feat(search): add unicode character removal for fuzzy matching#4360
feat(search): add unicode character removal for fuzzy matching#43604yinn wants to merge 30 commits intoFlow-Launcher:devfrom
Conversation
Introduced a string preprocessing step in FuzzySearch that removes unicode characters. This improves the search
experience by allowing users to find results regardless of accents or special formatting.
|
🥷 Code experts: Jack251970 Jack251970 has most 👩💻 activity in the files. See details
Activity based on git-commit:
Knowledge based on git-blame: ✨ Comment |
|
Be a legend 🏆 by adding a before and after screenshot of the changes you made, especially if they are around UI/UX. |
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughAdds a new SensitiveAccents setting and UI, implements an accent-normalization map and Normalize(string), and integrates accent-aware normalization into StringMatcher.FuzzyMatch to make fuzzy search optionally diacritic-sensitive. Changes
Sequence Diagram(s)mermaid User->>SettingsUI: Toggle SensitiveAccents Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested reviewers
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
Flow.Launcher.Infrastructure/StringMatcher.cs (1)
66-73: Consider adding an option to toggle diacritics-insensitive matching.The linked issue
#4149specifically requests an option to enable/disable diacritics-insensitive matching, similar to the Everything app. The current implementation always removes accents with no way to opt out.If this is an intentional scope reduction, consider documenting it. Otherwise, you could add a setting flag:
if (_settings.IgnoreDiacritics) { query = RemoveAccents(query); stringToCompare = RemoveAccents(stringToCompare); }🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Flow.Launcher.Infrastructure/StringMatcher.cs` around lines 66 - 73, FuzzyMatch currently always calls RemoveAccents on query and stringToCompare which forces diacritics-insensitive matching; add a configurable toggle (e.g. a boolean setting like _settings.IgnoreDiacritics) and only call RemoveAccents when that flag is true, updating FuzzyMatch's logic and any settings class to expose the option so consumers can enable/disable diacritics-insensitive matching; reference the RemoveAccents and FuzzyMatch methods and the MatchOption usage when making the change.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@Flow.Launcher.Infrastructure/StringMatcher.cs`:
- Around line 249-250: In StringMatcher.cs (inside the method using
Char.GetUnicodeCategory), rename the local variable unicodedCategory to
unicodeCategory to fix the typo and consistency, and also replace the type alias
use from Char.GetUnicodeCategory(c) to char.GetUnicodeCategory(c) for C# keyword
consistency; update any subsequent references to the variable name
(unicodeCategory) accordingly.
- Around line 72-73: The matching indices are computed on the post-RemoveAccents
normalized string but applied to the original input, causing mismatches when
combining marks change string length; update the flow in the method that calls
RemoveAccents to 1) keep the original stringToCompare (save
originalStringToCompare before normalization), 2) produce an index mapping from
normalized-to-original positions when RemoveAccents transforms the string
(similar to the existing TranslationMapping for alphabet translation), and 3)
before returning, remap all indices in indexList and any positions inside
acronymMatchData using that accent-removal mapping (in addition to the existing
TranslationMapping) so returned indices align with the original string. Ensure
the mapping logic is used wherever indices are translated back (same spot where
TranslationMapping is applied).
