refactor: prefix internal methods and modules with underscore

Author: mplatzer

mostlyai/qa/accuracy.py renamed to mostlyai/qa/_accuracy.py

@@ -26,7 +26,7 @@
 import scipy.stats
 from joblib import Parallel, delayed, parallel_config, cpu_count
 
-from mostlyai.qa.common import (
+from mostlyai.qa._common import (
     CHARTS_COLORS,
     CHARTS_FONTS,
     EMPTY_BIN,
@@ -46,7 +46,7 @@
 )
 from plotly import graph_objs as go
 
-from mostlyai.qa.filesystem import TemporaryWorkspace, Statistics
+from mostlyai.qa._filesystem import TemporaryWorkspace, Statistics
 
 _LOG = logging.getLogger(__name__)
 

mostlyai/qa/common.py renamed to mostlyai/qa/_common.py

@@ -19,7 +19,7 @@
 import pandas as pd
 from rich.progress import Progress
 
-from mostlyai.qa.filesystem import Statistics
+from mostlyai.qa._filesystem import Statistics
 
 _LOG = logging.getLogger(__name__)
 

mostlyai/qa/distances.py renamed to mostlyai/qa/_distances.py

@@ -17,11 +17,11 @@
 import numpy as np
 from joblib import cpu_count
 
-from mostlyai.qa.common import (
+from mostlyai.qa._common import (
     CHARTS_COLORS,
     CHARTS_FONTS,
 )
-from mostlyai.qa.filesystem import TemporaryWorkspace
+from mostlyai.qa._filesystem import TemporaryWorkspace
 from plotly import graph_objs as go
 from sklearn.neighbors import NearestNeighbors
 

mostlyai/qa/filesystem.py renamed to mostlyai/qa/_filesystem.py

@@ -19,8 +19,8 @@
 
 import pandas as pd
 from jinja2 import Environment, FileSystemLoader
-from mostlyai.qa.accuracy import trim_label, filter_uni_acc_for_plotting, filter_biv_acc_for_plotting
-from mostlyai.qa.filesystem import TemporaryWorkspace
+from mostlyai.qa._accuracy import trim_label, filter_uni_acc_for_plotting, filter_biv_acc_for_plotting
+from mostlyai.qa._filesystem import TemporaryWorkspace
 from mostlyai.qa.assets import (
     HTML_ASSETS_PATH,
     read_html_assets,

mostlyai/qa/html_report.py renamed to mostlyai/qa/_html_report.py

@@ -34,7 +34,7 @@
 import pyarrow as pa
 import torch
 
-from mostlyai.qa.common import (
+from mostlyai.qa._common import (
     CTX_COLUMN_PREFIX,
     TGT_COLUMN_PREFIX,
     NXT_COLUMN_PREFIX,

mostlyai/qa/sampling.py renamed to mostlyai/qa/_sampling.py

@@ -19,11 +19,11 @@
 from sklearn.decomposition import PCA
 from sklearn.metrics.pairwise import cosine_similarity
 
-from mostlyai.qa.common import (
+from mostlyai.qa._common import (
     CHARTS_FONTS,
     CHARTS_COLORS,
 )
-from mostlyai.qa.filesystem import TemporaryWorkspace
+from mostlyai.qa._filesystem import TemporaryWorkspace
 import scipy.stats
 from sklearn.model_selection import StratifiedKFold
 from sklearn.linear_model import LogisticRegression

mostlyai/qa/similarity.py renamed to mostlyai/qa/_similarity.py

@@ -20,8 +20,8 @@
 import pandas as pd
 from pandas.core.dtypes.common import is_numeric_dtype, is_datetime64_dtype
 
-from mostlyai.qa import distances, similarity, html_report
-from mostlyai.qa.accuracy import (
+from mostlyai.qa import _distances, _similarity, _html_report
+from mostlyai.qa._accuracy import (
     binning_data,
     calculate_correlations,
     plot_store_correlation_matrices,
@@ -37,8 +37,8 @@
     plot_store_bivariates,
 )
 from mostlyai.qa.metrics import Metrics, Accuracy, Similarity, Distances
-from mostlyai.qa.sampling import calculate_embeddings, pull_data_for_accuracy, pull_data_for_embeddings
-from mostlyai.qa.common import (
+from mostlyai.qa._sampling import calculate_embeddings, pull_data_for_accuracy, pull_data_for_embeddings
+from mostlyai.qa._common import (
     determine_data_size,
     ProgressCallback,
     PrerequisiteNotMetError,
@@ -49,7 +49,7 @@
     REPORT_CREDITS,
     ProgressCallbackWrapper,
 )
-from mostlyai.qa.filesystem import Statistics, TemporaryWorkspace
+from mostlyai.qa._filesystem import Statistics, TemporaryWorkspace
 
