Files
MDHelper/mdhelper/search.py
T
2026-06-14 09:51:16 +08:00

183 lines
5.8 KiB
Python

from __future__ import annotations
from dataclasses import dataclass
from pathlib import Path
from .document_tree import CODE_DOCUMENT_EXTENSIONS, DocumentEntry
from .text_io import DEFAULT_SEARCH_INDEX_BYTES, is_within_size_limit, read_text_with_fallback
@dataclass(frozen=True)
class SearchMatch:
entry: DocumentEntry
snippet: str
match_count: int = 0
match_kind: str = "正文"
class SearchIndex:
def __init__(self, entries: list[DocumentEntry] | None = None) -> None:
self._records: list[tuple[DocumentEntry, str, str]] = []
self._body_cache: dict[Path, tuple[int, int, str]] = {}
if entries:
self.rebuild(entries)
def rebuild(self, entries: list[DocumentEntry]) -> None:
self._records = [(entry, entry.title, entry.relative_path) for entry in entries]
current_paths = {entry.path for entry in entries}
self._body_cache = {path: value for path, value in self._body_cache.items() if path in current_paths}
def matches_entry(self, entry: DocumentEntry, query: str) -> bool:
normalized = query.strip().casefold()
if not normalized:
return True
for indexed_entry, title, relative_path in self._records:
if indexed_entry.path == entry.path:
if normalized in self._haystack(title, relative_path, "").casefold():
return True
return normalized in self._body_for(indexed_entry).casefold()
if normalized in self._haystack(entry.title, entry.relative_path, "").casefold():
return True
return normalized in self._body_for(entry).casefold()
def search(self, query: str, limit: int = 80) -> list[SearchMatch]:
normalized = query.strip().casefold()
if not normalized:
return []
matches: list[SearchMatch] = []
for entry, title, relative_path in self._records:
result = _match_details(title, relative_path, "", normalized)
if result is None:
result = _match_details(title, relative_path, self._body_for(entry), normalized)
if result is None:
continue
snippet, match_count, match_kind = result
matches.append(
SearchMatch(
entry=entry,
snippet=snippet,
match_count=match_count,
match_kind=match_kind,
)
)
if len(matches) >= limit:
break
return matches
def _haystack(self, title: str, relative_path: str, body: str) -> str:
return f"{title}\n{relative_path}\n{body}"
def _body_for(self, entry: DocumentEntry) -> str:
signature = _file_signature(entry.path)
if signature is None:
return ""
cached = self._body_cache.get(entry.path)
if cached is not None and cached[:2] == signature:
return cached[2]
body = _read_text(entry)
self._body_cache[entry.path] = (*signature, body)
return body
def matches_entry(entry: DocumentEntry, query: str) -> bool:
normalized = query.strip().casefold()
if not normalized:
return True
haystack = f"{entry.title}\n{entry.relative_path}\n{_read_text(entry)}".casefold()
return normalized in haystack
def search_entries(entries: list[DocumentEntry], query: str, limit: int = 80) -> list[SearchMatch]:
normalized = query.strip().casefold()
if not normalized:
return []
matches: list[SearchMatch] = []
for entry in entries:
text = _read_text(entry)
result = _match_details(entry.title, entry.relative_path, text, normalized)
if result is None:
continue
snippet, match_count, match_kind = result
matches.append(
SearchMatch(
entry=entry,
snippet=snippet,
match_count=match_count,
match_kind=match_kind,
)
)
if len(matches) >= limit:
break
return matches
def _read_text(entry: DocumentEntry) -> str:
supported_text_extensions = {
".md",
".markdown",
".txt",
".html",
".htm",
".xhtml",
".css",
".js",
".json",
".yaml",
".yml",
} | CODE_DOCUMENT_EXTENSIONS
if entry.path.suffix.lower() not in supported_text_extensions:
return ""
if not is_within_size_limit(entry.path, DEFAULT_SEARCH_INDEX_BYTES):
return ""
text = read_text_with_fallback(entry.path, max_bytes=DEFAULT_SEARCH_INDEX_BYTES)
return "" if text.startswith("读取失败:") else text
def _file_signature(path: Path) -> tuple[int, int] | None:
try:
stat = path.stat()
except OSError:
return None
return stat.st_mtime_ns, stat.st_size
def _snippet(text: str, index: int, length: int) -> str:
start = max(0, index - 42)
end = min(len(text), index + length + 86)
snippet = text[start:end].replace("\r", " ").replace("\n", " ")
snippet = " ".join(snippet.split())
if start > 0:
snippet = "..." + snippet
if end < len(text):
snippet += "..."
return snippet
def _match_details(title: str, relative_path: str, body: str, normalized_query: str) -> tuple[str, int, str] | None:
fields = (
("标题", title),
("路径", relative_path),
("正文", body),
)
total = sum(_count_matches(text, normalized_query) for _, text in fields)
if total <= 0:
return None
for kind, text in fields:
index = text.casefold().find(normalized_query)
if index >= 0:
return _snippet(text, index, len(normalized_query)), total, kind
return None
def _count_matches(text: str, normalized_query: str) -> int:
if not normalized_query:
return 0
return text.casefold().count(normalized_query)