GH-16453 - export no_progress_block from h2o module

h2o-py/h2o/__init__.py

@@ -26,7 +26,7 @@
 from h2o.h2o import (connect, init, api, connection, resume,
                      lazy_import, upload_file, import_file, import_sql_table, import_sql_select, import_hive_table,
                      parse_setup, parse_raw, assign, deep_copy, models, get_model, get_grid, get_frame,
-                     show_progress, no_progress, enable_expr_optimizations, is_expr_optimizations_enabled,
+                     show_progress, no_progress, no_progress_block, enable_expr_optimizations, is_expr_optimizations_enabled,
                      log_and_echo, remove, remove_all, rapids,
                      ls, frame, frames, create_frame, load_frame,
                      download_pojo, download_csv, download_all_logs, save_model, download_model, upload_model, load_model,

h2o-py/h2o/explanation/_explain.py

@@ -3,7 +3,6 @@
 import random
 import warnings
 from collections import OrderedDict, Counter, defaultdict
-from contextlib import contextmanager
 
 from io import StringIO
 
@@ -186,22 +185,6 @@ def _ipython_display_(self):
                 display(v)
 
 
-@contextmanager
-def no_progress_block():
-    """
-    A context manager that temporarily blocks showing the H2O's progress bar.
-    Used when a multiple models are evaluated.
-    """
-    progress = h2o.job.H2OJob.__PROGRESS_BAR__
-    if progress:
-        h2o.no_progress()
-    try:
-        yield
-    finally:
-        if progress:
-            h2o.show_progress()
-
-
 class NumpyFrame:
     """
     Simple class that very vaguely emulates Pandas DataFrame.
@@ -699,7 +682,7 @@ def shap_summary_plot(
         permutation = list(range(frame.nrow))
         random.shuffle(permutation)
 
-    with no_progress_block():
+    with h2o.no_progress_block():
         contributions = NumpyFrame(model.predict_contributions(frame, output_format="compact", background_frame=background_frame))
     frame = NumpyFrame(frame)
     contribution_names = contributions.columns
@@ -833,7 +816,7 @@ def shap_explain_row_plot(
         top_n_features = float("inf")
 
     row = frame[row_index, :]
-    with no_progress_block():
+    with h2o.no_progress_block():
         contributions = NumpyFrame(model.predict_contributions(row, output_format="compact", background_frame=background_frame))
     contribution_names = contributions.columns
     prediction = float(contributions.sum(axis=1))
@@ -842,7 +825,7 @@ def shap_explain_row_plot(
                            key=lambda pair: -abs(pair[1]))
 
     if plot_type == "barplot":
-        with no_progress_block():
+        with h2o.no_progress_block():
             prediction = model.predict(row)[0, "predict"]
         row = NumpyFrame(row)
 
@@ -1394,7 +1377,7 @@ def pd_ice_common(
 
     show_logodds = is_binomial and binary_response_scale == "logodds"
 
-    with no_progress_block():
+    with h2o.no_progress_block():
         plt.figure(figsize=figsize)
         if is_ice:
             res = _handle_ice(model, frame, colormap, plt, target, is_factor, column, show_logodds, centered,
@@ -1576,7 +1559,7 @@ def pd_multi_plot(
     models = [m if isinstance(m, h2o.model.ModelBase) else h2o.get_model(m) for m in models]
 
     colors = plt.get_cmap(colormap, len(models))(list(range(len(models))))
-    with no_progress_block():
+    with h2o.no_progress_block():
         plt.figure(figsize=figsize)
         is_factor = frame[column].isfactor()[0]
         if is_factor:
@@ -2332,7 +2315,7 @@ def model_correlation(
         models = list(_get_models_from_automl_or_leaderboard(models))
     is_classification = frame[models[0].actual_params["response_column"]].isfactor()[0]
     predictions = []
-    with no_progress_block():
+    with h2o.no_progress_block():
         for idx, model in enumerate(models):
             predictions.append(model.predict(frame)["predict"])
 
@@ -2411,7 +2394,7 @@ def residual_analysis_plot(
     plt = get_matplotlib_pyplot(False, raise_if_not_available=True)
     _, y = _get_xy(model)
 
-    with no_progress_block():
+    with h2o.no_progress_block():
         predicted = NumpyFrame(model.predict(frame)["predict"])
     actual = NumpyFrame(frame[y])
 
@@ -2964,7 +2947,7 @@ def _get_leaderboard(
     if row_index is not None:
         model_ids = [m[0] for m in
                      leaderboard["model_id"].as_data_frame(use_pandas=False, header=False)]
-        with no_progress_block():
+        with h2o.no_progress_block():
             preds = h2o.get_model(model_ids[0]).predict(frame[row_index, :])
             for model_id in model_ids[1:]:
                 preds = preds.rbind(h2o.get_model(model_id).predict(frame[row_index, :]))

h2o-py/h2o/h2o.py

@@ -10,6 +10,7 @@
 import tempfile
 import warnings
 import webbrowser
+from contextlib import contextmanager
 
 from .backend import H2OConnection
 from .backend import H2OConnectionConf
@@ -1216,6 +1217,31 @@ def no_progress():
     H2OJob.__PROGRESS_BAR__ = False
 