 _LOG = logging.getLogger(__name__)
 
@@ -166,7 +166,7 @@ def report(
         except PrerequisiteNotMetError as err:
             _LOG.info(err)
             statistics.mark_early_exit()
-            html_report.store_early_exit_report(report_path)
+            _html_report.store_early_exit_report(report_path)
             return report_path, None
 
         # prepare datasets for accuracy
@@ -217,7 +217,7 @@ def report(
             syn[col] = syn[col].astype(trn[col].dtype)
 
         _LOG.info("report accuracy and correlations")
-        acc_uni, acc_biv, corr_trn = report_accuracy_and_correlations(
+        acc_uni, acc_biv, corr_trn = _report_accuracy_and_correlations(
             trn=trn,
             syn=syn,
             statistics=statistics,
@@ -286,7 +286,7 @@ def report(
         progress.update(completed=80, total=100)
 
         _LOG.info("report similarity")
-        sim_cosine_trn_hol, sim_cosine_trn_syn, sim_auc_trn_hol, sim_auc_trn_syn = report_similarity(
+        sim_cosine_trn_hol, sim_cosine_trn_syn, sim_auc_trn_hol, sim_auc_trn_syn = _report_similarity(
             syn_embeds=syn_embeds,
             trn_embeds=trn_embeds,
             hol_embeds=hol_embeds,
@@ -296,15 +296,15 @@ def report(
         progress.update(completed=90, total=100)
 
         _LOG.info("report distances")
-        dcr_trn, dcr_hol = report_distances(
+        dcr_trn, dcr_hol = _report_distances(
             syn_embeds=syn_embeds,
             trn_embeds=trn_embeds,
             hol_embeds=hol_embeds,
             workspace=workspace,
         )
         progress.update(completed=99, total=100)
 
-        metrics = calculate_metrics(
+        metrics = _calculate_metrics(
             acc_uni=acc_uni,
             acc_biv=acc_biv,
             dcr_trn=dcr_trn,
@@ -329,7 +329,7 @@ def report(
             "report_extra_info": report_extra_info,
         }
         statistics.store_meta(meta=meta)
-        html_report.store_report(
+        _html_report.store_report(
             report_path=report_path,
             report_type="model_report",
             workspace=workspace,
@@ -343,7 +343,7 @@ def report(
         return report_path, metrics
 
 
-def calculate_metrics(
+def _calculate_metrics(
     *,
     acc_uni: pd.DataFrame | None = None,
     acc_biv: pd.DataFrame | None = None,
@@ -419,7 +419,7 @@ def calculate_metrics(
     )
 
 
-def report_accuracy_and_correlations(
+def _report_accuracy_and_correlations(
     *,
     trn: pd.DataFrame,
     syn: pd.DataFrame,
@@ -529,7 +529,7 @@ def report_accuracy_and_correlations(
     return acc_uni, acc_biv, trn_corr
 
 
-def report_similarity(
+def _report_similarity(
     *,
     syn_embeds: np.ndarray,
     trn_embeds: np.ndarray,
@@ -538,17 +538,17 @@ def report_similarity(
     statistics: Statistics,
 ) -> tuple[np.float64 | None, np.float64, np.float64 | None, np.float64]:
     _LOG.info("calculate centroid similarities")
-    sim_cosine_trn_hol, sim_cosine_trn_syn = similarity.calculate_cosine_similarities(
+    sim_cosine_trn_hol, sim_cosine_trn_syn = _similarity.calculate_cosine_similarities(
         syn_embeds=syn_embeds, trn_embeds=trn_embeds, hol_embeds=hol_embeds
     )
 
     _LOG.info("calculate discriminator AUC")
-    sim_auc_trn_hol, sim_auc_trn_syn = similarity.calculate_discriminator_auc(
+    sim_auc_trn_hol, sim_auc_trn_syn = _similarity.calculate_discriminator_auc(
         syn_embeds=syn_embeds, trn_embeds=trn_embeds, hol_embeds=hol_embeds
     )
 