---
Nitpick comments:
In `@Flow.Launcher.Infrastructure/StringMatcher.cs`:
- Around line 66-73: FuzzyMatch currently always calls RemoveAccents on query
and stringToCompare which forces diacritics-insensitive matching; add a
configurable toggle (e.g. a boolean setting like _settings.IgnoreDiacritics) and
only call RemoveAccents when that flag is true, updating FuzzyMatch's logic and
any settings class to expose the option so consumers can enable/disable
diacritics-insensitive matching; reference the RemoveAccents and FuzzyMatch
methods and the MatchOption usage when making the change.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 0172967a-b14a-495a-9c63-e4cd5ee587ab
📒 Files selected for processing (1)
Flow.Launcher.Infrastructure/StringMatcher.cs
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
2 issues found across 1 file
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="Flow.Launcher.Infrastructure/StringMatcher.cs">
<violation number="1" location="Flow.Launcher.Infrastructure/StringMatcher.cs:72">
P1: `query` is not revalidated after accent stripping, so mark-only Unicode input can become empty and crash at `querySubstrings[0]`.</violation>
<violation number="2" location="Flow.Launcher.Infrastructure/StringMatcher.cs:73">
P2: Match indices are computed after accent-stripping but never mapped back to original string positions, causing incorrect highlight offsets for decomposed Unicode text.</violation>
</file>
Since this is your first cubic review, here's how it works:
- cubic automatically reviews your code and comments on bugs and improvements
- Teach cubic by replying to its comments. cubic learns from your replies and gets better over time
- Add one-off context when rerunning by tagging
@cubic-dev-aiwith guidance or docs links (includingllms.txt) - Ask questions if you need clarification on any suggestion
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
Co-authored-by: cubic-dev-ai[bot] <191113872+cubic-dev-ai[bot]@users.noreply.github.com>
Jack251970
added
the
review in progress
Jack251970
left a comment
Jack251970
left a comment
There was a problem hiding this comment.
Thanks for your contribution! But from AI's review, this removal may change the length of the string which can cause some issues like incorrect highlights.
|
Please don't involve too much code format changes so we can trace changes in the future. |
jjw24
left a comment
jjw24
left a comment
There was a problem hiding this comment.
Thank you for your PR.
StringMatcher is performance-critical, so we must minimize additional overhead. Please address the following:
-
Consolidate Loops: The current implementation performs an initial loop to identify and remove Unicode characters, followed by a second loop for the main matching logic. Please merge it into the main loop to avoid redundant iteration.
-
Conditional Unicode Processing: We should make Unicode removal optional—for instance, by adding a toggle for non-English languages. This prevents unnecessary processing when the feature isn't required.
Yes, I can see the problem. I'll think about a way to implement it with better optimizations. |
… option Implemented a string normalization method to handle accented characters, improving search consistency and preventing query mismatches. Added an accent mapping dictionary for common diacritics Implemented normalization using Span<char> and stackalloc to reduce heap allocations and improve performance Introduced a user-controlled toggle (SensitiveAccents) to enable or disable normalization dynamically Prepared the system for cache-aware queries based on normalization settings
…github.com/egotting/Flow.Launcher into feature/4149-diacritics-insensitive-search
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
Flow.Launcher.Infrastructure/StringMatcher.cs (1)
27-31:⚠️ Potential issue | 🔴 CriticalNullReferenceException:
_settingsis null when using this constructor.The workaround constructor does not initialize
_settings, butFuzzyMatchunconditionally accesses_settings.SensitiveAccentson lines 92-93. This will throwNullReferenceExceptionfor all unit tests using this constructor.🐛 Proposed fix with null-conditional fallback
// This is a workaround to allow unit tests to set the instance public StringMatcher(IAlphabet alphabet) { _alphabet = alphabet; + _settings = null; // Explicitly null for tests - FuzzyMatch will use default behavior }And update the usage on lines 92-93:
-var fullStringToCompare = _settings.SensitiveAccents ? fullStringToCompareAndNormalize : fullStringToCompareWithoutCase; -var queryToCompare = _settings.SensitiveAccents ? queryWithoutCaseAndNormalize : queryWithoutCase; +var sensitiveAccents = _settings?.SensitiveAccents ?? false; +var fullStringToCompare = sensitiveAccents ? fullStringToCompareAndNormalize : fullStringToCompareWithoutCase; +var queryToCompare = sensitiveAccents ? queryWithoutCaseAndNormalize : queryWithoutCase;🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Flow.Launcher.Infrastructure/StringMatcher.cs` around lines 27 - 31, The workaround constructor StringMatcher(IAlphabet alphabet) leaves the field _settings null which causes FuzzyMatch to throw a NullReferenceException when accessing _settings.SensitiveAccents; fix by ensuring _settings is initialized or guarded: either initialize _settings to a sensible default inside the constructor (e.g., _settings = new MatchingSettings() or the default settings object used by the other ctor) or change FuzzyMatch to use a null-safe check like (_settings?.SensitiveAccents ?? <defaultBool>) and (_settings?.PreferDiacritics ?? <defaultBool>) where those flags are read; update the constructor or the FuzzyMatch usage accordingly (referencing the StringMatcher constructor, the _settings field, and the FuzzyMatch method).
