fanlusky
fae80660c1
在完整音节后敲反引号，接着的字母筛这个音节上的字：第一码大小写均可， 紧跟的大写字母是第二码，之后可以接着打，一个输入里可以有多段。开关 helpcode.shuangpin_mid_sentence_helpcode 默认关闭。 约束在整句解码时生效，而不是解完再筛： - Google 解码器在 add_char_qwerty 里送 extend_mtrx_nd 之前按约束过滤词条， 锁住的字带着上下文重新整句解码；约束变化时整段重解，不污染增量搜索。 - kenlm 词格建图时按约束筛边，受约束的跨度不截断地查词库；神经重排只在 词格 n-best 里挑，自动满足约束。 - 普通词条、云和 AI 候选按同一规则收尾校验；满足不了就不出整句。 反引号段在请求里换成手动分隔符，原串留给宿主回读与预编辑显示；选词推进 时剩余音节上的约束随原串保留。输入形状规则放在 contracts/mid_sentence_helpcode.h， TSF、Server、引擎共用，新增 Worker opcode MidSentenceHelpcodeChanged = 28。 googlepinyinime-rev 已与上游解绑，AGENTS 与 UPSTREAM.md 同步说明。
265 lines
9.8 KiB
C++
265 lines
9.8 KiB
C++
#include "input_session.h"
|
|
#include "../shuangpin/shuangpin_query.h"
|
|
#include "../shuangpin/shuangpin_utils.h"
|
|
#include <algorithm>
|
|
#include <cctype>
|
|
|
|
namespace metasequoia
|
|
{
|
|
std::string InputSession::editing_text() const
|
|
{
|
|
if (dedicated_english_mode_)
|
|
return dedicated_english_preedit_;
|
|
if (local_input_mode_ == LocalInputMode::TemporaryJapanese)
|
|
return "R" + engine_.get_request().raw_input_with_cases;
|
|
if (local_input_mode_ != LocalInputMode::None)
|
|
return local_preedit_;
|
|
return get_pinyin_sequence_with_cases();
|
|
}
|
|
|
|
std::size_t InputSession::caret_position() const
|
|
{
|
|
const auto size = editing_text().size();
|
|
return std::min(caret_.value_or(size), size);
|
|
}
|
|
|
|
namespace
|
|
{
|
|
// Unit boundaries of a quanpin spelling in raw coordinates. The visible preedit
|
|
// (BuildQuanpinAutocorrectDisplay) is always rebuilt from the raw letters -- the
|
|
// autocorrect cut and the alias layer only move or add separators -- so a
|
|
// separator in the display marks where the next raw unit starts. A display that
|
|
// cannot explain the raw letters yields no boundaries and the caller falls back
|
|
// to single-character editing.
|
|
std::vector<std::size_t> QuanpinRawBoundaries(const std::string &raw, const std::string &display)
|
|
{
|
|
std::vector<std::size_t> boundaries;
|
|
std::vector<std::size_t> raw_letter_offsets;
|
|
raw_letter_offsets.reserve(raw.size());
|
|
for (std::size_t index = 0; index < raw.size(); ++index)
|
|
{
|
|
if (raw[index] != '\'')
|
|
{
|
|
raw_letter_offsets.push_back(index);
|
|
}
|
|
}
|
|
if (raw_letter_offsets.empty())
|
|
{
|
|
return boundaries;
|
|
}
|
|
|
|
boundaries.push_back(0);
|
|
std::size_t letters_seen = 0;
|
|
for (const char ch : display)
|
|
{
|
|
if (ch != '\'')
|
|
{
|
|
++letters_seen;
|
|
continue;
|
|
}
|
|
// The next unit starts at the raw offset of the next letter; a
|
|
// trailing separator has no next letter and starts no unit.
|
|
if (letters_seen < raw_letter_offsets.size())
|
|
{
|
|
boundaries.push_back(raw_letter_offsets[letters_seen]);
|
|
}
|
|
}
|
|
if (letters_seen != raw_letter_offsets.size())
|
|
{
|
|
// The display no longer corresponds letter-for-letter: refuse to map
|
|
// rather than delete an arbitrary span.
