core: fzy finder singleton

Author: CMurtagh-LGTM

src/core/CMakeLists.txt

@@ -39,6 +39,7 @@ qt_add_library(quickshell-core STATIC
 	scriptmodel.cpp
 	colorquantizer.cpp
 	toolsupport.cpp
+	fzy.cpp
 )
 
 qt_add_qml_module(quickshell-core

src/core/fzy.cpp

@@ -0,0 +1,231 @@
+#include "fzy.h"
+
+#include <algorithm>
+#include <array>
+#include <bitset>
+#include <generator>
+#include <string_view>
+
+namespace {
+constexpr size_t MATCH_MAX_LEN = 1024;
+
+constexpr double SCORE_MAX = std::numeric_limits<double>::infinity();
+constexpr double SCORE_MIN = -std::numeric_limits<double>::infinity();
+
+constexpr double SCORE_GAP_LEADING = -0.005;
+constexpr double SCORE_GAP_TRAILING = -0.005;
+constexpr double SCORE_GAP_INNER = -0.01;
+constexpr double SCORE_MATCH_CONSECUTIVE = 1.0;
+constexpr double SCORE_MATCH_SLASH = 0.9;
+constexpr double SCORE_MATCH_WORD = 0.8;
+constexpr double SCORE_MATCH_CAPITAL = 0.7;
+constexpr double SCORE_MATCH_DOT = 0.6;
+
+struct ScoredResult {
+    double score{};
+    QString str;
+    QObject* obj = nullptr;
+};
+
+bool isUpper(char16_t ch) {
+    return 'A' <= ch && ch <= 'Z';
+}
+
+bool isOrdinary(char16_t ch) {
+    return
+        ('0' <= ch && ch <= '9') ||
+        ('a' <= ch && ch <= 'z') ||
+        ('A' <= ch && ch <= 'Z')
+    ;
+}
+
+// This is from llvm but with char16_t
+// But will return pointer to end rather than nullptr
+const char16_t* strpbrk(const char16_t *src, const char16_t *segment) {
+    std::bitset<256> bitset;
+
+  for (; *segment; ++segment) // NOLINT
+    bitset.set(*reinterpret_cast<const unsigned char *>(segment));
+  for (; *src && !bitset.test(*reinterpret_cast<const unsigned char *>(src));
+       ++src) // NOLINT
+    ;
+  return src; // NOLINT
+}
+
+const char16_t *strcasechr(const char16_t *s, char16_t c) {
+	const char16_t accept[3] = {c, static_cast<char16_t>(toupper(c)), 0}; // NOLINT
+	return strpbrk(s, accept);
+}
+
+bool hasMatch(std::u16string_view needle, std::u16string_view haystack) {
+    const auto *haystackIter = haystack.begin();
+    for (auto needleChar : needle){
+	haystackIter = strcasechr(haystackIter, needleChar);
+        if (haystackIter == haystack.end()) {
+            return false;
+        }
+        haystackIter++; // NOLINT
+    }
+    return true;
+}
+
+struct MatchStruct {
+    std::u16string lowerNeedle;
+    std::u16string lowerHaystack;
+
+    std::array<double, MATCH_MAX_LEN> matchBonus{};
+};
+
+
+double getBonus(char16_t ch, char16_t lastCh){
+    if (!isOrdinary(lastCh)) {
+        return 0.0;
+    }
+    switch (ch) {
+        case '/':
+            return SCORE_MATCH_SLASH;
+        case '-':
+        case '_':
+        case ' ': return SCORE_MATCH_WORD;
+        case '.': return SCORE_MATCH_DOT;
+        case 'a':
+        case 'b':
+        case 'c':
+        case 'd':
+        case 'e':
+        case 'f':
+        case 'g':
+        case 'h':
+        case 'i':
+        case 'j':
+        case 'k':
+        case 'l':
+        case 'm':
+        case 'n':
+        case 'o':
+        case 'p':
+        case 'q':
+        case 'r':
+        case 's':
+        case 't':
+        case 'u':
+        case 'v':
+        case 'w':
+        case 'x':
+        case 'y':
+        case 'z':
+            return isUpper(lastCh) ? SCORE_MATCH_CAPITAL : 0.0;
+        default: return 0.0;
+    }
+}
+
+void precomputeBonus(std::u16string_view haystack, std::span<double> matchBonus) {
+    /* Which positions are beginning of words */
+    char16_t lastCh = '/';
+    for (size_t index = 0; index < haystack.size(); index++) {
+        char16_t ch = haystack[index];
+        matchBonus[index] = getBonus(lastCh, ch);
+        lastCh = ch;
+    }
+}
+
+MatchStruct setupMatchStruct(std::u16string_view needle, std::u16string_view haystack) {
+    MatchStruct match{};
+
+    for (const auto nch : needle){
+	match.lowerNeedle.push_back(tolower(nch));
+    }
+    for (const auto hch : haystack){
+	match.lowerHaystack.push_back(tolower(hch));
+    }
+
+    precomputeBonus(haystack, match.matchBonus);
+
+    return match;
+}
+
+void matchRow(const MatchStruct& match, size_t row, std::span<double> currD, std::span<double> currM, std::span<const double> lastD, std::span<const double> lastM) {
+    size_t needleLen = match.lowerNeedle.size();
+    size_t haystackLen = match.lowerHaystack.size();
+
+    std::u16string_view lowerNeedle = match.lowerNeedle;
+    std::u16string_view lowerHaystack = match.lowerHaystack;
+    const auto& matchBonus = match.matchBonus;
+
+    double prevScore = SCORE_MIN;
+    double gapScore = row == needleLen - 1 ? SCORE_GAP_TRAILING : SCORE_GAP_INNER;
+
+    /* These will not be used with this value, but not all compilers see it */
+    double prevM = SCORE_MIN;
+    double prevD = SCORE_MIN;
+
+    for (size_t index = 0; index < haystackLen; index++) {
+        if (lowerNeedle[row] == lowerHaystack[index]) {
+            double score = SCORE_MIN;
+            if (!row) {
+                score = (static_cast<double>(index) * SCORE_GAP_LEADING) + matchBonus[index];
+            } else if (index) { /* row > 0 && index > 0*/
+                score = fmax(
+                        prevM + matchBonus[index],
+
+                        /* consecutive match, doesn't stack with match_bonus */
+                        prevD + SCORE_MATCH_CONSECUTIVE);
+            }
+            prevD = lastD[index];
+            prevM = lastM[index];
+            currD[index] = score;
+            currM[index] = prevScore = fmax(score, prevScore + gapScore);
+        } else {
+            prevD = lastD[index];
+            prevM = lastM[index];
+            currD[index] = SCORE_MIN;
+            currM[index] = prevScore = prevScore + gapScore;
+        }
+    }
+}
+
+double match(std::u16string_view needle, std::u16string_view haystack) {
+    if (needle.empty())
+        return SCORE_MIN;
+
+    if (haystack.size() > MATCH_MAX_LEN || needle.size() > haystack.size()) {
+        return SCORE_MIN;
+    } else if (haystack.size() == needle.size()){
+	return SCORE_MAX;
+    }
+
+    MatchStruct match = setupMatchStruct(needle, haystack);
+
+    /*
+     * D Stores the best score for this position ending with a match.
+     * M Stores the best possible score at this position.
+     */
+    std::array<double, MATCH_MAX_LEN> d{};
+    std::array<double, MATCH_MAX_LEN> m{};
+
+    for (size_t index = 0; index < needle.size(); index++) {
+        matchRow(match, index, d, m, d, m);
+    }
+
+    return m[haystack.size() - 1];
+}
+
+}
+
+namespace qs {
+
+QList<QObject*> FzyFinder::filter(const QString& needle, const QList<QObject*>& haystacks, const QString& name) {
+    std::vector<ScoredResult> list;
+    for (const auto& haystack : haystacks){
+	const auto h = haystack->property(name.toUtf8()).toString();
+	if (hasMatch(needle, h)) {
+	    list.emplace_back(match(needle, h), h, haystack);
+	}
+    }
+    std::ranges::stable_sort(list, std::ranges::greater(), &ScoredResult::score);
+    auto out = QList<QObject*>(static_cast<qsizetype>(list.size()));
+    std::ranges::transform(list, out.begin(), [](const ScoredResult& result) -> QObject* { return result.obj; });
+    return out; 
+}
+
+}