🧹 Nitpick comments (5)
Flow.Launcher/SettingPages/Views/SettingsPaneGeneral.xaml (1)
94-101: Missing Margin causes inconsistent spacing.The new SettingsCard lacks a
Marginattribute. Other cards in this file useMargin="0 4 0 0"for tight grouping orMargin="0 14 0 0"for section breaks. Additionally, this card is missing aHeaderIconwhich most other cards have.Consider whether this setting should be grouped with other search-related settings (e.g., near
querySearchPrecisionorShouldUsePinyin) rather than between notification and window positioning settings.🎨 Proposed fix to add margin and icon
<ui:SettingsCard + Margin="0 14 0 0" Description="{DynamicResource sensitiveAccentToolTip}" Header="{DynamicResource sensitiveAccent}"> + <ui:SettingsCard.HeaderIcon> + <ui:FontIcon Glyph="" /> + </ui:SettingsCard.HeaderIcon> <ui:ToggleSwitch IsOn="{Binding Settings.SensitiveAccents}" OffContent="{DynamicResource disable}" OnContent="{DynamicResource enable}" /> </ui:SettingsCard>🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Flow.Launcher/SettingPages/Views/SettingsPaneGeneral.xaml` around lines 94 - 101, The new SettingsCard for the SensitiveAccents toggle lacks the standard Margin and HeaderIcon used by other cards, causing inconsistent spacing and missing iconography; update the SettingsCard element (the one containing the ToggleSwitch bound to Settings.SensitiveAccents) to include a Margin attribute (use either "0 4 0 0" for tight grouping or "0 14 0 0" for a section break to match surrounding cards) and add the appropriate HeaderIcon attribute to match other cards, and consider relocating this SettingsCard near related search settings such as querySearchPrecision or ShouldUsePinyin if it logically groups with search-related options instead of being placed between notifications and window positioning.Flow.Launcher.Infrastructure/StringMatcher.cs (3)
257-267: Implementation is efficient but has a stack size assumption.Using
stackalloc char[value.Length]is performant for typical search queries. For extremely long strings, this could cause stack overflow, though this is unlikely in practice for search operations.If robustness is needed for arbitrary input lengths, consider a threshold:
Span<char> buffer = value.Length <= 256 ? stackalloc char[value.Length] : new char[value.Length];🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Flow.Launcher.Infrastructure/StringMatcher.cs` around lines 257 - 267, The Normalize method uses stackalloc with value.Length which can overflow the stack for very long inputs; modify Normalize to allocate the temporary buffer on the stack only for small inputs (e.g., <= 256 chars) and fall back to a heap-allocated char[] for larger inputs, then copy/use that buffer when mapping characters from AccentMap and returning the new string; update references inside Normalize to treat the heap array as a Span<char> so the rest of the logic (char.ToLowerInvariant, AccentMap.TryGetValue, new string(...)) is unchanged.
5-7: Unused imports can be removed.
System.GlobalizationandSystem.Textdon't appear to be used in the current implementation. These may be leftover from a previous approach usingNormalizationForm.FormD.🧹 Proposed cleanup
using CommunityToolkit.Mvvm.DependencyInjection; using Flow.Launcher.Plugin.SharedModels; using System; using System.Collections.Generic; -using System.Globalization; using System.Linq; -using System.Text; using Flow.Launcher.Infrastructure.UserSettings;🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Flow.Launcher.Infrastructure/StringMatcher.cs` around lines 5 - 7, Remove the unused using directives System.Globalization and System.Text from the top of StringMatcher.cs; locate the using block in the StringMatcher file (where System.Linq is present) and delete the two unused imports so only necessary namespaces remain, leaving System.Linq (and any other actually referenced namespaces) intact to avoid warnings and keep the file tidy.
246-256: AccentMap has limited language coverage.The map covers common Western European accents but misses characters from:
- Polish: ą→a, ę→e, ł→l, ń→n, ś→s, ź→z, ż→z
- Turkish: ğ→g, ş→s, ı→i
- Czech/Slovak: ř→r, š→s, ž→z, č→c, ě→e, ů→u
- Nordic: ø→o, æ→ae
- And others
This manual approach trades comprehensiveness for performance and avoids the index-mapping complexity of Unicode normalization (FormD). Consider documenting this limitation or expanding the map based on user feedback.