|
|
return {};
|
|
}
|
|
boundaries.push_back(raw.size());
|
|
boundaries.erase(std::unique(boundaries.begin(), boundaries.end()), boundaries.end());
|
|
return boundaries;
|
|
}
|
|
} // namespace
|
|
|
|
std::vector<std::size_t> InputSession::segment_raw_boundaries() const
|
|
{
|
|
// Local modes and the dedicated English scheme spell words, not syllables:
|
|
// the host treats these keys exactly like a plain Backspace (PRD R4).
|
|
if (dedicated_english_mode_ || local_input_mode_ != LocalInputMode::None)
|
|
{
|
|
return {};
|
|
}
|
|
|
|
const std::string raw_with_cases = get_pinyin_sequence_with_cases();
|
|
const std::string raw = get_pinyin_sequence();
|
|
if (raw_with_cases.empty())
|
|
{
|
|
return {};
|
|
}
|
|
|
|
if (current_scheme_type() == SchemeType::Shuangpin)
|
|
{
|
|
// 带句中辅助码的串没有逐音节的单元模型：按段删除、按段跳光标和光标前缀重算都退回逐字符。
|
|
if (shuangpin::has_mid_sentence_helpcode(raw_with_cases))
|
|
{
|
|
return {};
|
|
}
|
|
const std::size_t helpcode_length =
|
|
shuangpin::detect_active_double_helpcode_length(raw, raw_with_cases, shuangpin_profile_);
|
|
const std::string base =
|
|
helpcode_length > 0 ? shuangpin::trim_trailing_letters_preserve_delimiters(raw_with_cases, helpcode_length)
|
|
: raw_with_cases;
|
|
std::vector<std::size_t> boundaries = shuangpin::segment_raw_boundaries(base, shuangpin_profile_);
|
|
if (helpcode_length > 0 && !boundaries.empty())
|
|
{
|
|
// The active double helpcode is one editable unit of its own, the
|
|
// same boundary raw_segmentation draws before it.
|
|
boundaries.push_back(base.size());
|
|
boundaries.push_back(raw_with_cases.size());
|
|
std::sort(boundaries.begin(), boundaries.end());
|
|
boundaries.erase(std::unique(boundaries.begin(), boundaries.end()), boundaries.end());
|
|
}
|
|
return boundaries;
|
|
}
|
|
|
|
if (current_scheme_type() != SchemeType::Quanpin)
|
|
{
|
|
return {};
|
|
}
|
|
return QuanpinRawBoundaries(raw_with_cases, get_pinyin_segmentation_with_cases());
|
|
}
|
|
|
|
KeyResult InputSession::edit_at_caret(Command command)
|
|
{
|
|
auto text = editing_text();
|
|
auto caret = caret_position();
|
|
// Local-mode markers are commands, not editable payload. Backspace at the
|
|
// end of a marker-only composition retains the existing cancel behavior.
|
|
const std::size_t begin = local_input_mode_ == LocalInputMode::None ? 0 : 1;
|
|
switch (command)
|
|
{
|
|
case Command::MoveLeft:
|
|
caret = caret > begin ? caret - 1 : begin;
|
|
break;
|
|
case Command::MoveRight:
|
|
caret = std::min(caret + 1, text.size());
|
|
break;
|
|
case Command::MoveHome:
|
|
caret = begin;
|
|
break;
|
|
case Command::MoveEnd:
|
|
caret = text.size();
|
|
break;
|
|
case Command::Backspace:
|
|
if (caret <= begin)
|
|
return {true, std::nullopt, std::nullopt};
|
|
text.erase(--caret, 1);
|
|
return replace_editing_text(std::move(text), caret);
|
|
case Command::DeleteForward:
|
|
if (caret == text.size())
|
|
return {true, std::nullopt, std::nullopt};
|
|
text.erase(caret, 1);
|
|
return replace_editing_text(std::move(text), caret);
|
|
default:
|
|
return {};
|
|
}
|
|
caret_ = caret;
|
|
// Pure moves mutate the caret too, so they must re-evaluate the prefix boundary exactly
|
|
// like an insert/delete does; otherwise candidates() would keep serving the previous
|
|
// boundary's list until the next edit key (PRD R2: the boundary follows the caret).
