Merge tag '3.36.0' into stable

.github/workflows/test.yml

@@ -19,7 +19,7 @@ jobs:
       - name: Setup Python
         uses: actions/setup-python@v4
         with:
-          python-version: 3.8
+          python-version: 3.9
 
       - name: Install Tox
         run: |
@@ -36,21 +36,21 @@ jobs:
     strategy:
       fail-fast: False
       matrix:
-        os: [ubuntu-20.04]
-        python-version: [3.8, 3.9, '3.10']
+        os: [ubuntu-22.04]
+        python-version: [3.9, '3.10', '3.11']
         tox_env: [orange-released]
         name: [Released]
         include:
-          - os: ubuntu-20.04
-            python-version: '3.10'
+          - os: ubuntu-22.04
+            python-version: '3.11'
             tox_env: orange-latest
             name: Latest
           - os: ubuntu-20.04
-            python-version: 3.8
+            python-version: 3.9
             tox_env: orange-oldest
             name: Oldest dependencies
-          - os: ubuntu-20.04
-            python-version: 3.9
+          - os: ubuntu-22.04
+            python-version: '3.10'
             tox_env: pyqt6
             name: PyQt6
 
@@ -85,6 +85,10 @@ jobs:
           sudo apt-get update
           sudo apt-get install -y libxkbcommon-x11-0 libxcb-icccm4 libxcb-image0 libxcb-keysyms1 libxcb-randr0 libxcb-render-util0 libxcb-xinerama0 libxcb-xfixes0 libegl1-mesa libxcb-shape0 libxcb-cursor0
 
+      - name: Install glibc-tools dependency on Ubuntu-22.04
+        if: matrix.os == 'ubuntu-22.04'
+        run: sudo apt-get install -y glibc-tools  # required for catchsegv
+
       - name: Install Tox
         run: |
           python -m pip install --upgrade pip
@@ -116,16 +120,16 @@ jobs:
       fail-fast: false
       matrix:
         os: [macos-latest, windows-latest]
-        python-version: [3.8, 3.9, '3.10']
+        python-version: [3.9, '3.10', '3.11']
         tox_env: [orange-released]
         name: [Released]
         include:
           - os: windows-latest
-            python-version: '3.10'
+            python-version: '3.11'
             tox_env: orange-latest
             name: Latest
           - os: macos-latest
-            python-version: '3.10'
+            python-version: '3.11'
             tox_env: orange-latest
             name: Latest
           - os: windows-latest

CHANGELOG.md

@@ -4,6 +4,45 @@ Change Log
 [next] - TBA
 ------------
 
+[3.36.0] - 2023-09-08
+--------------------
+##### Enhancements
+* Update widget.json
+* feature-constructor2-stamped.png - resize and index
+* Formula documentation: Minor changes
+* index.rst - fix reference to formula
+* move accidentally displaced formula.md
+* SOM: output columns with coordinates and errors ([#6542](../../pull/6542))
+* Lazy signals for Hierarchical Clustering ([#6348](../../pull/6348))
+* PCA widget runs a separate thread ([#6528](../../pull/6528))
+* Edit Domain: Change types of multiple variables ([#6426](../../pull/6426))
+* owsavebase: Display a warning when auto save is disabled ([#6454](../../pull/6454))
+* Select Columns: Accept partially correct drops ([#6390](../../pull/6390))
+
+##### Bugfixes
+* Update widget.json
+* feature-constructor2-stamped.png - resize and index
+* Formula documentation: Minor changes
+* index.rst - fix reference to formula
+* move accidentally displaced formula.md
+* Predictions: Fix column size hinting ([#6563](../../pull/6563))
+* Naive Bayes: fix predictions with unknown values ([#6564](../../pull/6564))
+* Scatter Plot - Handle input features that are hidden in data ([#6531](../../pull/6531))
+* Adapt tests to pandas 2.1 and fix deprecations ([#6560](../../pull/6560))
+* Adapt to newly released pandas 2.1 ([#6559](../../pull/6559))
+* CSV Import - Change datetime format parsing ([#6539](../../pull/6539))
+* owdiscretize: Fix formatting display string when user role is undefined ([#6498](../../pull/6498))
+* Neural Network: Default learner name propagation ([#6526](../../pull/6526))
+* PCA - Output instance of table subclass when instance of table subclass on input ([#6536](../../pull/6536))
+* owpredictions: Fix a type error in report when using NoopDelegate ([#6537](../../pull/6537))
+* Scatter Plot - Fix Vizrank for hidden attributes ([#6530](../../pull/6530))
+* config: Replace and fix use of pkg_resources.parse_version ([#6523](../../pull/6523))
+* Fix classification trees for data with repeated feature values ([#6488](../../pull/6488))
+* Group By - Fix error with hidden attributes ([#6473](../../pull/6473))
+* Sql - Fix comparison with values trailing spaces ([#6470](../../pull/6470))
+* ListView - empty variable when all values unselected ([#6441](../../pull/6441))
+
+
 [3.35.0] - 2023-05-05
 ---------------------                                                                                       
 ##### Enhancements
@@ -1784,7 +1823,8 @@ Change Log
 