➕ Expanded AccentMap example
private static readonly Dictionary<char, char> AccentMap = new() { ['á'] = 'a', ['à'] = 'a', ['ã'] = 'a', ['â'] = 'a', ['ä'] = 'a', ['å'] = 'a', + ['ą'] = 'a', ['æ'] = 'a', ['é'] = 'e', ['è'] = 'e', ['ê'] = 'e', ['ë'] = 'e', + ['ę'] = 'e', ['ě'] = 'e', ['í'] = 'i', ['ì'] = 'i', ['î'] = 'i', ['ï'] = 'i', + ['ı'] = 'i', ['ó'] = 'o', ['ò'] = 'o', ['õ'] = 'o', ['ô'] = 'o', ['ö'] = 'o', + ['ø'] = 'o', ['ú'] = 'u', ['ù'] = 'u', ['û'] = 'u', ['ü'] = 'u', + ['ů'] = 'u', ['ç'] = 'c', + ['č'] = 'c', ['ñ'] = 'n', + ['ń'] = 'n', ['ý'] = 'y', ['ÿ'] = 'y', + ['ł'] = 'l', + ['ś'] = 's', ['š'] = 's', ['ş'] = 's', + ['ź'] = 'z', ['ż'] = 'z', ['ž'] = 'z', + ['ř'] = 'r', + ['ğ'] = 'g' };🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Flow.Launcher.Infrastructure/StringMatcher.cs` around lines 246 - 256, The AccentMap dictionary in StringMatcher (AccentMap) only covers a subset of Western European diacritics; update the AccentMap to include the missing language characters (Polish: ą, ę, ł, ń, ś, ź, ż; Turkish: ğ, ş, ı; Czech/Slovak: ř, š, ž, č, ě, ů; Nordic: ø, æ→map 'æ'→"ae" decision) or alternatively add a clear comment above the AccentMap field explaining its limited coverage and why Unicode FormD normalization was avoided; if expanding, ensure you add both lowercase and uppercase mappings where needed and keep the dictionary initialization in AccentMap consistent with existing style.Flow.Launcher/SettingPages/ViewModels/SettingsPaneGeneralViewModel.cs (1)
199-211: ViewModel property is unused — XAML binds directly to Settings.The XAML at line 98 binds to
{Binding Settings.SensitiveAccents}, bypassing this ViewModel property entirely. Either:
- Remove this unused property, or
- Update the XAML to bind to
{Binding SensitiveAccents}to use the ViewModel wrapperIf there's intent to add custom logic in the setter later (like
EnableDialogJumpdoes), keep the property and update the XAML binding.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Flow.Launcher/SettingPages/ViewModels/SettingsPaneGeneralViewModel.cs` around lines 199 - 211, The SensitiveAccents property on SettingsPaneGeneralViewModel is unused because the XAML binds directly to Settings.SensitiveAccents; either remove the SensitiveAccents property from SettingsPaneGeneralViewModel (and any references) or update the XAML binding at the control (currently {Binding Settings.SensitiveAccents}) to {Binding SensitiveAccents} so the ViewModel wrapper is used; if you intend to keep room for future setter logic (like EnableDialogJump) keep the property and change the XAML, otherwise delete the property to avoid dead code.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@Flow.Launcher.Infrastructure/StringMatcher.cs`:
- Around line 92-93: The boolean name _settings.SensitiveAccents is inverted vs
behavior: when true the code strips accents (accent-insensitive). Rename the
setting to a clear name like IgnoreAccents or AccentInsensitiveSearch (and
update its usages) or invert the logic where fullStringToCompare and
queryToCompare are selected; specifically update the Settings property, any
corresponding ViewModel property, related XAML binding keys, and localization
labels so the new name matches behavior, and replace references to
_settings.SensitiveAccents in StringMatcher (affecting fullStringToCompare and
queryToCompare) with the new property or inverted condition.
