambi_interpret__")
+
+if not logging.root.handlers:
+    logger.setLevel(logging.INFO)
+    if len(logger.handlers) == 0:
+        logger.addHandler(logging.StreamHandler())

bambi/interpret/logs.py

@@ -1,4 +1,6 @@
-from bambi.interpret import logger
+import logging
+
+from bambi import config
 
 
 def log_interpret_defaults(func):
@@ -10,11 +12,11 @@ def log_interpret_defaults(func):
     or 'wrt' of 'comparisons' and 'slopes', as well as the 'conditional'
     parameter of 'comparisons', 'predictions', and 'slopes'.
     """
-    interpret_logger = logger.get_logger("interpret")
+    logger = logging.getLogger("__bambi_interpret__")
 
     def wrapper(*args, **kwargs):
 
-        if not logger.messages:
+        if not config["INTERPRET_VERBOSE"]:
             return func(*args, **kwargs)
 
         arg_name = None
@@ -39,7 +41,7 @@ def wrapper(*args, **kwargs):
             covariate_name = args[0].name
 
         if arg_name:
-            interpret_logger.info("Default computed for %s variable: %s", arg_name, covariate_name)
+            logger.info("Default computed for %s variable: %s", arg_name, covariate_name)
 
         return func(*args, **kwargs)
 

tests/test_config.py

@@ -0,0 +1,21 @@
+import pytest
+
+from bambi.config import Config
+
+
+def test_config():
+    config = Config()
+
+    # Evaluate getters
+    assert config["INTERPRET_VERBOSE"] is True
+    assert config.INTERPRET_VERBOSE is True
+
+    config.INTERPRET_VERBOSE = False
+    assert config.INTERPRET_VERBOSE is False
+
+    # Evaluate setters
+    with pytest.raises(ValueError, match="anything is not a valid value for 'INTERPRET_VERBOSE'"):
+        config.INTERPRET_VERBOSE = "anything"
+
+    with pytest.raises(KeyError, match="'DOESNT_EXIST' is not a valid configuration option"):
+        config.DOESNT_EXIST = "anything"

tests/test_interpret_messages.py

@@ -1,13 +1,9 @@
-import numpy as np
+import bambi as bmb
 import pandas as pd
-import matplotlib.pyplot as plt
 import pytest
 
-import bambi as bmb
 from bambi.interpret import plot_comparisons, plot_predictions, plot_slopes
 
-bmb.interpret.logger.messages = True
-
 
 @pytest.fixture(scope="module")
 def mtcars():
@@ -22,7 +18,7 @@ def mtcars():
 
 def test_predictions_list(mtcars, caplog):
     model, idata = mtcars
-    caplog.set_level("INFO", logger="interpret")
+    caplog.set_level("INFO", logger="__bambi_interpret__")
 
     # List of values with no unspecified covariates
     conditional = ["hp", "drat", "am"]
@@ -39,7 +35,7 @@ def test_predictions_list(mtcars, caplog):
 
 def test_predictions_list_unspecified(mtcars, caplog):
     model, idata = mtcars
-    caplog.set_level("INFO", logger="interpret")
+    caplog.set_level("INFO", logger="__bambi_interpret__")
 
     # List of values with unspecified covariates
     conditional = ["hp", "drat"]
@@ -57,7 +53,7 @@ def test_predictions_list_unspecified(mtcars, caplog):
 
 def test_predictions_dict_unspecified(mtcars, caplog):
     model, idata = mtcars
-    caplog.set_level("INFO", logger="interpret")
+    caplog.set_level("INFO", logger="__bambi_interpret__")
 
     # User passed values with unspecified covariates
     conditional = {"hp": [110, 175], "am": [0, 1]}
@@ -79,7 +75,7 @@ def test_predictions_dict_unspecified(mtcars, caplog):
 
 def test_comparisons_contrast_default(mtcars, caplog):
     model, idata = mtcars
-    caplog.set_level("INFO", logger="interpret")
+    caplog.set_level("INFO", logger="__bambi_interpret__")
 
     # List of values with no unspecified covariates
     plot_comparisons(model, idata, "hp", conditional=None, average_by="am")
@@ -93,7 +89,7 @@ def test_comparisons_contrast_default(mtcars, caplog):
 
 def test_slopes_wrt_default(mtcars, caplog):
     model, idata = mtcars
-    caplog.set_level("INFO", logger="interpret")
+    caplog.set_level("INFO", logger="__bambi_interpret__")
 
     # List of values with no unspecified covariates
     plot_slopes(model, idata, "hp", conditional=None, average_by="am")