 
+@contextmanager
+def no_progress_block():
+    """
+    A context manager that temporarily blocks showing the H2O's progress bar.
+
+    :examples:
+
+    >>> with h2o.no_progress_block():
+    >>>     airlines= h2o.import_file("https://s3.amazonaws.com/h2o-public-test-data/smalldata/airlines/allyears2k_headers.zip")
+    >>>     x = ["DayofMonth", "Month"]
+    >>>     model = H2OGeneralizedLinearEstimator(family="binomial",
+    ...                                           alpha=0,
+    ...                                           Lambda=1e-5)
+    >>>     model.train(x=x, y="IsDepDelayed", training_frame=airlines) 
+    """
+    progress = H2OJob.__PROGRESS_BAR__
+    if progress:
+        no_progress()
+    try:
+        yield
+    finally:
+        if progress:
+            show_progress()
+
+
 def show_progress():
     """Enable the progress bar (it is enabled by default).
 

h2o-py/h2o/model/extensions/fairness.py

@@ -106,7 +106,6 @@ def fair_pd_plot(self, frame, column, protected_columns, figsize=(16, 9), autosc
         if not can_use_numpy():
             raise ImportError("numpy is required for fair_pd_plot.")
         import numpy as np
-        from h2o.explanation._explain import no_progress_block
         from h2o.plot import get_matplotlib_pyplot
         from h2o.utils.typechecks import assert_is_type, is_type
 
@@ -121,7 +120,7 @@ def fair_pd_plot(self, frame, column, protected_columns, figsize=(16, 9), autosc
         plt.figure(figsize=figsize)
         results = []
         maxes = []
-        with no_progress_block():
+        with h2o.no_progress_block():
             for pg in pgs:
                 pg = [p[0] for p in pg]
                 filtered_hdf = frame
@@ -349,7 +348,6 @@ def fair_shap_plot(self, frame, column, protected_columns, autoscale=True, figsi
         if not can_use_numpy():
             raise ImportError("numpy is required for fair_shap_plot.")
         import numpy as np
-        from h2o.explanation._explain import no_progress_block
         from h2o.explanation import H2OExplanation
         from h2o.explanation._explain import NumpyFrame
         from h2o.explanation._explain import _density
@@ -370,7 +368,7 @@ def fair_shap_plot(self, frame, column, protected_columns, autoscale=True, figsi
         results = defaultdict(list)
         maxes = []
         contr_columns = [column]
-        with no_progress_block():
+        with h2o.no_progress_block():
             for pg in pgs:
                 pg = [p[0] for p in pg]
                 filtered_hdf = frame

h2o-py/tests/testdir_misc/explain/pyunit_SHAP.py

@@ -21,7 +21,6 @@ def tqdm(x, *args, **kwargs):
         return x
 
 from h2o.estimators import *
-from h2o.explanation._explain import no_progress_block
 
 seed = 6
 K = 2
@@ -72,7 +71,7 @@ def test_local_accuracy(
         print("Testing local accuracy in output space...")
     else:
         print("Testing local accuracy...")
-    with no_progress_block():
+    with h2o.no_progress_block():
         cf = mod.predict_contributions(
             test, background_frame=train, output_format=output_format, output_space=output_space,
             output_per_reference=True
@@ -314,7 +313,7 @@ def test_contributions_against_naive(mod, y, train, test, link=False, eps=1e-6):
                 [train[brow, k] == "NA" for k, v in train.types.items() if v == "enum"]
             ):
                 continue
-            with no_progress_block():
+            with h2o.no_progress_block():
                 naive_contr = naiveBSHAP(mod, y, train, test, xrow, brow, link=link)
                 contr = mod.predict_contributions(
                     test[xrow, :],

h2o-py/tests/testdir_misc/explain/pyunit_SHAP_NOPASS.py

@@ -21,7 +21,6 @@ def tqdm(x, *args, **kwargs):
         return x
 
 from h2o.estimators import *
-from h2o.explanation._explain import no_progress_block
 
 seed = 6
 K = 5
@@ -57,7 +56,7 @@ def test_local_accuracy(
         print("Testing local accuracy in output space...")
     else:
         print("Testing local accuracy...")
-    with no_progress_block():
+    with h2o.no_progress_block():
         cf = mod.predict_contributions(
             test, background_frame=train, output_format=output_format, output_space=output_space,
             output_per_reference=True
@@ -299,7 +298,7 @@ def test_contributions_against_naive(mod, y, train, test, link=False, eps=1e-6):
                 [train[brow, k] == "NA" for k, v in train.types.items() if v == "enum"]
             ):
                 continue
-            with no_progress_block():
+            with h2o.no_progress_block():
                 naive_contr = naiveBSHAP(mod, y, train, test, xrow, brow, link=link)
                 contr = mod.predict_contributions(
                     test[xrow, :],