[Enhancement] *Multi*dict

data/constants/keycodes/keycodes_0.0.4_quantum.hjson

@@ -2,9 +2,9 @@
     "keycodes": {
        "0x7C7B": {
             "group": "quantum",
-            "key": "QK_AUTOCORRECT_BANK_TOGGLE",
+            "key": "QK_AUTOCORRECT_DICT_CYCLE",
             "aliases": [
-                "AC_BANK"
+                "AC_DICT"
             ]
         }
     }

docs/feature_autocorrect.md

@@ -55,10 +55,15 @@ This file will look like this:
 // ouput         -> output
 // widht         -> width
 
-#define AUTOCORRECT_MIN_LENGTH 5  // "ouput"
-#define AUTOCORRECT_MAX_LENGTH 6  // ":thier"
+#define N_DICTS 1
+
+static const uint16_t autocorrect_offsets[N_DICTS] PROGMEM     = {0};
+static const uint16_t autocorrect_min_lengths[N_DICTS] PROGMEM = {5};
+static const uint16_t autocorrect_max_lengths[N_DICTS] PROGMEM = {6};
+static const uint16_t autocorrect_sizes[N_DICTS] PROGMEM       = {74};
 
 #define DICTIONARY_SIZE 74
+#define TYPO_BUFFER_SIZE 6
 
 static const uint8_t autocorrect_data[DICTIONARY_SIZE] PROGMEM = {85, 7, 0, 23, 35, 0, 0, 8, 0, 76, 16, 0, 15, 25, 0, 0,
     11, 23, 44, 0, 130, 101, 105, 114, 0, 23, 12, 9, 0, 131, 108, 116, 101, 114, 0, 75, 42, 0, 24, 64, 0, 0, 71, 49, 0,
@@ -86,11 +91,13 @@ The `qmk generate-autocorrect-data` commands can make an effort to check for ent
 
 ### Using multiple dictionaries
 
-Including an additional dictionary under the file `autocorrect_data_alt.h` allows for on the fly switching between two sets of autocorrection rules,
-useful for bilingual users or for running context-specific rulesets.
-`QK_AUTOCORRECT_BANK_TOGGLE` can then be used to toggle the active dictionary and persist the selection to eeprom.
+Including additional dictionaries allows for on-the-fly switching between sets of autocorrection rules, useful for bilingual users or for running context-specific rulesets.
+`QK_AUTOCORRECT_DICT_CYCLE` can then be used to cycle the active dictionary and persist the selection to eeprom.
 
-To use this feature, `autocorrect_data_alt.h` should be generated using the flag `-a`, which constructs an alternate dictionary with appropriately named constants
+To use this feature, you can input several files into `qmk generate-autocorrect-data`.
+```sh
+qmk generate-autocorrect-data dict1.txt dict2.txt
+```
 
 ## Overriding Autocorrect
 
@@ -106,12 +113,12 @@ Additionally, you can use the `AC_TOGG` keycode to toggle the on/off status for
 
 ### Keycodes :id=keycodes
 
-|Keycode                     |Aliases  |Description                                                                           |
-|----------------------------|---------|--------------------------------------------------------------------------------------|
-|`QK_AUTOCORRECT_ON`         |`AC_ON`  |Turns on the Autocorrect feature.                                                     |
-|`QK_AUTOCORRECT_OFF`        |`AC_OFF` |Turns off the Autocorrect feature.                                                    |
-|`QK_AUTOCORRECT_TOGGLE`     |`AC_TOGG`|Toggles the status of the Autocorrect feature.                                        |
-|`QK_AUTOCORRECT_BANK_TOGGLE`|`AC_BANK`|Toggles the dictionary in use by the Autocorrect feature (if an alternate is present).|
+|Keycode                     |Aliases  |Description                                                      |
+|----------------------------|---------|-----------------------------------------------------------------|
+|`QK_AUTOCORRECT_ON`         |`AC_ON`  |Turns on the Autocorrect feature.                                |
+|`QK_AUTOCORRECT_OFF`        |`AC_OFF` |Turns off the Autocorrect feature.                               |
+|`QK_AUTOCORRECT_TOGGLE`     |`AC_TOGG`|Toggles the status of the Autocorrect feature.                   |
+|`QK_AUTOCORRECT_DICT_CYCLE` |`AC_DICT`|Cycle through dictionaries, reverse direction when Shift is held.|
 