---
Outside diff comments:
In `@Flow.Launcher.Infrastructure/StringMatcher.cs`:
- Around line 27-31: The workaround constructor StringMatcher(IAlphabet
alphabet) leaves the field _settings null which causes FuzzyMatch to throw a
NullReferenceException when accessing _settings.SensitiveAccents; fix by
ensuring _settings is initialized or guarded: either initialize _settings to a
sensible default inside the constructor (e.g., _settings = new
MatchingSettings() or the default settings object used by the other ctor) or
change FuzzyMatch to use a null-safe check like (_settings?.SensitiveAccents ??
<defaultBool>) and (_settings?.PreferDiacritics ?? <defaultBool>) where those
flags are read; update the constructor or the FuzzyMatch usage accordingly
(referencing the StringMatcher constructor, the _settings field, and the
FuzzyMatch method).
---
Nitpick comments:
In `@Flow.Launcher.Infrastructure/StringMatcher.cs`:
- Around line 257-267: The Normalize method uses stackalloc with value.Length
which can overflow the stack for very long inputs; modify Normalize to allocate
the temporary buffer on the stack only for small inputs (e.g., <= 256 chars) and
fall back to a heap-allocated char[] for larger inputs, then copy/use that
buffer when mapping characters from AccentMap and returning the new string;
update references inside Normalize to treat the heap array as a Span<char> so
the rest of the logic (char.ToLowerInvariant, AccentMap.TryGetValue, new
string(...)) is unchanged.
- Around line 5-7: Remove the unused using directives System.Globalization and
System.Text from the top of StringMatcher.cs; locate the using block in the
StringMatcher file (where System.Linq is present) and delete the two unused
imports so only necessary namespaces remain, leaving System.Linq (and any other
actually referenced namespaces) intact to avoid warnings and keep the file tidy.
- Around line 246-256: The AccentMap dictionary in StringMatcher (AccentMap)
only covers a subset of Western European diacritics; update the AccentMap to
include the missing language characters (Polish: ą, ę, ł, ń, ś, ź, ż; Turkish:
ğ, ş, ı; Czech/Slovak: ř, š, ž, č, ě, ů; Nordic: ø, æ→map 'æ'→"ae" decision) or
alternatively add a clear comment above the AccentMap field explaining its
limited coverage and why Unicode FormD normalization was avoided; if expanding,
ensure you add both lowercase and uppercase mappings where needed and keep the
dictionary initialization in AccentMap consistent with existing style.
In `@Flow.Launcher/SettingPages/ViewModels/SettingsPaneGeneralViewModel.cs`:
- Around line 199-211: The SensitiveAccents property on
SettingsPaneGeneralViewModel is unused because the XAML binds directly to
Settings.SensitiveAccents; either remove the SensitiveAccents property from
SettingsPaneGeneralViewModel (and any references) or update the XAML binding at
the control (currently {Binding Settings.SensitiveAccents}) to {Binding
SensitiveAccents} so the ViewModel wrapper is used; if you intend to keep room
for future setter logic (like EnableDialogJump) keep the property and change the
XAML, otherwise delete the property to avoid dead code.
