245 lines
8.2 KiB
Python
245 lines
8.2 KiB
Python
import re
|
|
from typing import List, Tuple, Dict, Literal, Type
|
|
|
|
from fastapi import HTTPException
|
|
from tortoise import Tortoise, Model
|
|
from tortoise.expressions import Q
|
|
|
|
from app.api.search_dict.search_schemas import SearchRequest, ProverbSearchRequest
|
|
from app.models import WordlistFr, WordlistJp, KangjiMapping
|
|
from app.utils.all_kana import all_in_kana
|
|
from app.utils.textnorm import normalize_text
|
|
from settings import TORTOISE_ORM
|
|
|
|
|
|
async def detect_language(text: str) -> Tuple[str, str, bool]:
|
|
"""
|
|
自动检测输入语言:
|
|
- zh: 简体中文
|
|
- jp: 日语（含假名或旧字体）
|
|
- fr: 拉丁字母（法语等）
|
|
- other: 其他
|
|
|
|
返回:
|
|
(映射或原文本, 语言代码, 是否为“含汉字且命中映射表”的情况)
|
|
"""
|
|
JAPANESE_HIRAGANA = r"[\u3040-\u309F]"
|
|
JAPANESE_KATAKANA = r"[\u30A0-\u30FF\u31F0-\u31FF]"
|
|
|
|
text = text.strip()
|
|
if not text:
|
|
return "", "other", False
|
|
|
|
# ✅ Step 1: 全部假名（无汉字）
|
|
if re.fullmatch(f"(?:{JAPANESE_HIRAGANA}|{JAPANESE_KATAKANA})+", text):
|
|
return text, "jp", False
|
|
|
|
# ✅ Step 2: 汉字检测
|
|
if re.search(r"[\u4e00-\u9fff]", text):
|
|
# 优先判断是否为日语汉字
|
|
jp_match = await KangjiMapping.get_or_none(kangji=text).only("kangji")
|
|
if jp_match:
|
|
return text, "jp", True # 含汉字且命中日语列
|
|
|
|
# 再检查是否为中文汉字
|
|
zh_match = await KangjiMapping.get_or_none(hanzi=text).only("hanzi", "kangji")
|
|
if zh_match:
|
|
return zh_match.kangji, "zh", True # 含汉字且命中中文列
|
|
|
|
# 若都不在映射表中，则为未映射的中文
|
|
return text, "zh", False
|
|
|
|
# ✅ Step 3: 拉丁字母检测（如法语）
|
|
if re.search(r"[a-zA-ZÀ-ÿ]", text):
|
|
return text, "fr", False
|
|
|
|
# ✅ Step 4: 其他情况（符号、空格等）
|
|
return text, "other", False
|
|
|
|
|
|
async def accurate_idiom_proverb(search_id: int, model: Type[Model], only_fields: List[str] = None):
|
|
if "freq" not in only_fields:
|
|
only_fields.append("freq")
|
|
result = await model.get_or_none(id=search_id).only(*only_fields)
|
|
if not result:
|
|
raise HTTPException(status_code=404, detail="Target not found")
|
|
result.freq = result.freq + 1
|
|
await result.save(update_fields=["freq"])
|
|
return result
|
|
|
|
|
|
async def suggest_proverb(
|
|
query: str,
|
|
lang: Literal["fr", "zh", "jp"],
|
|
model: Type[Model],
|
|
search_field: str = "search_text",
|
|
target_field: str = "text",
|
|
chi_exp_field: str = "chi_exp",
|
|
limit: int = 10,
|
|
) -> List[Dict[str, str]]:
|
|
keyword = query.strip()
|
|
if not keyword:
|
|
return []
|
|
|
|
# ✅ 搜索条件：中文时双字段联合匹配
|
|
if lang == "zh":
|
|
start_condition = Q(**{f"{chi_exp_field}__istartswith": keyword}) | Q(
|
|
**{f"{search_field}__istartswith": keyword})
|
|
contain_condition = Q(**{f"{chi_exp_field}__icontains": keyword}) | Q(**{f"{search_field}__icontains": keyword})
|
|
else:
|
|
start_condition = Q(**{f"{search_field}__istartswith": keyword})
|
|
contain_condition = Q(**{f"{search_field}__icontains": keyword})
|
|
|
|
# ✅ 1. 开头匹配
|
|
start_matches = await (
|
|
model.filter(start_condition)
|
|
.order_by("-freq", "id")
|
|
.limit(limit)
|
|
.values("id", target_field, chi_exp_field, "search_text")
|
|
)
|
|
|
|
# ✅ 2. 包含匹配（但不是开头）
|
|
contain_matches = await (
|
|
model.filter(contain_condition & ~start_condition)
|
|
.order_by("-freq", "id")
|
|
.limit(limit)
|
|
.values("id", target_field, chi_exp_field, "search_text")
|
|
)
|
|
|
|
# ✅ 3. 合并去重保持顺序
|
|
results = []
|
|
seen_ids = set()
|
|
for row in start_matches + contain_matches:
|
|
if row["id"] not in seen_ids:
|
|
seen_ids.add(row["id"])
|
|
results.append({
|
|
"id": row["id"],
|
|
"proverb": row[target_field],
|
|
"search_text": row["search_text"],
|
|