 ## User Callback Functions
 
@@ -254,15 +261,13 @@ bool apply_autocorrect(uint8_t backspaces, const char *str, char *typo, char *co
 
 Additional user callback functions to manipulate Autocorrect:
 
-| Function                   | Description                                                                      |
-|----------------------------|----------------------------------------------------------------------------------|
-| `autocorrect_enable()`     | Turns Autocorrect on.                                                            |
-| `autocorrect_disable()`    | Turns Autocorrect off.                                                           |
-| `autocorrect_toggle()`     | Toggles Autocorrect.                                                             |
-| `autocorrect_is_enabled()` | Returns true if Autocorrect is currently on.                                     |
-| `autocorrect_bank_toggle()`| Toggles and initializes Autocorrect dictionary (if multiple present)             |
-| `autocorrect_init_bank()`  | Initializes current dictionary according so selected bank  (if multiple present) |
-
+| Function                              | Description                                                    |
+|---------------------------------------|----------------------------------------------------------------|
+| `autocorrect_enable()`                | Turns Autocorrect on.                                          |
+| `autocorrect_disable()`               | Turns Autocorrect off.                                         |
+| `autocorrect_toggle()`                | Toggles Autocorrect.                                           |
+| `autocorrect_is_enabled()`            | Returns true if Autocorrect is currently on.                   |
+| `autocorrect_dict_cycle(bool forward)`| Cycles through dictionaries (parameter controls the direction).|
 
 ## Appendix: Trie binary data format :id=appendix
 

docs/keycodes.md

@@ -270,13 +270,12 @@ See also: [Auto Shift](feature_auto_shift.md)
 
 See also: [Autocorrect](feature_autocorrect.md)
 
-|Keycode                     |Aliases  |Description                                                                           |
-|----------------------------|---------|--------------------------------------------------------------------------------------|
-|`QK_AUTOCORRECT_ON`         |`AC_ON`  |Turns on the Autocorrect feature.                                                     |
-|`QK_AUTOCORRECT_OFF`        |`AC_OFF` |Turns off the Autocorrect feature.                                                    |
-|`QK_AUTOCORRECT_TOGGLE`     |`AC_TOGG`|Toggles the status of the Autocorrect feature.                                        |
-|`QK_AUTOCORRECT_BANK_TOGGLE`|`AC_BANK`|Toggles the dictionary in use by the Autocorrect feature (if an alternate is present).|
-
+|Keycode                     |Aliases  |Description                                                      |
+|----------------------------|---------|-----------------------------------------------------------------|
+|`QK_AUTOCORRECT_ON`         |`AC_ON`  |Turns on the Autocorrect feature.                                |
+|`QK_AUTOCORRECT_OFF`        |`AC_OFF` |Turns off the Autocorrect feature.                               |
+|`QK_AUTOCORRECT_TOGGLE`     |`AC_TOGG`|Toggles the status of the Autocorrect feature.                   |
+|`QK_AUTOCORRECT_DICT_CYCLE` |`AC_DICT`|Cycle through dictionaries, reverse direction when Shift is held.|
 
 ## Backlighting :id=backlighting
 

lib/python/qmk/cli/generate/autocorrect_data.py

@@ -1,4 +1,4 @@
 # Copyright 2021 Google LLC
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -29,6 +29,7 @@
 
 import sys
 import textwrap
+from pathlib import Path
 from typing import Any, Dict, Iterator, List, Tuple
 
 from milc import cli
@@ -249,17 +250,33 @@
     return f'0x{b:02X}'
 
