Unicode, Unicodemap and UCIS refactor (#21659)

6 files changed, 371 insertions, 15 deletions

diff --git a/quantum/unicode/ucis.c b/quantum/unicode/ucis.c
new file mode 100644
index 0000000000..32be785206
--- /dev/null
+++ b/quantum/unicode/ucis.c
@@ -0,0 +1,96 @@
+// Copyright 2023 QMK
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include "ucis.h"
+#include "unicode.h"
+#include "action.h"
+
+uint8_t count                        = 0;
+bool    active                       = false;
+char    input[UCIS_MAX_INPUT_LENGTH] = {0};
+
+void ucis_start(void) {
+    count  = 0;
+    active = true;
+
+    register_unicode(0x2328); // ⌨
+}
+
+bool ucis_active(void) {
+    return active;
+}
+
+uint8_t ucis_count(void) {
+    return count;
+}
+
+static char keycode_to_char(uint16_t keycode) {
+    if (keycode >= KC_A && keycode <= KC_Z) {
+        return 'a' + (keycode - KC_A);
+    } else if (keycode >= KC_1 && keycode <= KC_9) {
+        return '1' + (keycode - KC_1);
+    } else if (keycode == KC_0) {
+        return '0';
+    }
+    return 0;
+}
+
+bool ucis_add(uint16_t keycode) {
+    char c = keycode_to_char(keycode);
+    if (c) {
+        input[count++] = c;
+        return true;
+    }
+    return false;
+}
+
+bool ucis_remove_last(void) {
+    if (count) {
+        count--;
+        return true;
+    }
+
+    return false;
+}
+
+static bool match_mnemonic(char *mnemonic) {
+    for (uint8_t i = 0; input[i]; i++) {
+        if (i > count || input[i] != mnemonic[i]) {
+            return false;
+        }
+    }
+    return true;
+}
+
+void ucis_finish(void) {
+    uint8_t i     = 0;
+    bool    found = false;
+    for (; ucis_symbol_table[i].mnemonic; i++) {
+        if (match_mnemonic(ucis_symbol_table[i].mnemonic)) {
+            found = true;
+            break;
+        }
+    }
+
+    if (found) {
+        for (uint8_t j = 0; j <= count; j++) {
+            tap_code(KC_BACKSPACE);
+        }
+        register_ucis(i);
+    }
+
+    active = false;
+}
+
+void ucis_cancel(void) {
+    count  = 0;
+    active = false;
+}
+
+void register_ucis(uint8_t index) {
+    const uint32_t *code_points = ucis_symbol_table[index].code_points;
+
+    for (int i = 0; i < UCIS_MAX_CODE_POINTS && code_points[i]; i++) {
+        register_unicode(code_points[i]);
+    }
+}
diff --git a/quantum/unicode/ucis.h b/quantum/unicode/ucis.h
new file mode 100644
index 0000000000..5a4fa26784
--- /dev/null
+++ b/quantum/unicode/ucis.h
@@ -0,0 +1,97 @@
+// Copyright 2023 QMK
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include <stdbool.h>
+#include <stdint.h>
+
+/**
+ * \file
+ *
+ * \defgroup ucis UCIS
+ * \{
+ */
+
+#ifndef UCIS_MAX_INPUT_LENGTH
+#    define UCIS_MAX_INPUT_LENGTH 32
+#endif
+
+#ifndef UCIS_MAX_CODE_POINTS
+#    define UCIS_MAX_CODE_POINTS 3
+#endif
+
+typedef struct {
+    char*    mnemonic;
+    uint32_t code_points[UCIS_MAX_CODE_POINTS];
+} ucis_symbol_t;
+
+// clang-format off
+
+#define UCIS_TABLE(...) { \
+    __VA_ARGS__,          \
+    { NULL, {} }          \
+}
+
+#define UCIS_SYM(name, ...) {    \
+    .mnemonic    = name,         \
+    .code_points = {__VA_ARGS__} \
+}
+
+// clang-format on
+
+extern const ucis_symbol_t ucis_symbol_table[];
+
+/**
+ * \brief Begin the input sequence.
+ */
+void ucis_start(void);
+
+/**
+ * \brief Whether UCIS is currently active.
+ *
+ * \return `true` if UCIS is active.
+ */
+bool ucis_active(void);
+
+/**
+ * \brief Get the number of characters in the input sequence buffer.
+ *
+ * \return The current input sequence buffer length.
+ */
+uint8_t ucis_count(void);
+
+/**
+ * \brief Add the given keycode to the input sequence buffer.
+ *
+ * \param keycode The keycode to add. Must be between `KC_A` and `KC_Z`, or `KC_1` and `KC_0`.
+ *
+ * \return `true` if the keycode was added.
+ */
+bool ucis_add(uint16_t keycode);
+
+/**
+ * \brief Remove the last character from the input sequence.
+ *
+ * \return `true` if the sequence was not empty.
+ */
+bool ucis_remove_last(void);
+
+/**
+ * Mark the input sequence as complete, and attempt to match.
+ */
+void ucis_finish(void);
+
+/**
+ * \brief Cancel the input sequence.
+ */
+void ucis_cancel(void);
+
+/**
+ * Send the code point(s) for the given UCIS index.
+ *
+ * \param index The index into the UCIS symbol table.
+ */
+void register_ucis(uint8_t index);
+
+/** \} */
diff --git a/quantum/unicode/unicode.c b/quantum/unicode/unicode.c
index f92ba51984..78a4cad585 100644
--- a/quantum/unicode/unicode.c
+++ b/quantum/unicode/unicode.c
@@ -73,16 +73,8 @@ static int8_t  selected_count = ARRAY_SIZE(selected);
 static int8_t  selected_index;
 #endif
 