In `@Flow.Launcher/SettingPages/Views/SettingsPaneGeneral.xaml`:
- Around line 94-101: The new SettingsCard for the SensitiveAccents toggle lacks
the standard Margin and HeaderIcon used by other cards, causing inconsistent
spacing and missing iconography; update the SettingsCard element (the one
containing the ToggleSwitch bound to Settings.SensitiveAccents) to include a
Margin attribute (use either "0 4 0 0" for tight grouping or "0 14 0 0" for a
section break to match surrounding cards) and add the appropriate HeaderIcon
attribute to match other cards, and consider relocating this SettingsCard near
related search settings such as querySearchPrecision or ShouldUsePinyin if it
logically groups with search-related options instead of being placed between
notifications and window positioning.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: f7968958-51cd-4ab8-b534-538cd66e811a
📒 Files selected for processing (6)
Flow.Launcher.Infrastructure/StringMatcher.csFlow.Launcher.Infrastructure/UserSettings/Settings.csFlow.Launcher/Languages/en.xamlFlow.Launcher/Languages/pt-br.xamlFlow.Launcher/SettingPages/ViewModels/SettingsPaneGeneralViewModel.csFlow.Launcher/SettingPages/Views/SettingsPaneGeneral.xaml
✅ Files skipped from review due to trivial changes (2)
- Flow.Launcher/Languages/en.xaml
- Flow.Launcher/Languages/pt-br.xaml
|
Sorry for the delay — I’ve been thinking about ways to implement this and reached a possible solution, but I got a bit stuck in one situation and wanted to ask for your help if possible. I implemented this approach to handle the issue: private static readonly Dictionary<char, char> AccentMap = new()
{
['á'] = 'a', ['à'] = 'a', ['ã'] = 'a', ['â'] = 'a', ['ä'] = 'a', ['å'] = 'a',
['é'] = 'e', ['è'] = 'e', ['ê'] = 'e', ['ë'] = 'e',
['í'] = 'i', ['ì'] = 'i', ['î'] = 'i', ['ï'] = 'i',
['ó'] = 'o', ['ò'] = 'o', ['õ'] = 'o', ['ô'] = 'o', ['ö'] = 'o',
['ú'] = 'u', ['ù'] = 'u', ['û'] = 'u', ['ü'] = 'u',
['ç'] = 'c',
['ñ'] = 'n',
['ý'] = 'y', ['ÿ'] = 'y'
};
public static string Normalize(string value)
{
Span<char> buffer = stackalloc char[value.Length];
for (int i = 0; i < value.Length; i++)
{
var c = char.ToLowerInvariant(value[i]);
buffer[i] = AccentMap.TryGetValue(c, out var mapped) ? mapped : c;
}
return new string(buffer);
}And I’m using it like this inside method FuzzyMatch: var fullStringToCompareAndNormalize = opt.IgnoreCase ? Normalize(stringToCompare) : stringToCompare;
var queryWithoutCaseAndNormalize = opt.IgnoreCase ? Normalize(query) : query;
var fullStringToCompareWithoutCase = opt.IgnoreCase ? stringToCompare.ToLower() : stringToCompare;
var queryWithoutCase = opt.IgnoreCase ? query.ToLower() : query;
var fullStringToCompare = _settings.SensitiveAccents ? fullStringToCompareAndNormalize : fullStringToCompareWithoutCase;
var queryToCompare = _settings.SensitiveAccents ? queryWithoutCaseAndNormalize : queryWithoutCase;This logic is only triggered when the user enables the option via a button in the general settings. However, I realized that since the apps/results are cached in memory during queries, I would need to reset the cache whenever the user enables or disables this option so that the behavior updates correctly. I got a bit stuck on how to handle this situation properly and would appreciate any guidance on the best approach. |
Ensures the system restarts automatically when the Sensitive Accents option is enabled or disabled Improves user experience by making the restart action explicit
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@Flow.Launcher/Languages/pt-br.xaml`:
- Line 117: The resource entry with x:Key="typingStartEn" contains an accidental
trailing space after the closing </system:String> tag; edit the XAML element for
typingStartEn to remove the trailing whitespace so the line ends exactly with
</system:String> with no extra space, preserving the existing text content
"Sempre Começar Digitando em Modo Inglês".