 
-@cli.argument('filename', type=normpath, help='The autocorrection database file')
+class AutocorrectDict:
+    """Class to generate autocorrect data from a file."""
+    def __init__(self, path: Path):
+        self.path = path
+        self.autocorrections = parse_file(self.path)
+
+        self.trie = make_trie(self.autocorrections)
+        self.data = serialize_trie(self.autocorrections, self.trie)
+        assert all(0 <= b <= 255 for b in self.data)
+
+        self.min_typo = min(self.autocorrections, key=typo_len)[0]
+        self.max_typo = max(self.autocorrections, key=typo_len)[0]
+
+
+@cli.argument('filenames', type=normpath, help='The autocorrection database file(s)', nargs='+')
 @cli.argument('-kb', '--keyboard', type=keyboard_folder, completer=keyboard_completer, help='The keyboard to build a firmware for. Ignored when a configurator export is supplied.')
 @cli.argument('-km', '--keymap', completer=keymap_completer, help='The keymap to build a firmware for. Ignored when a configurator export is supplied.')
 @cli.argument('-o', '--output', arg_only=True, type=normpath, help='File to write to')
 @cli.argument('-q', '--quiet', arg_only=True, action='store_true', help="Quiet mode, only output error messages")
 @cli.argument('-a', '--alternate', arg_only=True, action='store_true', help="Create an alternate dictionary file")
 @cli.subcommand('Generate the autocorrection data file from a dictionary file.')
 def generate_autocorrect_data(cli):
-    autocorrections = parse_file(cli.args.filename)
-    trie = make_trie(autocorrections)
-    data = serialize_trie(autocorrections, trie)
+    if len(cli.args.filenames) > 8:
+        cli.log.error("Current EEPROM settings can only index up to 8 dicts")
+        sys.exit(1)
+
+    autocorrect_dicts = [AutocorrectDict(path) for path in cli.args.filenames]
 
     current_keyboard = cli.args.keyboard or cli.config.user.keyboard or cli.config.generate_autocorrect_data.keyboard
     current_keymap = cli.args.keymap or cli.config.user.keymap or cli.config.generate_autocorrect_data.keymap
@@ -271,22 +288,53 @@
     if current_keyboard and current_keymap:
         cli.args.output = locate_keymap(current_keyboard, current_keymap).parent / file_name
 
-    assert all(0 <= b <= 255 for b in data)
-
-    min_typo = min(autocorrections, key=typo_len)[0]
-    max_typo = max(autocorrections, key=typo_len)[0]
-
     # Build the autocorrect_data.h file.
     autocorrect_data_h_lines = [GPL2_HEADER_C_LIKE, GENERATED_HEADER_C_LIKE, '#pragma once', '']
 
-    autocorrect_data_h_lines.append(f'// Autocorrection dictionary ({len(autocorrections)} entries):')
-    for typo, correction in autocorrections:
-        autocorrect_data_h_lines.append(f'//   {typo:<{len(max_typo)}} -> {correction}')
+    n_entries = sum(len(dict_.autocorrections) for dict_ in autocorrect_dicts)
+    autocorrect_data_h_lines.append(f'// Autocorrection dictionary ({n_entries} entries):')
 
+    # collect all information, and write the corrections as comments
+    data = []
+    maxs = []
+    mins = []
+    sizes = []
+    for dict_ in autocorrect_dicts:
+        autocorrect_data_h_lines.append(f'// From {dict_.path}')
+        for typo, correction in dict_.autocorrections:
+            autocorrect_data_h_lines.append(f'//   {typo:<{len(dict_.max_typo)}} -> {correction}')
+        autocorrect_data_h_lines.append('// ' + '-' * 15)
+
+        data.extend(dict_.data)
+        maxs.append(len(dict_.max_typo))
+        mins.append(len(dict_.min_typo))
+        sizes.append(len(dict_.data))
+
+    assert(sum(sizes) == len(data))
+
+    offsets = [0]
+    for size in sizes[:-1]:
+        offsets.append(offsets[-1] + size)
+
+    def list_to_str(x: list) -> str:
+        """Helper to stringify the lists"""
+        return ', '.join(map(str, x))
+
+    offsets_str = list_to_str(offsets)
+    maxs_str = list_to_str(maxs)
+    mins_str = list_to_str(mins)
+    sizes_str = list_to_str(sizes)
+
+    autocorrect_data_h_lines.append('')
+    autocorrect_data_h_lines.append(f'#define N_DICTS {len(autocorrect_dicts)}')
+    autocorrect_data_h_lines.append('')
+    autocorrect_data_h_lines.append(f'static const uint16_t autocorrect_offsets[N_DICTS] PROGMEM     = {{{offsets_str}}};')
+    autocorrect_data_h_lines.append(f'static const uint16_t autocorrect_min_lengths[N_DICTS] PROGMEM = {{{mins_str}}};')
+    autocorrect_data_h_lines.append(f'static const uint16_t autocorrect_max_lengths[N_DICTS] PROGMEM = {{{maxs_str}}};')
+    autocorrect_data_h_lines.append(f'static const uint16_t autocorrect_sizes[N_DICTS] PROGMEM       = {{{sizes_str}}};')
     autocorrect_data_h_lines.append('')
-    autocorrect_data_h_lines.append(f'#define AUTOCORRECT_MIN_LENGTH{defines_suffix} {len(min_typo)} // "{min_typo}"')
-    autocorrect_data_h_lines.append(f'#define AUTOCORRECT_MAX_LENGTH{defines_suffix} {len(max_typo)} // "{max_typo}"')
-    autocorrect_data_h_lines.append(f'#define DICTIONARY_SIZE{defines_suffix} {len(data)}')
+    autocorrect_data_h_lines.append(f'#define DICTIONARY_SIZE {sum(sizes)}')
+    autocorrect_data_h_lines.append(f'#define TYPO_BUFFER_SIZE {max(maxs)}')
     autocorrect_data_h_lines.append('')
     autocorrect_data_h_lines.append(f'static const uint8_t autocorrect_data{static_suffix}[DICTIONARY_SIZE{defines_suffix}] PROGMEM = {{')
     autocorrect_data_h_lines.append(textwrap.fill('    %s' % (', '.join(map(to_hex, data))), width=100, subsequent_indent='    '))