-/** \brief unicode input mode set at user level
- *
- * Run user code on unicode input mode change
- */
 __attribute__((weak)) void unicode_input_mode_set_user(uint8_t input_mode) {}
 
-/** \brief unicode input mode set at keyboard level
- *
- *  Run keyboard code on unicode input mode change
- */
 __attribute__((weak)) void unicode_input_mode_set_kb(uint8_t input_mode) {
     unicode_input_mode_set_user(input_mode);
 }
@@ -172,6 +164,10 @@ uint8_t get_unicode_input_mode(void) {
     return unicode_config.input_mode;
 }
 
+static void persist_unicode_input_mode(void) {
+    eeprom_update_byte(EECONFIG_UNICODEMODE, unicode_config.input_mode);
+}
+
 void set_unicode_input_mode(uint8_t mode) {
     unicode_config.input_mode = mode;
     persist_unicode_input_mode();
@@ -182,26 +178,34 @@ void set_unicode_input_mode(uint8_t mode) {
     dprintf("Unicode input mode set to: %u\n", unicode_config.input_mode);
 }
 
-void cycle_unicode_input_mode(int8_t offset) {
+static void cycle_unicode_input_mode(int8_t offset) {
 #if UNICODE_SELECTED_MODES != -1
     selected_index = (selected_index + offset) % selected_count;
     if (selected_index < 0) {
         selected_index += selected_count;
     }
+
     unicode_config.input_mode = selected[selected_index];
+
 #    if UNICODE_CYCLE_PERSIST
     persist_unicode_input_mode();
 #    endif
+
 #    ifdef AUDIO_ENABLE
     unicode_play_song(unicode_config.input_mode);
 #    endif
+
     unicode_input_mode_set_kb(unicode_config.input_mode);
     dprintf("Unicode input mode cycle to: %u\n", unicode_config.input_mode);
 #endif
 }
 
-void persist_unicode_input_mode(void) {
-    eeprom_update_byte(EECONFIG_UNICODEMODE, unicode_config.input_mode);
+void unicode_input_mode_step(void) {
+    cycle_unicode_input_mode(1);
+}
+
+void unicode_input_mode_step_reverse(void) {
+    cycle_unicode_input_mode(-1);
 }
 
 __attribute__((weak)) void unicode_input_start(void) {
diff --git a/quantum/unicode/unicode.h b/quantum/unicode/unicode.h
index fc0e9d3d2b..90a54c8b18 100644
--- a/quantum/unicode/unicode.h
+++ b/quantum/unicode/unicode.h
@@ -19,6 +19,13 @@
 #include <stdint.h>
 #include "unicode_keycodes.h"
 
+/**
+ * \file
+ *
+ * \defgroup unicode Unicode
+ * \{
+ */
+
 typedef union {
     uint8_t raw;
     struct {
@@ -41,21 +48,87 @@ enum unicode_input_modes {
     UNICODE_MODE_COUNT // Number of available input modes (always leave at the end)
 };
 
-void    unicode_input_mode_init(void);
+void unicode_input_mode_init(void);
+
+/**
+ * \brief Get the current Unicode input mode.
+ *
+ * \return The currently active Unicode input mode.
+ */
 uint8_t get_unicode_input_mode(void);
-void    set_unicode_input_mode(uint8_t mode);
-void    cycle_unicode_input_mode(int8_t offset);
-void    persist_unicode_input_mode(void);
 