|
|
if (!dedicated_english_mode_ && local_input_mode_ == LocalInputMode::None)
|
|
{
|
|
update_mixed_candidates();
|
|
}
|
|
return {true, std::nullopt, std::nullopt};
|
|
}
|
|
|
|
KeyResult InputSession::insert_at_caret(char character)
|
|
{
|
|
auto text = editing_text();
|
|
const auto caret = caret_position();
|
|
const bool lower = character >= 'a' && character <= 'z';
|
|
const bool upper = character >= 'A' && character <= 'Z';
|
|
bool accepted = lower || upper;
|
|
if (!dedicated_english_mode_)
|
|
{
|
|
switch (local_input_mode_)
|
|
{
|
|
case LocalInputMode::Unicode:
|
|
accepted = (character >= '0' && character <= '9') || (character >= 'a' && character <= 'f') ||
|
|
(character >= 'A' && character <= 'F') ||
|
|
(character == '+' && caret == 1 && text.find('+') == std::string::npos);
|
|
break;
|
|
case LocalInputMode::QuickPhrase:
|
|
accepted = lower;
|
|
break;
|
|
case LocalInputMode::DateTime:
|
|
accepted = false;
|
|
break;
|
|
case LocalInputMode::None:
|
|
accepted = lower || (upper && ((scheme() == SchemeType::Quanpin && quanpin_helpcode_enabled_) ||
|
|
(scheme() == SchemeType::Shuangpin && shuangpin_helpcode_enabled_)));
|
|
if (character == ';' && scheme() == SchemeType::Shuangpin &&
|
|
ShuangpinProfileUsesSemicolonFinal(shuangpin_profile_))
|
|
{
|
|
const auto separator = caret == 0 ? std::string::npos : text.rfind('\'', caret - 1);
|
|
const auto start = separator == std::string::npos ? 0 : separator + 1;
|
|
accepted = (caret - start) % 2 == 1;
|
|
}
|
|
if (character == '\'' && scheme() != SchemeType::Wubi)
|
|
accepted = caret > 0;
|
|
break;
|
|
case LocalInputMode::Emoji:
|
|
case LocalInputMode::Kaomoji:
|
|
case LocalInputMode::TemporaryJapanese:
|
|
accepted = accepted || character == '\'';
|
|
break;
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
if (!accepted)
|
|
return {};
|
|
if (character == '\'' && ((caret > 0 && text[caret - 1] == '\'') || text[caret] == '\''))
|
|
return {true, std::nullopt, std::nullopt};
|
|
text.insert(caret, 1, character);
|
|
return replace_editing_text(std::move(text), caret + 1);
|
|
}
|
|
|
|
KeyResult InputSession::replace_editing_text(std::string text, std::size_t caret)
|
|
{
|
|
std::optional<std::string> diagnostic;
|
|
if (dedicated_english_mode_)
|
|
{
|
|
dedicated_english_preedit_ = std::move(text);
|
|
update_dedicated_english_candidates();
|
|
}
|
|
else if (local_input_mode_ != LocalInputMode::None && local_input_mode_ != LocalInputMode::TemporaryJapanese)
|
|
{
|
|
local_preedit_ = std::move(text);
|
|
diagnostic = update_local_candidates();
|
|
}
|
|
else
|
|
{
|
|
const auto payload = local_input_mode_ == LocalInputMode::TemporaryJapanese ? text.substr(1) : text;
|
|
auto normalized = payload;
|
|
std::transform(normalized.begin(), normalized.end(), normalized.begin(),
|
|
[](unsigned char value) { return static_cast<char>(std::tolower(value)); });
|
|
set_pinyin_sequence(normalized);
|
|
set_pinyin_sequence_with_cases(payload);
|
|
apply_pending_sequence();
|
|
if (local_input_mode_ == LocalInputMode::TemporaryJapanese)
|
|
{
|
|
local_preedit_ = "R" + engine_.get_preedit();
|
|
local_candidates_ = engine_.get_candidates();
|
|
}
|
|
}
|
|
caret_ = caret;
|
|
online_requests_.invalidate();
|
|
discard_abandoned_phrase_progress();
|
|
// apply_pending_sequence() already refreshed candidates while the caret was cleared; now
|
|
// that it is back in place, re-evaluate the caret prefix boundary (a caret at the end of
|
|
// the text deactivates the prefix decode, mid-text re-decodes it).
|
|
if (!dedicated_english_mode_ && local_input_mode_ == LocalInputMode::None)
|
|
{
|
|
update_mixed_candidates();
|
|
}
|
|
return {true, std::nullopt, std::move(diagnostic)};
|
|
}
|
|
} // namespace metasequoia
|