quantum/keycode_config.h

@@ -1,4 +1,4 @@
 /* Copyright 2016 Jack Humbert
  *
  * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -44,7 +44,7 @@
         bool oneshot_enable : 1;
         bool swap_escape_capslock : 1;
         bool autocorrect_enable : 1;
-        bool autocorrect_bank : 1;
+        uint8_t autocorrect_curr_dict : 3;
     };
 } keymap_config_t;
 

quantum/keycodes.h

@@ -723,7 +723,7 @@ enum qk_keycode_defines {
     QK_TRI_LAYER_UPPER = 0x7C78,
     QK_REPEAT_KEY = 0x7C79,
     QK_ALT_REPEAT_KEY = 0x7C7A,
-    QK_AUTOCORRECT_BANK_TOGGLE = 0x7C7B,
+    QK_AUTOCORRECT_DICT_CYCLE = 0x7C7B,
     QK_KB_0 = 0x7E00,
     QK_KB_1 = 0x7E01,
     QK_KB_2 = 0x7E02,
@@ -1367,7 +1367,7 @@ enum qk_keycode_defines {
     TL_UPPR    = QK_TRI_LAYER_UPPER,
     QK_REP     = QK_REPEAT_KEY,
     QK_AREP    = QK_ALT_REPEAT_KEY,
-    AC_BANK    = QK_AUTOCORRECT_BANK_TOGGLE,
+    AC_DICT    = QK_AUTOCORRECT_DICT_CYCLE,
 };
 
 // Range Helpers
@@ -1419,7 +1419,7 @@ enum qk_keycode_defines {
 #define IS_MACRO_KEYCODE(code) ((code) >= QK_MACRO_0 && (code) <= QK_MACRO_31)
 #define IS_BACKLIGHT_KEYCODE(code) ((code) >= QK_BACKLIGHT_ON && (code) <= QK_BACKLIGHT_TOGGLE_BREATHING)
 #define IS_RGB_KEYCODE(code) ((code) >= RGB_TOG && (code) <= RGB_MODE_TWINKLE)
-#define IS_QUANTUM_KEYCODE(code) ((code) >= QK_BOOTLOADER && (code) <= QK_AUTOCORRECT_BANK_TOGGLE)
+#define IS_QUANTUM_KEYCODE(code) ((code) >= QK_BOOTLOADER && (code) <= QK_AUTOCORRECT_DICT_CYCLE)
 #define IS_KB_KEYCODE(code) ((code) >= QK_KB_0 && (code) <= QK_KB_31)
 #define IS_USER_KEYCODE(code) ((code) >= QK_USER_0 && (code) <= QK_USER_31)
 