+/**
+ * \brief Set the Unicode input mode.
+ *
+ * \param mode The input mode to set.
+ */
+void set_unicode_input_mode(uint8_t mode);
+
+/**
+ * \brief Change to the next Unicode input mode.
+ */
+void unicode_input_mode_step(void);
+
+/**
+ * \brief Change to the previous Unicode input mode.
+ */
+void unicode_input_mode_step_reverse(void);
+
+/**
+ * \brief User-level callback, invoked when the input mode is changed.
+ *
+ * \param input_mode The new input mode.
+ */
 void unicode_input_mode_set_user(uint8_t input_mode);
+
+/**
+ * \brief Keyboard-level callback, invoked when the input mode is changed.
+ *
+ * \param input_mode The new input mode.
+ */
 void unicode_input_mode_set_kb(uint8_t input_mode);
 
+/**
+ * \brief Begin the Unicode input sequence. The exact behavior depends on the currently selected input mode.
+ */
 void unicode_input_start(void);
+
+/**
+ * \brief Complete the Unicode input sequence. The exact behavior depends on the currently selected input mode.
+ */
 void unicode_input_finish(void);
+
+/**
+ * \brief Cancel the Unicode input sequence. The exact behavior depends on the currently selected input mode.
+ */
 void unicode_input_cancel(void);
 
+/**
+ * \brief Send a 16-bit hex number.
+ *
+ * \param hex The number to send.
+ */
 void register_hex(uint16_t hex);
+
+/**
+ * \brief Send a 32-bit hex number.
+ *
+ * \param hex The number to send.
+ */
 void register_hex32(uint32_t hex);
+
+/**
+ * \brief Input a single Unicode character. A surrogate pair will be sent if required by the input mode.
+ *
+ * \param code_point The code point of the character to send.
+ */
 void register_unicode(uint32_t code_point);
 
+/**
+ * \brief Send a string containing Unicode characters.
+ *
+ * \param str The string to send.
+ */
 void send_unicode_string(const char *str);
+
+/** \} */
diff --git a/quantum/unicode/unicodemap.c b/quantum/unicode/unicodemap.c
new file mode 100644
index 0000000000..2f618056a2
--- /dev/null
+++ b/quantum/unicode/unicodemap.c
@@ -0,0 +1,43 @@
+// Copyright 2023 QMK
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include "unicodemap.h"
+#include "unicode.h"
+#include "keycodes.h"
+#include "quantum_keycodes.h"
+#include "modifiers.h"
+#include "host.h"
+#include "action_util.h"
+
+uint8_t unicodemap_index(uint16_t keycode) {
+    if (keycode >= QK_UNICODEMAP_PAIR) {
+        // Keycode is a pair: extract index based on Shift / Caps Lock state
+        uint16_t index;
+
+        uint8_t mods = get_mods() | get_weak_mods();
+#ifndef NO_ACTION_ONESHOT
+        mods |= get_oneshot_mods();
+#endif
+
+        bool shift = mods & MOD_MASK_SHIFT;
+        bool caps  = host_keyboard_led_state().caps_lock;
+        if (shift ^ caps) {
+            index = QK_UNICODEMAP_PAIR_GET_SHIFTED_INDEX(keycode);
+        } else {
+            index = QK_UNICODEMAP_PAIR_GET_UNSHIFTED_INDEX(keycode);
+        }
+
+        return index;
+    } else {
+        // Keycode is a regular index
+        return QK_UNICODEMAP_GET_INDEX(keycode);
+    }
+}
+
+uint32_t unicodemap_get_code_point(uint8_t index) {
+    return pgm_read_dword(unicode_map + index);
+}
+
+void register_unicodemap(uint8_t index) {
+    register_unicode(unicodemap_get_code_point(index));
+}
diff --git a/quantum/unicode/unicodemap.h b/quantum/unicode/unicodemap.h
new file mode 100644
index 0000000000..00d2aec8fa
--- /dev/null
+++ b/quantum/unicode/unicodemap.h
@@ -0,0 +1,43 @@
+// Copyright 2023 QMK
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#pragma once
+
+#include <stdint.h>
+#include "progmem.h"
+
+/**
+ * \file
+ *
+ * \defgroup unicodemap Unicode Map
+ * \{
+ */
+
+extern const uint32_t unicode_map[] PROGMEM;
+
+/**
+ * \brief Get the index into the `unicode_map` array for the given keycode, respecting shift state for pair keycodes.
+ *
+ * \param keycode The Unicode Map keycode to get the index of.
+ *
+ * \return An index into the `unicode_map` array.
+ */
+uint8_t unicodemap_index(uint16_t keycode);
+
+/**
+ * \brief Get the code point for the given index in the `unicode_map` array.
+ *
+ * \param index The index into the `unicode_map` array.
+ *
+ * \return A Unicode code point value.
+ */
+uint32_t unicodemap_get_code_point(uint8_t index);
+
+/**
+ * \brief Send the code point for the given index in the `unicode_map` array.
+ *
+ * \param index The index into the `unicode_map` array.
+ */
+void register_unicodemap(uint8_t index);
+
+/** \} */