"chi_exp": row[chi_exp_field],
|
|
})
|
|
|
|
# ✅ 截断最终返回数量
|
|
return results[:limit]
|
|
|
|
|
|
async def suggest_autocomplete(query: SearchRequest, limit: int = 10):
|
|
"""
|
|
|
|
:param query: 当前用户输入的内容
|
|
:param limit: 返回列表限制长度
|
|
:return: 联想的单词列表（非完整信息，单纯单词）
|
|
"""
|
|
if query.language == 'fr':
|
|
query_word = normalize_text(query.query)
|
|
exact = await (
|
|
WordlistFr
|
|
.get_or_none(search_text=query.query)
|
|
.values("text", "freq")
|
|
)
|
|
if exact:
|
|
exact_word = [(exact.get("text"), exact.get("freq"))]
|
|
else:
|
|
exact_word = []
|
|
|
|
qs_prefix = (
|
|
WordlistFr
|
|
.filter(Q(search_text__startswith=query_word) | Q(text__startswith=query.query))
|
|
.exclude(search_text=query.query)
|
|
.only("text", "freq")
|
|
)
|
|
prefix_objs = await qs_prefix[:limit]
|
|
prefix: List[Tuple[str, int]] = [(o.text, o.freq) for o in prefix_objs]
|
|
|
|
need = max(0, limit - len(prefix))
|
|
contains: List[Tuple[str, int]] = []
|
|
|
|
if need > 0:
|
|
qs_contain = (
|
|
WordlistFr
|
|
.filter(Q(search_text__icontains=query_word) | Q(text__icontains=query.query))
|
|
.exclude(Q(search_text__startswith=query_word) | Q(text__startswith=query.query) | Q(text=query.query))
|
|
.only("text", "freq")
|
|
.only("text", "freq")
|
|
)
|
|
contains_objs = await qs_contain[: need * 2]
|
|
contains = [(o.text, o.freq) for o in contains_objs]
|
|
|
|
seen_text, out = set(), []
|
|
for text, freq in list(exact_word) + list(prefix) + list(contains):
|
|
key = text
|
|
if key not in seen_text:
|
|
seen_text.add(key)
|
|
out.append((text, freq))
|
|
if len(out) >= limit:
|
|
break
|
|
out = sorted(out, key=lambda w: (-w[2], len(w[0]), w[0]))
|
|
return [text for text, _ in out]
|
|
|
|
else:
|
|
query_word = all_in_kana(query.query)
|
|
exact = await (
|
|
WordlistJp
|
|
.get_or_none(
|
|
text=query.query
|
|
)
|
|
.only("text", "hiragana", "freq")
|
|
)
|
|
if exact:
|
|
exact_word = [(exact.text, exact.hiragana, exact.freq)]
|
|
else:
|
|
exact_word = []
|
|
|
|
qs_prefix = (
|
|
WordlistJp
|
|
.filter(Q(hiragana__startswith=query_word) | Q(text__startswith=query.query))
|
|
.exclude(text=query.query)
|
|
.only("text", "hiragana", "freq")
|
|
)
|
|
prefix_objs = await qs_prefix[:limit]
|
|
prefix: List[Tuple[str, str, int]] = [(o.text, o.hiragana, o.freq) for o in prefix_objs]
|
|
|
|
need = max(0, limit - len(prefix))
|
|
contains: List[Tuple[str, str, int]] = []
|
|
|
|
if need > 0:
|
|
qs_contain = await (
|
|
WordlistJp
|
|
.filter(Q(hiragana__icontains=query_word) | Q(text__icontains=query.query))
|
|
.exclude(Q(hiragana__startswith=query_word) | Q(text__startswith=query.query) | Q(text=query.query))
|
|
.only("text", "hiragana", "freq")
|
|
)
|
|
contains_objs = qs_contain[:need * 2]
|
|
contains: List[Tuple[str, str, int]] = [(o.text, o.hiragana, o.freq) for o in contains_objs]
|
|
|
|
seen_text, out = set(), []
|
|
for text, hiragana, freq in list(exact_word) + list(prefix) + list(contains):
|
|
key = (text, hiragana)
|
|
if key not in seen_text:
|
|
seen_text.add(key)
|
|
out.append((text, hiragana, freq))
|
|
if len(out) >= limit:
|
|
break
|
|
out = sorted(out, key=lambda w: (-w[2], len(w[0]), w[0]))
|
|
return [(text, hiragana) for text, hiragana, _ in out]
|
|
|
|
|
|
async def __test():
|
|
query_word: str = '棋逢'
|
|
return await (
|
|
suggest_proverb(
|
|
query=ProverbSearchRequest(query=query_word),
|
|
lang='zh'
|
|
)
|
|
)
|
|
|
|
|
|
async def __main():
|
|
await Tortoise.init(config=TORTOISE_ORM)
|
|
print(await __test())
|
|
|
|
|
|
if __name__ == '__main__':
|
|
# asyncio.run(__main())
|
|
print(detect_language(text="ahsjdasd"))
|