@@ -1441,6 +1441,6 @@ enum qk_keycode_defines {
 #define MACRO_KEYCODE_RANGE                 QK_MACRO_0 ... QK_MACRO_31
 #define BACKLIGHT_KEYCODE_RANGE             QK_BACKLIGHT_ON ... QK_BACKLIGHT_TOGGLE_BREATHING
 #define RGB_KEYCODE_RANGE                   RGB_TOG ... RGB_MODE_TWINKLE
-#define QUANTUM_KEYCODE_RANGE               QK_BOOTLOADER ... QK_AUTOCORRECT_BANK_TOGGLE
+#define QUANTUM_KEYCODE_RANGE               QK_BOOTLOADER ... QK_AUTOCORRECT_DICT_CYCLE
 #define KB_KEYCODE_RANGE                    QK_KB_0 ... QK_KB_31
 #define USER_KEYCODE_RANGE                  QK_USER_0 ... QK_USER_31

quantum/process_keycode/autocorrect_data_default.h

@@ -72,10 +72,15 @@
 //   udpate     -> update
 //   widht      -> width
 
-#define AUTOCORRECT_MIN_LENGTH 5  // ":ture"
-#define AUTOCORRECT_MAX_LENGTH 10 // "accomodate"
+#define N_DICTS 1
+
+static const uint16_t autocorrect_offsets[N_DICTS] PROGMEM     = {0};
+static const uint16_t autocorrect_min_lengths[N_DICTS] PROGMEM = {5};
+static const uint16_t autocorrect_max_lengths[N_DICTS] PROGMEM = {10};
+static const uint16_t autocorrect_sizes[N_DICTS] PROGMEM       = {1104};
 
 #define DICTIONARY_SIZE 1104
+#define TYPO_BUFFER_SIZE 10
 
 static const uint8_t autocorrect_data[DICTIONARY_SIZE] PROGMEM = {108, 43,  0,   6,   71, 0,  7,   81, 0,   8,   199, 0,   9,   240, 1,  10,  250, 1,  11,  26,  2,   17,  53,  2,   18, 190, 2,   19,  202, 2,   21,  212, 2,   22,  20,  3,   23,  67,  3,   28,  16,  4,   0,  72,  50,  0,   22,  60,  0,   0,   11,  23,  44, 8,   11, 23,  44,  0,   132, 0,   8,   22,  18,  18,  15,  0,  132, 115, 101, 115, 0,   11,  23,  12,  26,  22,  0,   129, 99,  104, 0,   68,  94,  0,   8,   106, 0,   15, 174, 0,   21, 187, 0,   0,   12,  15,  25,  17,  12,  0,   131, 97,  108, 105, 100, 0,   74,  119, 0,   12,  129, 0,   21,  140, 0,   24,  165, 0,   0,   17,  12,  22,  0,   131, 103, 110, 101, 100, 0,   25,  21, 8,   7,   0,   131, 105, 118, 101, 100, 0,   72,  147, 0,  24,  156, 0,  0,   9,   8,   21,  0,   129, 114, 101, 100, 0,   6,   6,   18,  0,   129, 114, 101, 100, 0,   15,  6,   17,  12,  0,   129, 100, 101, 0,   18, 22,  8,   21,  11,  23,  0,   130, 104, 111,
                                                                   108, 100, 0,   4,   26, 18, 9,   0,  131, 114, 119, 97,  114, 100, 0,  68,  233, 0,  6,   246, 0,   7,   4,   1,   8,  16,  1,   10,  52,  1,   15,  81,  1,   21,  90,  1,   22,  117, 1,   23,  144, 1,   24, 215, 1,   25,  228, 1,   0,   6,   19,  22,  8,  16,  4,  17,  0,   130, 97,  99,  101, 0,   19,  4,   22,  8,  16,  4,   17,  0,   131, 112, 97,  99,  101, 0,   12,  21,  8,   25,  18,  0,   130, 114, 105, 100, 101, 0,  23,  0,   68, 25,  1,   17,  36,  1,   0,   21,  4,   24,  10,  0,   130, 110, 116, 101, 101, 0,   4,   21,  24,  4,   10,  0,   135, 117, 97,  114, 97,  110, 116, 101, 101, 0,   68,  59,  1,   7,   69,  1,   0,  24,  10,  44,  0,   131, 97,  117, 103, 101, 0,   8,   15, 12,  25,  12, 21,  19,  0,   130, 103, 101, 0,   22,  4,   9,   0,   130, 108, 115, 101, 0,   76,  97,  1,   24,  109, 1,   0,   24,  20,  4,   0,   132, 99, 113, 117, 105, 114, 101, 0,   23,  44,  0,