from __future__ import annotations
-from collections import defaultdict
import enum
import sys
-from typing import (
- Any,
- Dict,
- List,
- NamedTuple,
- Optional,
- Set,
- Sequence,
- Tuple,
- Union,
-)
+from collections import defaultdict
+from typing import Any, Dict, List, NamedTuple, Optional, Sequence, Set, Tuple, Union
class ParseError(Exception):
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
+ properties: List[Tuple[str, str]] # an optional set of key->value properties
reference: Any # an optional reference to something else
... reference=None,
... )
... )
+ >>> c.add_doc(Document(
+ ... docid=3,
+ ... tags=set(['urgent']),
+ ... properties=[
+ ... ('author', 'Scott'),
+ ... ('subject', 'car turning in front of you')
+ ... ],
+ ... reference=None,
+ ... )
+ ... )
>>> c.query('author:Scott and important')
{1}
"""
def __init__(self) -> None:
self.docids_by_tag: Dict[str, Set[str]] = defaultdict(set)
- self.docids_by_property: Dict[Tuple[str, str], Set[str]] = defaultdict(
- set
- )
+ self.docids_by_property: Dict[Tuple[str, str], Set[str]] = defaultdict(set)
self.docids_with_property: Dict[str, Set[str]] = defaultdict(set)
self.documents_by_docid: Dict[str, Document] = {}
"""Invert a set of docids."""
return set(
- [
- docid
- for docid in self.documents_by_docid.keys()
- if docid not in original
- ]
+ [docid for docid in self.documents_by_docid.keys() if docid not in original]
)
def get_doc(self, docid: str) -> Optional[Document]:
ok = True
break
if not ok:
- raise ParseError(
- "Unbalanced parenthesis in query expression"
- )
+ raise ParseError("Unbalanced parenthesis in query expression")
# and, or, not
else:
raise ParseError(f"Unexpected query {tag}")
elif self.op is Operation.DISJUNCTION:
if len(evaled_operands) != 2:
- raise ParseError(
- "Operation.DISJUNCTION (or) expects two operands."
- )
+ raise ParseError("Operation.DISJUNCTION (or) expects two operands.")
retval.update(evaled_operands[0])
retval.update(evaled_operands[1])
elif self.op is Operation.CONJUNCTION:
if len(evaled_operands) != 2:
- raise ParseError(
- "Operation.CONJUNCTION (and) expects two operands."
- )
+ raise ParseError("Operation.CONJUNCTION (and) expects two operands.")
retval.update(evaled_operands[0])
retval = retval.intersection(evaled_operands[1])
elif self.op is Operation.INVERSION:
if len(evaled_operands) != 1:
- raise ParseError(
- "Operation.INVERSION (not) expects one operand."
- )
+ raise ParseError("Operation.INVERSION (not) expects one operand.")
_ = evaled_operands[0]
if isinstance(_, set):
retval.update(self.corpus.invert_docid_set(_))
if __name__ == '__main__':
import doctest
+
doctest.testmod()