 
 [next]: https://github.com/biolab/orange3/compare/3.35.0..HEAD
-[3.34.1]: https://github.com/biolab/orange3/compare/3.34.1...3.35.0
+[3.36.0]: https://github.com/biolab/orange3/compare/3.35.0...3.36.0
+[3.35.0]: https://github.com/biolab/orange3/compare/3.34.1...3.35.0
 [3.34.1]: https://github.com/biolab/orange3/compare/3.34.0...3.34.1
 [3.34.0]: https://github.com/biolab/orange3/compare/3.33.0...3.34.0
 [3.33.0]: https://github.com/biolab/orange3/compare/3.32.0...3.33.0

Orange/canvas/config.py

@@ -5,14 +5,12 @@
 import random
 import uuid
 import warnings
-
 import os
 import sys
-import itertools
-from distutils.version import LooseVersion
 
 from typing import Dict, Any, Optional, Iterable, List
 
+import packaging.version
 import pkg_resources
 import requests
 
@@ -112,8 +110,8 @@ def splash_screen():
 
         version = Config.ApplicationVersion
         if version:
-            version_parsed = LooseVersion(version)
-            version_comp = version_parsed.version
+            version_parsed = packaging.version.Version(version)
+            version_comp = version_parsed.release
             version = ".".join(map(str, version_comp[:2]))
         size = 13
         font = QFont("Helvetica")

Orange/canvas/run.py

@@ -14,6 +14,14 @@
 # Imported to make webwidget addons work
 from Orange.widgets.utils.webview import WebviewWidget  # pylint: disable=unused-import
 
+LOG_LEVELS = [
+    logging.CRITICAL + 10,
+    logging.CRITICAL,
+    logging.ERROR,
+    logging.WARN,
+    logging.INFO,
+    logging.DEBUG
+]
 
 def main(argv=None):
     app = QApplication(list(argv) if argv else [])
@@ -28,13 +36,13 @@ def main(argv=None):
         )
     )
     parser.add_argument("--log-level", "-l", metavar="LEVEL", type=int,
-                        default=logging.CRITICAL, dest="log_level")
+                        default=3, dest="log_level")
     parser.add_argument("--config", default="Orange.canvas.config.Config",
                         type=str)
     parser.add_argument("file")
     args = parser.parse_args(argv[1:])
 
-    log_level = args.log_level
+    log_level = LOG_LEVELS[args.log_level]
     filename = args.file
     logging.basicConfig(level=log_level)
 
@@ -71,6 +79,9 @@ def on_finished():
                 if msg.contents and msg.severity == msg.Error:
                     print(msg.contents, msg.message_id, file=sys.stderr)
                     severity = msg.Error
+                elif msg.contents and msg.severity == msg.Warning and \
+                        log_level <= logging.WARNING:
+                    print(msg.contents, file=sys.stderr)
         if severity == UserMessage.Error:
             app.exit(1)
         else:

Orange/classification/_tree_scorers.pyx

@@ -79,7 +79,7 @@ def find_threshold_entropy(const double[:] x, const double[:] y,
             curr_y = <int>y[idx[i]]
             distr[curr_y] -= 1
             distr[n_classes + curr_y] += 1
-            if curr_y != y[idx[i + 1]] and x[idx[i]] != x[idx[i + 1]]:
+            if x[idx[i]] != x[idx[i + 1]]:
                 entro = (i + 1) * log(i + 1) + (N - i - 1) * log(N - i - 1)
                 for j in range(2 * n_classes):
                     if distr[j]:

Orange/classification/naive_bayes.py

@@ -98,7 +98,7 @@ def _dense_probs(self, data, probs):
         zeros = np.zeros((1, probs.shape[1]))
         for col, attr_prob in zip(data.T, self.log_cont_prob):
             col = col.copy()
-            col[np.isnan(col)] = attr_prob.shape[1] - 1
+            col[np.isnan(col)] = attr_prob.shape[1]
             col = col.astype(int)
             probs0 = np.vstack((attr_prob.T, zeros))
             probs += probs0[col]
@@ -113,6 +113,7 @@ def _sparse_probs(self, data, probs):
             p0 = p.T[0].copy()
             probs[:] += p0
             log_prob[i, :p.shape[1]] = p.T - p0
+            log_prob[i, n_vals-1] = -p0
 
         dat = data.data.copy()
         dat[np.isnan(dat)] = n_vals - 1

Orange/data/filter.py

@@ -419,18 +419,18 @@ class FilterString(ValueFilter):
 
         The operator; should be `FilterString.Equal`, `NotEqual`, `Less`,
         `LessEqual`, `Greater`, `GreaterEqual`, `Between`, `Outside`,
-        `Contains`, `StartsWith`, `EndsWith` or `IsDefined`.
+        `Contains`, `NotContain`, `StartsWith`, `NotStartsWith`, `EndsWith`, `NotEndsWith`, `IsDefined` or `NotIsDefined`.
 
     .. attribute:: case_sensitive
 
         Tells whether the comparisons are case sensitive
     """
     Type = Enum('FilterString',
                 'Equal, NotEqual, Less, LessEqual, Greater,'
-                'GreaterEqual, Between, Outside, Contains,'
-                'StartsWith, EndsWith, IsDefined')
+                'GreaterEqual, Between, Outside, Contains, NotContain,'
+                'StartsWith, NotStartsWith, EndsWith, NotEndsWith, IsDefined, NotIsDefined')
     (Equal, NotEqual, Less, LessEqual, Greater, GreaterEqual,
-     Between, Outside, Contains, StartsWith, EndsWith, IsDefined) = Type
+     Between, Outside, Contains, NotContain, StartsWith, NotStartsWith, EndsWith, NotEndsWith, IsDefined, NotIsDefined) = Type
 
     def __init__(self, position, oper, ref=None, max=None,
                  case_sensitive=True, **a):
@@ -460,6 +460,8 @@ def __call__(self, inst):
         value = inst[inst.domain.index(self.column)]
         if self.oper == self.IsDefined:
             return not np.isnan(value)
+        if self.oper == self.NotIsDefined:
+            return np.isnan(value)
         if self.case_sensitive:
             value = str(value)
             refval = str(self.ref)
@@ -480,10 +482,16 @@ def __call__(self, inst):
             return value >= refval
         if self.oper == self.Contains:
             return refval in value
+        if self.oper == self.NotContain:
+            return refval not in value
         if self.oper == self.StartsWith:
             return value.startswith(refval)
+        if self.oper == self.NotStartsWith:
+            return not value.startswith(refval)
         if self.oper == self.EndsWith:
             return value.endswith(refval)
+        if self.oper == self.NotEndsWith:
+            return not value.endswith(refval)
         high = self.max if self.case_sensitive else self.max.lower()
         if self.oper == self.Between:
             return refval <= value <= high