src/core/fzy.h

@@ -0,0 +1,36 @@
+#pragma once
+
+#include <qlist.h>
+#include <qobject.h>
+#include <qqmlintegration.h>
+#include <qstring.h>
+
+namespace qs {
+
+///! A fzy finder.
+/// A fzy finder.
+///
+/// You can use this singleton to filter desktop entries like below.
+///
+/// ```qml
+/// model: ScriptModel {
+///   values: FzyFinder.filter(search.text, @@DesktopEntries.applications.values, "name");
+/// }
+/// ```
+class FzyFinder : public QObject {
+	Q_OBJECT;
+	QML_SINGLETON;
+	QML_ELEMENT;
+
+public:
+	explicit FzyFinder(QObject* parent = nullptr): QObject(parent) {}
+
+	/// Filters the list haystacks that don't contain the needle.
+	/// `needle` is the query to search with.
+	/// `haystacks` is what will be searched.
+	/// `name` is a property of each object in `haystacks` if `haystacks[n].name` is not a `string` then it will be treated as an empty string.
+	/// The returned list is the objects that contain the query in fzy score order.
+	Q_INVOKABLE [[nodiscard]] static QList<QObject*> filter(const QString& needle, const QList<QObject*>& haystacks, const QString& name);
+};
+
+}

src/core/module.md

@@ -30,5 +30,6 @@ headers = [
 	"clock.hpp",
 	"scriptmodel.hpp",
 	"colorquantizer.hpp",
+	"fzy.hpp",
 ]
 -----