X-Git-Url: https://wannabe.guru.org/gitweb/?a=blobdiff_plain;f=logical_search.py;h=c85b2628a0a25a7be6b013ec963e997799bddfc5;hb=cce8d58af187c0a7fb7585eab5bda9fed731b719;hp=b55e68901501ad2e68a7f45df034866072c3f61b;hpb=e8fbbb7306430478dec55d2c963eed116d8330cc;p=python_utils.git diff --git a/logical_search.py b/logical_search.py index b55e689..c85b262 100644 --- a/logical_search.py +++ b/logical_search.py @@ -9,7 +9,8 @@ from __future__ import annotations import enum import sys from collections import defaultdict -from typing import Any, Dict, List, NamedTuple, Optional, Sequence, Set, Tuple, Union +from dataclasses import dataclass, field +from typing import Any, Dict, List, Optional, Sequence, Set, Tuple, Union class ParseError(Exception): @@ -20,13 +21,22 @@ class ParseError(Exception): self.message = message -class Document(NamedTuple): - """A tuple representing a searchable document.""" +@dataclass +class Document: + """A class representing a searchable document.""" - docid: str # a unique idenfier for the document - tags: Set[str] # an optional set of tags - properties: List[Tuple[str, str]] # an optional set of key->value properties - reference: Any # an optional reference to something else + # A unique identifier for each document. + docid: str = '' + + # A set of tag strings for this document. May be empty. + tags: Set[str] = field(default_factory=set) + + # A list of key->value strings for this document. May be empty. + properties: List[Tuple[str, str]] = field(default_factory=list) + + # An optional reference to something else; interpreted only by + # caller code, ignored here. + reference: Optional[Any] = None class Operation(enum.Enum): @@ -173,7 +183,7 @@ class Corpus(object): def invert_docid_set(self, original: Set[str]) -> Set[str]: """Invert a set of docids.""" - return set([docid for docid in self.documents_by_docid.keys() if docid not in original]) + return {docid for docid in self.documents_by_docid if docid not in original} def get_doc(self, docid: str) -> Optional[Document]: """Given a docid, retrieve the previously added Document."""