- Around line 124-128: Add the five missing resource keys (sensitiveAccent,
sensitiveAccentToolTip, sensitiveAccentRestartTitle,
sensitiveAccentRestartMessage, sensitiveAccentRestartButton) into each of the
listed language resource XAML files (the ones currently missing these keys) as
system:String x:Key entries; if you don't have translations yet, populate them
with English fallback text matching the intent in the pt-br example so WPF
DynamicResource lookups in SettingsPaneGeneral.xaml resolve and do not render
blank UI; ensure keys are spelled exactly as shown and placed in the same
resource dictionary structure as other language entries.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 8cc778ed-764e-4122-b3b8-a8bff44783de
📒 Files selected for processing (4)
Flow.Launcher/Languages/en.xamlFlow.Launcher/Languages/pt-br.xamlFlow.Launcher/SettingPages/ViewModels/SettingsPaneGeneralViewModel.csFlow.Launcher/SettingPages/Views/SettingsPaneGeneral.xaml
🚧 Files skipped from review as they are similar to previous changes (3)
- Flow.Launcher/SettingPages/Views/SettingsPaneGeneral.xaml
- Flow.Launcher/SettingPages/ViewModels/SettingsPaneGeneralViewModel.cs
- Flow.Launcher/Languages/en.xaml
Flow.Launcher/Languages/pt-br.xaml
Outdated
| <system:String x:Key="sensitiveAccent">Ativar a distinção de acentuação na consulta de programas.</system:String> | ||
| <system:String x:Key="sensitiveAccentToolTip">Quando esta opção for alterada, será necessária uma reinicialização. Quando ativada, você poderá encontrar programas que contêm caracteres acentuados com mais facilidade.</system:String> | ||
| <system:String x:Key="sensitiveAccentRestartTitle">Reinicialização Necessária</system:String> | ||
| <system:String x:Key="sensitiveAccentRestartMessage">Alterar a sensibilidade de acentos requer a reinicialização do Flow Launcher para surtir efeito.</system:String> | ||
| <system:String x:Key="sensitiveAccentRestartButton">Reiniciar Agora</system:String> |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Description: Find language files missing the sensitiveAccent* keys
echo "=== Language files missing sensitiveAccent keys ==="
for file in $(fd -e xaml . Flow.Launcher/Languages/); do
missing=""
for key in sensitiveAccent sensitiveAccentToolTip sensitiveAccentRestartTitle sensitiveAccentRestartMessage sensitiveAccentRestartButton; do
if ! grep -q "x:Key=\"$key\"" "$file"; then
missing="$missing $key"
fi
done
if [ -n "$missing" ]; then
echo "$file: missing$missing"
fi
doneRepository: Flow-Launcher/Flow.Launcher
Length of output: 4324
Add sensitiveAccent keys to 25 other language files to prevent blank UI text.*
The five new keys (sensitiveAccent, sensitiveAccentToolTip, sensitiveAccentRestartTitle, sensitiveAccentRestartMessage, sensitiveAccentRestartButton) are missing from 25 language files. When WPF cannot resolve DynamicResource bindings in SettingsPaneGeneral.xaml, it renders blank text for those UI elements, creating a poor experience for users of those languages.
The missing files are: ar, cs, da, de, es, es-419, fr, he, it, ja, ko, nb, nb-NO, nl, pl, pt-pt, ru, sk, sr, sr-Cyrl-RS, tr, uk-UA, vi, zh-cn, zh-tw.
Add these keys to all language files—at minimum with English fallback text if translations are not yet available.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@Flow.Launcher/Languages/pt-br.xaml` around lines 124 - 128, Add the five
missing resource keys (sensitiveAccent, sensitiveAccentToolTip,
sensitiveAccentRestartTitle, sensitiveAccentRestartMessage,
sensitiveAccentRestartButton) into each of the listed language resource XAML
files (the ones currently missing these keys) as system:String x:Key entries; if
you don't have translations yet, populate them with English fallback text
matching the intent in the pt-br example so WPF DynamicResource lookups in
SettingsPaneGeneral.xaml resolve and do not render blank UI; ensure keys are
spelled exactly as shown and placed in the same resource dictionary structure as
other language entries.
There was a problem hiding this comment.
2 issues found across 3 files (changes from recent commits).
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="Flow.Launcher/Languages/en.xaml">
<violation number="1" location="Flow.Launcher/Languages/en.xaml:128">
P3: The label text says "Enable accent sensitivity", but the setting is named "ignoreAccents" and the tooltip describes accent-insensitive matching. This contradicts the actual behavior and can mislead users.</violation>
</file>
<file name="Flow.Launcher.Infrastructure/StringMatcher.cs">
<violation number="1" location="Flow.Launcher.Infrastructure/StringMatcher.cs:94">
P2: IgnoreAccents now forces lowercasing even when IgnoreCase is false, so case-sensitive searches become effectively case-insensitive whenever accent removal is enabled.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
Move the "Ignore Accents" settings card and its InfoBar to appear after the "showAtTopmost" card and before the "language" card in SettingsPaneGeneral.xaml. Adjust margin values for proper spacing. No functional changes to the controls.
