Reorganize code. Remove circular depenency in ResultData

tea/api.py

@@ -1,19 +1,16 @@
-from tea.build import (load_data, load_data_from_url, const,
+from .build import (load_data, load_data_from_url, const,
                     ordinal, isordinal,
                     nominal, isnominal,
                     ratio, isratio,
                     interval, isinterval, isnumeric,
                     select, compare, relate, predict,
                     get_var_from_list
-                    # , nominal, interval, ratio, load_data, model, 
-                    # mean, median, standard_deviation, variance, kurtosis, skew, normality, frequency,
-                    # between_experiment, within_experiment, mixed_experiment, model, equation,
-                    # load_data_arrs, hypothesis, experiment_design
                    )
-from tea.evaluate import evaluate
-from tea.evaluate_helper_methods import determine_study_type, assign_roles
-from tea.evaluate_data_structures import BivariateData, MultivariateData
-from .solver import which_props
+from .evaluate import evaluate
+import tea.helpers
+import tea.runtimeDataStructures
+import tea.z3_solver
+
 from typing import Dict
 from .global_vals import *
 

tea/ast.py

@@ -4,8 +4,6 @@
 
 from enum import Enum
 from typing import Dict, Union
-# from collections import OrderedDict
-# from pandas.api.types import CategoricalDtype
 
 class Node(object): 
     def relate(self, other):

tea/build.py

@@ -1,9 +1,11 @@
-from typing import Dict, Union
-from collections import OrderedDict
+from tea.runtimeDataStructures.dataset import Dataset
 from tea.ast import (  Variable, DataType, Literal,
                     Relate, Relationship
                 )
-from tea.dataset import Dataset
+
+
+from typing import Dict, Union
+from collections import OrderedDict
 
 iv_identifier = 'independent variable'
 dv_identifier = 'dependent variable'

tea/evaluate.py

@@ -1,21 +1,27 @@
-from tea.ast import *
-from tea.dataset import Dataset
-from tea.evaluate_data_structures import VarData, BivariateData, MultivariateData, ResultData # runtime data structures
-# from tea.evaluate_result_data_structures import ResultData
-from tea.evaluate_helper_methods import determine_study_type, assign_roles, add_paired_property, compute_data_properties, compute_combined_data_properties, execute_test, correct_multiple_comparison
-from .solver import synthesize_tests
-# from tea.solver import find_applicable_bivariate_tests
+from tea.ast import (   Node, Variable, Literal, 
+                        Equal, NotEqual, LessThan, 
+                        LessThanEqual, GreaterThan, GreaterThanEqual,
+                        Relate
+                    )
+from tea.runtimeDataStructures.dataset import Dataset
+from tea.runtimeDataStructures.varData import VarData
+from tea.runtimeDataStructures.bivariateData import BivariateData
+from tea.runtimeDataStructures.multivariateData import MultivariateData
+from tea.runtimeDataStructures.resultData import ResultData
+from tea.helpers.evaluateHelperMethods import determine_study_type, assign_roles, add_paired_property, execute_test, correct_multiple_comparison
+from tea.z3_solver.solver import synthesize_tests
 
 import attr
 from typing import Any
 from types import SimpleNamespace # allows for dot notation access for dictionaries
+from typing import Dict
 
 from scipy import stats # Stats library used
 import statsmodels.api as sm
 import statsmodels.formula.api as smf
 import numpy as np # Use some stats from numpy instead
 import pandas as pd
-# import bootstrapped as bs
+
 
 # TODO: Pass participant_id as part of experimental design, not load_data
 def evaluate(dataset: Dataset, expr: Node, assumptions: Dict[str, str], design: Dict[str, str]=None):
@@ -385,7 +391,6 @@ def evaluate(dataset: Dataset, expr: Node, assumptions: Dict[str, str], design:
         if len(tests) == 0: 
             tests.append('bootstrap') # Default to bootstrap
 
-
         for test in tests: 
             test_result = execute_test(dataset, design, expr.predictions, combined_data, test)
             results[test] = test_result
@@ -400,7 +405,7 @@ def evaluate(dataset: Dataset, expr: Node, assumptions: Dict[str, str], design:
             # There are multiple comparisons
             # if len(preds > 1): 
             # FOR DEBUGGING: 
-            if len(preds >= 1): 
+            if len(preds) >= 1: 
                 correct_multiple_comparison(res_data)
 