     _LOG.info("plot and store PCA similarity contours")
-    pca_model, _, trn_pca, hol_pca = similarity.plot_store_similarity_contours(
+    pca_model, _, trn_pca, hol_pca = _similarity.plot_store_similarity_contours(
         syn_embeds=syn_embeds, trn_embeds=trn_embeds, hol_embeds=hol_embeds, workspace=workspace
     )
 
@@ -566,15 +566,15 @@ def report_similarity(
     )
 
 
-def report_distances(
+def _report_distances(
     *,
     syn_embeds: np.ndarray,
     trn_embeds: np.ndarray,
     hol_embeds: np.ndarray | None,
     workspace: TemporaryWorkspace,
 ) -> tuple[np.ndarray, np.ndarray | None]:
-    dcr_trn, dcr_hol = distances.calculate_distances(
+    dcr_trn, dcr_hol = _distances.calculate_distances(
         syn_embeds=syn_embeds, trn_embeds=trn_embeds, hol_embeds=hol_embeds
     )
-    distances.plot_store_distances(dcr_trn, dcr_hol, workspace)
+    _distances.plot_store_distances(dcr_trn, dcr_hol, workspace)
     return dcr_trn, dcr_hol

mostlyai/qa/report.py

@@ -18,11 +18,9 @@
 import numpy as np
 import pandas as pd
 
-import mostlyai.qa.accuracy
-import mostlyai.qa.sampling
-from mostlyai.qa import accuracy, similarity, html_report
-from mostlyai.qa.sampling import pull_data_for_embeddings, calculate_embeddings
-from mostlyai.qa.common import (
+from mostlyai.qa import _accuracy, _sampling, _similarity, _html_report
+from mostlyai.qa._sampling import pull_data_for_embeddings, calculate_embeddings
+from mostlyai.qa._common import (
     ProgressCallback,
     PrerequisiteNotMetError,
     check_min_sample_size,
@@ -31,7 +29,7 @@
     REPORT_CREDITS,
     ProgressCallbackWrapper,
 )
-from mostlyai.qa.filesystem import Statistics, TemporaryWorkspace
+from mostlyai.qa._filesystem import Statistics, TemporaryWorkspace
 
 _LOG = logging.getLogger(__name__)
 
@@ -72,7 +70,7 @@ def report_from_statistics(
             check_statistics_prerequisite(statistics)
             check_min_sample_size(syn_sample_size, 100, "synthetic")
         except PrerequisiteNotMetError:
-            html_report.store_early_exit_report(report_path)
+            _html_report.store_early_exit_report(report_path)
             return report_path
 
         meta = statistics.load_meta()
@@ -87,7 +85,7 @@ def report_from_statistics(
 
         # prepare data
         _LOG.info("sample synthetic data started")
-        syn = mostlyai.qa.sampling.pull_data_for_accuracy(
+        syn = _sampling.pull_data_for_accuracy(
             df_tgt=syn_tgt_data,
             df_ctx=syn_ctx_data,
             ctx_primary_key=ctx_primary_key,
@@ -98,7 +96,7 @@ def report_from_statistics(
         progress.update(completed=20, total=100)
 
         # calculate and plot accuracy and correlations
-        acc_uni, acc_biv, corr_trn = report_accuracy_and_correlations_from_statistics(
+        acc_uni, acc_biv, corr_trn = _report_accuracy_and_correlations_from_statistics(
             syn=syn,
             statistics=statistics,
             workspace=workspace,
@@ -120,7 +118,7 @@ def report_from_statistics(
         )
 
         _LOG.info("report similarity")
-        report_similarity_from_statistics(
+        _report_similarity_from_statistics(
             syn_embeds=syn_embeds,
             workspace=workspace,
             statistics=statistics,
@@ -136,7 +134,7 @@ def report_from_statistics(
         }
 
         # HTML report
-        html_report.store_report(
+        _html_report.store_report(
             report_path=report_path,
             report_type="data_report",
             workspace=workspace,
@@ -150,7 +148,7 @@ def report_from_statistics(
         return report_path
 
 
-def report_accuracy_and_correlations_from_statistics(
+def _report_accuracy_and_correlations_from_statistics(
     *,
     syn: pd.DataFrame,
     statistics: Statistics,
@@ -161,7 +159,7 @@ def report_accuracy_and_correlations_from_statistics(
     syn = syn[bins.keys()].copy()
 
