feat: add smartcase and globless path searches (#743)

* fix: return path_strings in session

* feat: add smartcase and globless path search

Known issues: failing tests, sluggish autocomplete

* fix: all operational searches

* fix: limit path autocomplete to 100 items

* tests: add test cases

Author: CyanVoxel

tagstudio/src/core/library/alchemy/library.py

@@ -716,13 +716,15 @@ def has_path_entry(self, path: Path) -> bool:
         with Session(self.engine) as session:
             return session.query(exists().where(Entry.path == path)).scalar()
 
-    def get_paths(self, glob: str | None = None) -> list[str]:
+    def get_paths(self, glob: str | None = None, limit: int = -1) -> list[str]:
         path_strings: list[str] = []
         with Session(self.engine) as session:
-            paths = session.scalars(select(Entry.path)).unique()
+            if limit > 0:
+                paths = session.scalars(select(Entry.path).limit(limit)).unique()
+            else:
+                paths = session.scalars(select(Entry.path)).unique()
             path_strings = list(map(lambda x: x.as_posix(), paths))
-
-        return path_strings
+            return path_strings
 
     def search_library(
         self,

tagstudio/src/core/library/alchemy/visitors.py

@@ -2,19 +2,21 @@
 # Licensed under the GPL-3.0 License.
 # Created for TagStudio: https://github.com/CyanVoxel/TagStudio
 
+import re
 from typing import TYPE_CHECKING
 
 import structlog
 from sqlalchemy import ColumnElement, and_, distinct, func, or_, select, text
 from sqlalchemy.orm import Session
+from sqlalchemy.sql.operators import ilike_op
 from src.core.media_types import FILETYPE_EQUIVALENTS, MediaCategories
 from src.core.query_lang import BaseVisitor
 from src.core.query_lang.ast import ANDList, Constraint, ConstraintType, Not, ORList, Property
 
 from .joins import TagEntry
 from .models import Entry, Tag, TagAlias
 
-# workaround to have autocompletion in the Editor
+# Only import for type checking/autocompletion, will not be imported at runtime.
 if TYPE_CHECKING:
     from .library import Library
 else:
@@ -97,7 +99,29 @@ def visit_constraint(self, node: Constraint) -> ColumnElement[bool]:
         elif node.type == ConstraintType.TagID:
             return self.__entry_matches_tag_ids([int(node.value)])
         elif node.type == ConstraintType.Path:
-            return Entry.path.op("GLOB")(node.value)
+            ilike = False
+            glob = False
+
+            # Smartcase check
+            if node.value == node.value.lower():
+                ilike = True
+            if node.value.startswith("*") or node.value.endswith("*"):
+                glob = True
+
+            if ilike and glob:
+                logger.info("ConstraintType.Path", ilike=True, glob=True)
+                return func.lower(Entry.path).op("GLOB")(f"{node.value.lower()}")
+            elif ilike:
+                logger.info("ConstraintType.Path", ilike=True, glob=False)
+                return ilike_op(Entry.path, f"%{node.value}%")
+            elif glob:
+                logger.info("ConstraintType.Path", ilike=False, glob=True)
+                return Entry.path.op("GLOB")(node.value)
+            else:
+                logger.info(
+                    "ConstraintType.Path", ilike=False, glob=False, re=re.escape(node.value)
+                )
+                return Entry.path.regexp_match(re.escape(node.value))
         elif node.type == ConstraintType.MediaType:
             extensions: set[str] = set[str]()
             for media_cat in MediaCategories.ALL_CATEGORIES:

tagstudio/src/qt/ts_qt.py

@@ -1505,7 +1505,9 @@ def update_completions_list(self, text: str) -> None:
         elif query_type == "tag_id":
             completion_list = list(map(lambda x: prefix + "tag_id:" + str(x.id), self.lib.tags))
         elif query_type == "path":
-            completion_list = list(map(lambda x: prefix + "path:" + x, self.lib.get_paths()))
+            completion_list = list(
+                map(lambda x: prefix + "path:" + x, self.lib.get_paths(limit=100))
+            )
         elif query_type == "mediatype":
             single_word_completions = map(
                 lambda x: prefix + "mediatype:" + x.name,

tagstudio/tests/test_library.py

@@ -414,6 +414,24 @@ class TestPrefs(DefaultEnum):
         assert TestPrefs.BAR.value
 
 
+def test_path_search_ilike(library: Library):
+    results = library.search_library(FilterState.from_path("bar.md"))
+    assert results.total_count == 1
+    assert len(results.items) == 1
+
+
+def test_path_search_like(library: Library):
+    results = library.search_library(FilterState.from_path("BAR.MD"))
+    assert results.total_count == 0
+    assert len(results.items) == 0
+
+
+def test_path_search_default_with_sep(library: Library):
+    results = library.search_library(FilterState.from_path("one/two"))
+    assert results.total_count == 1
+    assert len(results.items) == 1
+
+
 def test_path_search_glob_after(library: Library):
     results = library.search_library(FilterState.from_path("foo*"))
     assert results.total_count == 1
@@ -432,6 +450,50 @@ def test_path_search_glob_both_sides(library: Library):
     assert len(results.items) == 1
 
 
+def test_path_search_ilike_glob_equality(library: Library):
+    results_ilike = library.search_library(FilterState.from_path("one/two"))
+    results_glob = library.search_library(FilterState.from_path("*one/two*"))
+    assert [e.id for e in results_ilike.items] == [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+    results_ilike = library.search_library(FilterState.from_path("bar.md"))
+    results_glob = library.search_library(FilterState.from_path("*bar.md*"))
+    assert [e.id for e in results_ilike.items] == [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+    results_ilike = library.search_library(FilterState.from_path("bar"))
+    results_glob = library.search_library(FilterState.from_path("*bar*"))
+    assert [e.id for e in results_ilike.items] == [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+    results_ilike = library.search_library(FilterState.from_path("bar.md"))
+    results_glob = library.search_library(FilterState.from_path("*bar.md*"))
+    assert [e.id for e in results_ilike.items] == [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+
+def test_path_search_like_glob_equality(library: Library):
+    results_ilike = library.search_library(FilterState.from_path("ONE/two"))
+    results_glob = library.search_library(FilterState.from_path("*ONE/two*"))
+    assert [e.id for e in results_ilike.items] == [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+    results_ilike = library.search_library(FilterState.from_path("BAR.MD"))
+    results_glob = library.search_library(FilterState.from_path("*BAR.MD*"))
+    assert [e.id for e in results_ilike.items] == [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+    results_ilike = library.search_library(FilterState.from_path("BAR.MD"))
+    results_glob = library.search_library(FilterState.from_path("*bar.md*"))
+    assert [e.id for e in results_ilike.items] != [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+    results_ilike = library.search_library(FilterState.from_path("bar.md"))
+    results_glob = library.search_library(FilterState.from_path("*BAR.MD*"))
+    assert [e.id for e in results_ilike.items] != [e.id for e in results_glob.items]
+    results_ilike, results_glob = None, None
+
+
 @pytest.mark.parametrize(["filetype", "num_of_filetype"], [("md", 1), ("txt", 1), ("png", 0)])
 def test_filetype_search(library, filetype, num_of_filetype):
     results = library.search_library(FilterState.from_filetype(filetype))