add meta function to clear caches (#57)

adbar · Jul 18, 2022 · 42644fc · 42644fc
1 parent 542b422
commit 42644fc
Show file tree

Hide file tree

Showing 2 changed files with 44 additions and 0 deletions.
diff --git a/htmldate/meta.py b/htmldate/meta.py
@@ -0,0 +1,34 @@
+"""
+Meta-functions to be applied module-wide.
+"""
+
+try:
+    from charset_normalizer.cd import encoding_languages
+    from charset_normalizer.md import is_suspiciously_successive_range
+    from charset_normalizer.utils import is_accentuated
+# prevent possible changes in function names
+except ImportError:
+    pass
+
+from .core import compare_reference
+from .extractors import try_date_expr
+from .settings import MAX_FILE_SIZE, MIN_FILE_SIZE
+from .validators import date_validator, filter_ymd_candidate
+
+
+def reset_caches() -> None:
+    """Reset all known LRU caches used to speed-up processing.
+    This may release some memory."""
+    # htmldate
+    compare_reference.cache_clear()
+    date_validator.cache_clear()
+    filter_ymd_candidate.cache_clear()
+    try_date_expr.cache_clear()
+    # charset_normalizer
+    try:
+        encoding_languages.cache_clear()
+        is_suspiciously_successive_range.cache_clear()
+        is_accentuated.cache_clear()
+    # prevent possible changes in function names
+    except NameError:
+        pass
diff --git a/tests/unit_tests.py b/tests/unit_tests.py
@@ -53,6 +53,7 @@
     regex_parse,
     try_date_expr,
 )
+from htmldate.meta import reset_caches
 from htmldate.settings import MIN_DATE, LATEST_POSSIBLE
 from htmldate.utils import (
     decode_response,
@@ -216,6 +217,11 @@ def test_sanity():
         mytree, ".//p", OUTPUTFORMAT, True, MIN_DATE, LATEST_POSSIBLE
     )
     assert result is None
+    mytree = html.fromstring("<html><body><p>1999/03/05</p></body></html>")
+    result = examine_date_elements(
+        mytree, ".//p", OUTPUTFORMAT, True, MIN_DATE, LATEST_POSSIBLE
+    )
+    assert result is not None
     # wrong field values in output format
     assert output_format_validator("%Y-%m-%d") is True
     assert output_format_validator("%M-%Y") is True
@@ -229,6 +235,10 @@ def test_sanity():
         )
     )
     assert len(discarded) == 1
+    # reset caches: examine_date_elements used above
+    old_values = try_date_expr.cache_info()
+    reset_caches()
+    assert try_date_expr.cache_info() != old_values
 
 
 def test_no_date():