     _LOG.info("calculate synthetic bins")
-    syn_bin, _ = mostlyai.qa.accuracy.bin_data(syn, bins)
+    syn_bin, _ = _accuracy.bin_data(syn, bins)
 
     _LOG.info("load univariates and bivariates")
     acc_uni = statistics.load_univariate_accuracies()
@@ -180,31 +178,31 @@ def report_accuracy_and_correlations_from_statistics(
     corr_trn = statistics.load_correlations()
 
     _LOG.info("calculate synthetic correlations")
-    corr_syn = mostlyai.qa.accuracy.calculate_correlations(binned=syn_bin, corr_cols=corr_trn.columns)
+    corr_syn = _accuracy.calculate_correlations(binned=syn_bin, corr_cols=corr_trn.columns)
 
     _LOG.info("plot correlations")
-    mostlyai.qa.accuracy.plot_store_correlation_matrices(corr_trn=corr_trn, corr_syn=corr_syn, workspace=workspace)
+    _accuracy.plot_store_correlation_matrices(corr_trn=corr_trn, corr_syn=corr_syn, workspace=workspace)
 
     _LOG.info("filter columns for plotting")
     syn = syn[acc_uni["column"]]
     acc_cols = list(set(acc_uni["column"]) | set(acc_biv["col1"]) | set(acc_biv["col2"]))
     syn_bin = syn_bin[acc_cols]
 
     _LOG.info("calculate numeric KDEs for synthetic")
-    syn_num_kdes = accuracy.calculate_numeric_uni_kdes(df=syn, trn_kdes=trn_num_kdes)
+    syn_num_kdes = _accuracy.calculate_numeric_uni_kdes(df=syn, trn_kdes=trn_num_kdes)
 
     _LOG.info("calculate categorical counts for synthetic")
-    syn_cat_uni_cnts = accuracy.calculate_categorical_uni_counts(
+    syn_cat_uni_cnts = _accuracy.calculate_categorical_uni_counts(
         df=syn,
         trn_col_counts=trn_cat_uni_cnts,
         hash_rare_values=False,
     )
 
     _LOG.info("calculate bin counts for synthetic")
-    syn_bin_cnts_uni, syn_bin_cnts_biv = accuracy.calculate_bin_counts(syn_bin)
+    syn_bin_cnts_uni, syn_bin_cnts_biv = _accuracy.calculate_bin_counts(syn_bin)
 
     _LOG.info("plot univariates")
-    accuracy.plot_store_univariates(
+    _accuracy.plot_store_univariates(
         trn_num_kdes=trn_num_kdes,
         syn_num_kdes=syn_num_kdes,
         trn_cat_cnts=trn_cat_uni_cnts,
@@ -217,7 +215,7 @@ def report_accuracy_and_correlations_from_statistics(
     )
 
     _LOG.info("plot bivariates")
-    accuracy.plot_store_bivariates(
+    _accuracy.plot_store_bivariates(
         trn_cnts_uni=trn_bin_cnts_uni,
         syn_cnts_uni=syn_bin_cnts_uni,
         trn_cnts_biv=trn_bin_cnts_biv,
@@ -230,7 +228,7 @@ def report_accuracy_and_correlations_from_statistics(
     return acc_uni, acc_biv, corr_trn
 
 
-def report_similarity_from_statistics(
+def _report_similarity_from_statistics(
     *,
     syn_embeds: np.ndarray,
     statistics: Statistics,
@@ -246,6 +244,6 @@ def report_similarity_from_statistics(
     trn_pca, hol_pca = statistics.load_trn_hol_pcas()
 
     _LOG.info("plot and store PCA similarity contours")
-    similarity.plot_store_similarity_contours(
+    _similarity.plot_store_similarity_contours(
         pca_model=pca_model, trn_pca=trn_pca, hol_pca=hol_pca, syn_embeds=syn_embeds, workspace=workspace
     )

mostlyai/qa/report_from_statistics.py

@@ -18,7 +18,7 @@
 import pandas as pd
 import pytest
 
-from mostlyai.qa.filesystem import TemporaryWorkspace
+from mostlyai.qa._filesystem import TemporaryWorkspace
 
 
 @pytest.fixture()