Clarified the "ignoreAccents" setting and tooltip to indicate that accents are ignored when searching all results, not just programs. Updated wording from "programs" to "results".
Jack251970
left a comment
Jack251970
left a comment
There was a problem hiding this comment.
Can you tell me why we need to restart Flow when change accent option?
I think _settings.IgnoreAccents in StringMatcher.cs can be changed during runtime and we do not need to restart Flow for taking this into effect.
I also wanted it to work that way, but I couldn’t find a way to do it without needing to restart. I had thought about using the method to reset the memory cache that exists in Main in the PluginProgram, but I wasn’t able to use it or call it in any way. |
This is because Program plugin constructs memory cache for program results. Flow.Launcher/Plugins/Flow.Launcher.Plugin.Program/Main.cs Lines 38 to 39 in 6d8c35c |
|
One possible solution is to reload plugins when this option changes |
|
@jjw24 |
I had noticed that the StringMatcher method is implemented as a singleton, so it ends up being a bit complicated. That’s why I was thinking about restarting the flow for this option, since it involves the query mechanism. |
StringMatcher now listens to Settings.PropertyChanged and updates UserSettingSearchPrecision when QuerySearchPrecision changes. Removed direct assignment from Settings setter to centralize update logic within StringMatcher.
Introduce StringMatcherBehaviorChanged event in Settings, triggered on changes to string matching properties. Expose the event via PublicAPIInstance for plugin subscriptions. Update Program plugin to reset cache on event, and ensure proper event unsubscription in Dispose. Refactor IgnoreAccents property for consistency.
|
@4yinn I added a new API event to handle this. Since I do not have any programs with accent-contained names, could you please help me check if this option can take effect for Program plugin without restarting? Note that you need to wait for a moment after you changed this option and test. |
Why not use the same architecture of PinyinAlphabet? It's doesn't require a restart. |
I think 5yinn is talking about the cache behavior of Program plugin. |
|
Program plugin will cache the query and results in memory. And we meed to reset it when these settings change. |
Oh sorry I forget there's a memory cache. |
That works. |
Removed restart note from ignoreAccents tooltip. Added a new constructor and UpdateApp method to SettingsPaneGeneralViewModel for triggering asynchronous app updates.
Jack251970
left a comment
Jack251970
left a comment
There was a problem hiding this comment.
LGTM! Thanks for your contribution! Please wait for jjw24's review
|
@jjw24 Hi, this PR looks good to me! Could you please finally review it? Thanks! |
Jack251970
removed
review in progress
Added a HeaderIcon to the "Ignore Accents" SettingsCard in SettingsPaneGeneral.xaml, using a TextBlock with "áé" to visually represent the feature.
4 participants
Description
Added a preprocessing step to
FuzzySearchto normalize Unicode characters and remove diacritics before matching.This allows searches to be accent-insensitive, improving usability. For example, searching for
camerawill also matchcâmera.Related Issue
Closes #4149
Summary by cubic
Adds accent-insensitive fuzzy matching behind a settings toggle to improve results for queries with diacritics. Also wires matcher behavior to settings and exposes a public event so plugins can reset caches when matching rules change (relates to #4149).
StringMatcheronly normalizes whenSettings.IgnoreAccentsis true; still applies case folding whenopt.IgnoreCaseis true; listens toSettings.PropertyChangedand updatesUserSettingSearchPrecisionwhenQuerySearchPrecisionchanges; updated General Settings copy to say “all results.”Settings.IgnoreAccentswith a General Settings toggle and English strings; header icon (“áé”) for the Ignore Accents card;Settings.StringMatcherBehaviorChangedevent exposed viaIPublicAPI/PublicAPIInstance;Plugins/Flow.Launcher.Plugin.Programsubscribes to reset its cache when matcher behavior changes; fastNormalize(string)using a static accent map with null/empty guards andstackalloc/ArrayPool<char>; tests updated to passSettingsintoStringMatcher.Settings.QuerySearchPrecisionsetter (now handled byStringMatcher); previously added non-English translations for the new setting.ArrayPool<char>for large inputs; plugin handlers unsubscribed inDisposeto avoid leaks.StringMatcherwithSettings.Written for commit 7607844. Summary will update on new commits.