         return res_data

tea/helpers/evaluate_helper_methods.py → tea/helpers/evaluateHelperMethods.py

@@ -1,28 +1,32 @@
+# Tea
 from tea.global_vals import *
-from tea.ast import *
-from tea.dataset import Dataset
-from tea.evaluate_data_structures import VarData, CombinedData, BivariateData, MultivariateData, ResultData
-# from tea.evaluate_result_data_structures import ResultData
-# from tea.solver import Tests, Assumptions
-
-import attr
-from typing import Any, Dict, List
-from types import SimpleNamespace # allows for dot notation access for dictionaries
-from collections import namedtuple
-from enum import Enum
-import copy
-
+from tea.ast import DataType, GreaterThan
+from tea.runtimeDataStructures.dataset import Dataset
+from tea.runtimeDataStructures.varData import VarData
+from tea.runtimeDataStructures.combinedData import CombinedData
+from tea.runtimeDataStructures.bivariateData import BivariateData
+from tea.runtimeDataStructures.multivariateData import MultivariateData
+from tea.runtimeDataStructures.resultData import ResultData
+
+# Stats
 from scipy import stats # Stats library used
 from sklearn import preprocessing # for creating interaction effects
 import statsmodels.api as sm
 import statsmodels.formula.api as smf
 from statsmodels.formula.api import ols
-# import numpy as np # Use some stats from numpy instead
+
 import pandas as pd
 from statsmodels.stats.anova import AnovaRM
 import bootstrapped.bootstrap as bs
 import bootstrapped.stats_functions as bs_stats
 
+# Other
+import attr
+from typing import Any, Dict, List
+from types import SimpleNamespace # allows for dot notation access for dictionaries
+from collections import namedtuple
+from enum import Enum
+import copy
 
 def determine_study_type(vars_data: list, design: Dict[str, str]):
     if design: 
@@ -837,77 +841,4 @@ def execute_test(dataset, design, predictions, combined_data: CombinedData, test
 # Correct for multiple comparisons
 def correct_multiple_comparison(res_data: ResultData): 
     # TODO: refactor ResultData first. 
-    pass    
-
-"""
-def explanatory_strings_for_assumptions(assumptions: Assumptions) -> List[str]:
-    explanation = []
-    if assumptions & Assumptions.INDEPENDENT_OBSERVATIONS:
-        explanation.append("Assumes independent observations.")
-        assumptions &= ~Assumptions.INDEPENDENT_OBSERVATIONS
-
-    if assumptions & Assumptions.NORMALLY_DISTRIBUTED_VARIABLES:
-        explanation.append("Assumes samples are normally distributed.")
-        assumptions &= ~Assumptions.NORMALLY_DISTRIBUTED_VARIABLES
-
-    if assumptions & Assumptions.NORMALLY_DISTRIBUTED_DIFFERENCE_BETWEEN_VARIABLES:
-        explanation.append("Assumes difference between paired values is normally distributed.")
-        assumptions &= ~Assumptions.NORMALLY_DISTRIBUTED_DIFFERENCE_BETWEEN_VARIABLES
-
-    if assumptions & Assumptions.SYMMETRICALLY_DISTRIBUTED_DIFFERENCE_BETWEEN_VARIABLES:
-        explanation.append("Assumes difference between paired values is symmetrically distributed.")
-        assumptions &= ~Assumptions.SYMMETRICALLY_DISTRIBUTED_DIFFERENCE_BETWEEN_VARIABLES
-
-    if assumptions & Assumptions.SIMILAR_VARIANCES:
-        explanation.append("Assumes samples have similar variances.")
-        assumptions &= ~Assumptions.SIMILAR_VARIANCES
-
-    if assumptions & Assumptions.LARGE_SAMPLE_SIZE:
-        explanation.append("Assumes a large enough sample size.")
-        assumptions &= ~Assumptions.LARGE_SAMPLE_SIZE
-
-    if assumptions & Assumptions.VALUES_ARE_FREQUENCIES:
-        explanation.append("Assumes values are frequencies (and not, e.g., percentages).")
-        assumptions &= ~Assumptions.VALUES_ARE_FREQUENCIES
-
-    if assumptions & Assumptions.PAIRED_OBSERVATIONS:
-        explanation.append("Assumes observations are paired (e.g. within subjects).")
-        assumptions &= ~Assumptions.PAIRED_OBSERVATIONS
-
-    if assumptions & Assumptions.NO_OUTLIERS:
-        explanation.append("Assumes there are no outliers in the data.")
-        assumptions &= ~Assumptions.NO_OUTLIERS
-
-    if assumptions & Assumptions.NO_OUTLIERS_IN_DIFFERENCE_BETWEEN_VARIABLES:
-        explanation.append("Assumes there are no outliers in the difference between paired values.")
-        assumptions &= ~Assumptions.NO_OUTLIERS_IN_DIFFERENCE_BETWEEN_VARIABLES
-
-    if assumptions & Assumptions.LINEAR_RELATIONSHIP:
-        explanation.append("Assumes there is a linear relationship between the variables.")
-        assumptions &= ~Assumptions.LINEAR_RELATIONSHIP
-
-    if assumptions & Assumptions.BIVARIATE_NORMAL_VARIABLES:
-        explanation.append("Assumes the two variables have a bivariate normal distribution.")
-        assumptions &= ~Assumptions.BIVARIATE_NORMAL_VARIABLES
-
-    if assumptions & Assumptions.RELATED_SAMPLES:
-        explanation.append("Assumes the samples come from related sources (e.g. within subjects).")
-        assumptions &= ~Assumptions.RELATED_SAMPLES
-
-    if assumptions & Assumptions.MONOTONIC_RELATIONSHIP:
-        explanation.append("Assumes there is a monotonic relationship between the variables.")
-        assumptions &= ~Assumptions.MONOTONIC_RELATIONSHIP
-
-    if assumptions & Assumptions.ALL_VARIABLES_CONTINUOUS_OR_ORDINAL:
-        explanation.append("Assumes all variables are continuous or ordinal.")
-        assumptions &= ~Assumptions.ALL_VARIABLES_CONTINUOUS_OR_ORDINAL
-
-    if assumptions & Assumptions.DEPENDENT_VARIABLE_CONTINUOUS_OR_ORDINAL:
-        explanation.append("Assumes the dependent variable is continuous or ordinal.")
-        assumptions &= ~Assumptions.DEPENDENT_VARIABLE_CONTINUOUS_OR_ORDINAL
-
-    assert assumptions == Assumptions.NONE, \
-        "Not all assumptions have a corresponding explanatory string: %s" % assumptions
-
-    return explanation
-"""
+    pass