Orange/data/io_base.py

@@ -654,7 +654,7 @@ def formatter(cls, var):
         if var.is_time:
             return var.repr_val
         elif var.is_continuous:
-            return lambda value: "" if isnan(value) else value
+            return lambda value: "" if isnan(value) else var.repr_val(value)
         elif var.is_discrete:
             return lambda value: "" if isnan(value) else var.values[int(value)]
         elif var.is_string:

Orange/data/pandas_compat.py

@@ -6,10 +6,11 @@
 from scipy.sparse import csr_matrix
 import pandas as pd
 from pandas.core.arrays import SparseArray
-from pandas.core.arrays.sparse.dtype import SparseDtype
 from pandas.api.types import (
-    is_categorical_dtype, is_object_dtype,
-    is_datetime64_any_dtype, is_numeric_dtype, is_integer_dtype
+    is_object_dtype,
+    is_datetime64_any_dtype,
+    is_numeric_dtype,
+    is_integer_dtype,
 )
 
 from Orange.data import (
@@ -166,9 +167,11 @@ def _reset_index(df: pd.DataFrame) -> pd.DataFrame:
 
 
 def _is_discrete(s, force_nominal):
-    return (is_categorical_dtype(s) or
-            is_object_dtype(s) and (force_nominal or
-                                    s.nunique() < s.size ** .666))
+    return (
+        isinstance(s.dtype, pd.CategoricalDtype)
+        or is_object_dtype(s)
+        and (force_nominal or s.nunique() < s.size**0.666)
+    )
 
 
 def _is_datetime(s):
@@ -297,7 +300,7 @@ def vars_from_df(df, role=None, force_nominal=False):
         elif not any(a_expr):
             # if all c in columns table will share memory with dataframe
             a_df = df if all(c in a_cols for c in df.columns) else df[a_cols]
-            if all(isinstance(a, SparseDtype) for a in a_df.dtypes):
+            if all(isinstance(a, pd.SparseDtype) for a in a_df.dtypes):
                 arr = csr_matrix(a_df.sparse.to_coo())
             else:
                 arr = np.asarray(a_df)

Orange/data/sql/backend/mssql.py

@@ -158,10 +158,12 @@ def distinct_values_query(self, field_name: str, table_name: str) -> str:
         return self.create_sql_query(
             table_name,
             [field],
-            # workaround for collations that are not case sensitive and
-            # UTF characters sensitive - in the domain we still want to
-            # have all values (collation independent)
-            group_by=[f"{field}, Cast({field} as binary)"],
+            # Cast - workaround for collations that are not case-sensitive and
+            # UTF characters sensitive
+            # DATALENGTH - workaround for string comparison that ignore trailing
+            # spaces, two strings that differ only in space in the end would
+            # group together if DATALENGTH wouldn't be used
+            group_by=[f"{field}, Cast({field} as binary), DATALENGTH({field})"],
             order_by=[field],
             limit=21,
         )

Orange/data/sql/table.py

@@ -1,6 +1,7 @@
 """
 Support for example tables wrapping data stored on a PostgreSQL server.
 """
+import contextlib
 import functools
 import logging
 import threading
@@ -669,3 +670,26 @@ def get_nan_frequency_attribute(self):
 
     def get_nan_frequency_class(self):
         return self.__get_nan_frequency(self.domain.class_vars)
+
+    def __getstate__(self):
+        # avoids locking magic in Table.__getstate__
+        return self.__dict__
+
+    def __setstate__(self, state):
+        # avoid locking magic in Table.__setstate__
+        self.__dict__.update(state)
+
+        # if X is defined then it was already downloaded
+        # thus ids exist to, rewrite them
+        if self._X is not None:
+            self._init_ids(self)
+
+    # pylint: disable=unused-argument
+    def _update_locks(self, *args, **kwargs):
+        # avoid locking inherited from Table
+        return
+
+    # pylint: disable=unused-argument
+    def unlocked(self, *parts):
+        # avoid locking inherited from Table
+        return contextlib.nullcontext()