From fd8b9a798be7b0fb7620c35d9c8f54e3a0473133 Mon Sep 17 00:00:00 2001
From: Osma Suominen <osma.suominen@helsinki.fi>
Date: Thu, 21 Nov 2024 17:14:44 +0200
Subject: [PATCH 1/2] smarter initialization of optional analyzers; use
 importlib to test optional feature availability

---
 annif/analyzer/__init__.py | 24 +++++++-----------------
 annif/analyzer/analyzer.py |  5 +++++
 annif/analyzer/spacy.py    |  7 +++++++
 annif/analyzer/voikko.py   | 10 ++++++++--
 4 files changed, 27 insertions(+), 19 deletions(-)

diff --git a/annif/analyzer/__init__.py b/annif/analyzer/__init__.py
index 27a2cd792..81f525116 100644
--- a/annif/analyzer/__init__.py
+++ b/annif/analyzer/__init__.py
@@ -8,7 +8,7 @@
 import annif
 from annif.util import parse_args
 
-from . import simple, simplemma, snowball
+from . import simple, simplemma, snowball, spacy, voikko
 
 if TYPE_CHECKING:
     from annif.analyzer.analyzer import Analyzer
@@ -17,7 +17,10 @@
 
 
 def register_analyzer(analyzer):
-    _analyzers[analyzer.name] = analyzer
+    if analyzer.is_available():
+        _analyzers[analyzer.name] = analyzer
+    else:
+        annif.logger.debug(f"{analyzer.name} analyzer not available, not enabling it")
 
 
 def get_analyzer(analyzerspec: str) -> Analyzer:
@@ -37,18 +40,5 @@ def get_analyzer(analyzerspec: str) -> Analyzer:
 register_analyzer(simple.SimpleAnalyzer)
 register_analyzer(snowball.SnowballAnalyzer)
 register_analyzer(simplemma.SimplemmaAnalyzer)
-
-# Optional analyzers
-try:
-    from . import voikko
-
-    register_analyzer(voikko.VoikkoAnalyzer)
-except ImportError:
-    annif.logger.debug("voikko not available, not enabling voikko analyzer")
-
-try:
-    from . import spacy
-
-    register_analyzer(spacy.SpacyAnalyzer)
-except ImportError:
-    annif.logger.debug("spaCy not available, not enabling spacy analyzer")
+register_analyzer(voikko.VoikkoAnalyzer)
+register_analyzer(spacy.SpacyAnalyzer)
diff --git a/annif/analyzer/analyzer.py b/annif/analyzer/analyzer.py
index 129b882ab..612234d95 100644
--- a/annif/analyzer/analyzer.py
+++ b/annif/analyzer/analyzer.py
@@ -22,6 +22,11 @@ class Analyzer(metaclass=abc.ABCMeta):
     name = None
     token_min_length = 3  # default value, can be overridden in instances
 
+    @staticmethod
+    def is_available() -> bool:
+        """Return True if the analyzer is available for use, False if not."""
+        return True  # can be overridden in implementations if necessary
+
     def __init__(self, **kwargs) -> None:
         if _KEY_TOKEN_MIN_LENGTH in kwargs:
             self.token_min_length = int(kwargs[_KEY_TOKEN_MIN_LENGTH])
diff --git a/annif/analyzer/spacy.py b/annif/analyzer/spacy.py
index 184f03ffc..407f0d9c7 100644
--- a/annif/analyzer/spacy.py
+++ b/annif/analyzer/spacy.py
@@ -2,6 +2,8 @@
 
 from __future__ import annotations
 
+import importlib
+
 import annif.util
 from annif.exception import OperationFailedException
 
@@ -13,6 +15,11 @@
 class SpacyAnalyzer(analyzer.Analyzer):
     name = "spacy"
 
+    @staticmethod
+    def is_available() -> bool:
+        # return True iff spaCy is installed
+        return importlib.util.find_spec("spacy") is not None
+
     def __init__(self, param: str, **kwargs) -> None:
         import spacy
 
diff --git a/annif/analyzer/voikko.py b/annif/analyzer/voikko.py
index b3e7d5007..3722d6b4b 100644
--- a/annif/analyzer/voikko.py
+++ b/annif/analyzer/voikko.py
@@ -3,8 +3,7 @@
 from __future__ import annotations
 
 import functools
-
-import voikko.libvoikko
+import importlib
 
 from . import analyzer
 
@@ -12,6 +11,11 @@
 class VoikkoAnalyzer(analyzer.Analyzer):
     name = "voikko"
 
+    @staticmethod
+    def is_available() -> bool:
+        # return True iff Voikko is installed
+        return importlib.util.find_spec("voikko") is not None
+
     def __init__(self, param: str, **kwargs) -> None:
         self.param = param
         self.voikko = None
@@ -26,6 +30,8 @@ def __getstate__(self) -> dict[str, str | None]:
 
     @functools.lru_cache(maxsize=500000)
     def _normalize_word(self, word: str) -> str:
+        import voikko.libvoikko
+
         if self.voikko is None:
             self.voikko = voikko.libvoikko.Voikko(self.param)
         result = self.voikko.analyze(word)

From 419f8df671f2640332d231cd76a0399561e4a448 Mon Sep 17 00:00:00 2001
From: Osma Suominen <osma.suominen@helsinki.fi>
Date: Thu, 21 Nov 2024 17:28:33 +0200
Subject: [PATCH 2/2] fix test skips for voikko and spacy analyzers

---
 tests/test_analyzer_spacy.py  | 5 ++++-
 tests/test_analyzer_voikko.py | 5 ++++-
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/tests/test_analyzer_spacy.py b/tests/test_analyzer_spacy.py
index bf14a9ae0..e3012af05 100644
--- a/tests/test_analyzer_spacy.py
+++ b/tests/test_analyzer_spacy.py
@@ -3,9 +3,12 @@
 import pytest
 
 import annif.analyzer
+import annif.analyzer.spacy
 from annif.exception import OperationFailedException
 
-spacy = pytest.importorskip("spacy")
+pytestmark = pytest.mark.skipif(
+    not annif.analyzer.spacy.SpacyAnalyzer.is_available(), reason="spaCy is required"
+)
 
 
 def test_spacy_model_not_found():
diff --git a/tests/test_analyzer_voikko.py b/tests/test_analyzer_voikko.py
index 6cba156f7..8598482b1 100644
--- a/tests/test_analyzer_voikko.py
+++ b/tests/test_analyzer_voikko.py
@@ -3,8 +3,11 @@
 import pytest
 
 import annif.analyzer
+import annif.analyzer.voikko
 
-voikko = pytest.importorskip("annif.analyzer.voikko")
+pytestmark = pytest.mark.skipif(
+    not annif.analyzer.voikko.VoikkoAnalyzer.is_available(), reason="voikko is required"
+)
 
 
 def test_voikko_getstate():