tea/runtime_data_structures/bivariate_data.py → tea/runtimeDataStructures/bivariateData.py

@@ -1,4 +1,5 @@
 import attr
+from .combinedData import CombinedData
 
 @attr.s(init=True, auto_attribs=True)
 class BivariateData(CombinedData):

tea/runtime_data_structures/combined_data.py → tea/runtimeDataStructures/combinedData.py

@@ -1,5 +1,7 @@
+from tea.global_vals import *
+from tea.ast import *
 import attr
-import value
+from .value import Value
 
 # CombinedData is the runtime data structure used to unify experimental design and variable declarations
 @attr.s(init=True)

tea/runtime_data_structures/multivariate_data.py → tea/runtimeDataStructures/multivariateData.py

@@ -1,5 +1,6 @@
 import attr
-import value
+from .value import Value
+from .combinedData import CombinedData
 
 @attr.s(init=True, auto_attribs=True)
 class MultivariateData(CombinedData):

tea/z3_solver/result_data.py → tea/runtimeDataStructures/resultData.py

@@ -1,5 +1,7 @@
+# from tea.z3_solver.solver import __ALL_TESTS__
+from tea.runtimeDataStructures.value import Value
+
 import attr
-from .solver import __ALL_TESTS__
 
 @attr.s(init=False, repr=False, str=False)
 class ResultData(Value):
@@ -9,22 +11,22 @@ class ResultData(Value):
     def __init__(self, test_to_results):
         self.test_to_results = test_to_results
         self.test_to_assumptions = {}
-        for test in __ALL_TESTS__:
-            if test.name in test_to_results:
-                print(test.name)
-                test_assumptions = []
-                # TODO: The names get stale if hypothesize() is called multiple times in a row.
-                for applied_prop in test._properties:
-                    assumption = f"{applied_prop.property.description}: "
-                    for stat_var in applied_prop.vars:
-                        assumption += f"{stat_var.name}, "
-
-                    assumption = assumption[:-2]
-                    if applied_prop.property_test_results is not None:
-                        assumption += f": {applied_prop.property_test_results}"
-                    test_assumptions.append(assumption)
-
-                self.test_to_assumptions[test.name] = test_assumptions
+        # for test in __ALL_TESTS__:
+        #     if test.name in test_to_results:
+        #         print(test.name)
+        #         test_assumptions = []
+        #         # TODO: The names get stale if hypothesize() is called multiple times in a row.
+        #         for applied_prop in test._properties:
+        #             assumption = f"{applied_prop.property.description}: "
+        #             for stat_var in applied_prop.vars:
+        #                 assumption += f"{stat_var.name}, "
+
+        #             assumption = assumption[:-2]
+        #             if applied_prop.property_test_results is not None:
+        #                 assumption += f": {applied_prop.property_test_results}"
+        #             test_assumptions.append(assumption)
+
+        #         self.test_to_assumptions[test.name] = test_assumptions
 
     def _pretty_print(self):
         output = "\nResults:\n--------------"

tea/runtime_data_structures/var_data.py → tea/runtimeDataStructures/varData.py

@@ -1,5 +1,8 @@
+from tea.global_vals import *
+from .value import Value
+from tea.ast import DataType
+
 import attr
-import value
 
 @attr.s(init=True)
 class VarData(Value):

tea/z3_solver/solver.py

@@ -1,12 +1,12 @@
+from tea.global_vals import *
+from tea.runtimeDataStructures.dataset import Dataset
+from tea.runtimeDataStructures.varData import VarData
+from tea.runtimeDataStructures.combinedData import CombinedData
+from tea.runtimeDataStructures.bivariateData import BivariateData
+from tea.helpers.evaluateHelperMethods import get_data, compute_normal_distribution, compute_eq_variance
+
 import attr
-# from z3 import *
 import z3
-# from z3 import BoolSort
-# from z3 import z3.is_true 
-from tea.dataset import Dataset
-from tea.evaluate_data_structures import VarData, CombinedData, BivariateData, MultivariateData
-from tea.evaluate_helper_methods import get_data, compute_normal_distribution, compute_eq_variance
-from tea.global_vals import *
 from typing import Dict, List
 
 # Prog -> List[StatisticalTest] -> Query