Mercurial > repos > bimib > cobraxy

comparison COBRAxy/src/utils/rule_parsing.py @ 539:2fb97466e404 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author francesco_lapi
date Sat, 25 Oct 2025 14:55:13 +0000
parents
children fcdbc81feb45
comparison
equal deleted inserted replaced
538:fd53d42348bd 539:2fb97466e404
1 """
2 Parsing utilities for gene rules (GPRs).
3
4 This module provides:
5 - RuleErr: structured errors for malformed rules
6 - RuleOp: valid logical operators (AND/OR)
7 - OpList: nested list structure representing parsed rules with explicit operator
8 - RuleStack: helper stack to build nested OpLists during parsing
9 - parseRuleToNestedList: main entry to parse a rule string into an OpList
10 """
11 from enum import Enum
12 import utils.general_utils as utils
13 from typing import List, Union, Optional
14
15 class RuleErr(utils.CustomErr):
16 """
17 Error type for rule syntax errors.
18 """
19 errName = "Rule Syntax Error"
20 def __init__(self, rule :str, msg = "no further details provided") -> None:
21 super().__init__(
22 f"rule \"{rule}\" is malformed, {msg}",
23 "please verify your input follows the validity guidelines")
24
25 class RuleOp(Enum):
26 """
27 Valid logical operators for gene rules.
28 """
29 OR = "or"
30 AND = "and"
31
32 @classmethod
33 def isOperator(cls, op :str) -> bool:
34 return op.upper() in cls.__members__
35
36 def __str__(self) -> str: return self.value
37
38 class OpList(List[Union[str, "OpList"]]):
39 """
40 Parsed rule structure: a list with an associated operator for that level.
41 """
42 def __init__(self, op :Optional[RuleOp] = None) -> None:
43 """
44 (Private) Initializes an instance of OpList.
45
46 Args:
47 op (str): Operator to be assigned to the OpList. Defaults to "".
48
49 Returns:
50 None : practically, an OpList instance.
51 """
52 self.op = op
53
54 def setOpIfMissing(self, op :RuleOp) -> None:
55 """
56 Sets the operator of the OpList if it's missing.
57
58 Args:
59 op (str): Operator to be assigned to the OpList.
60
61 Returns:
62 None
63 """
64 if not self.op: self.op = op
65
66 def __repr__(self, indent = "") -> str:
67 """
68 (Private) Returns a string representation of the current OpList instance.
69
70 Args:
71 indent (str): Indentation level . Defaults to "".
72
73 Returns:
74 str: A string representation of the current OpList instance.
75 """
76 nextIndent = indent + " "
77 return f"<{self.op}>[\n" + ",\n".join([
78 f"{nextIndent}{item.__repr__(nextIndent) if isinstance(item, OpList) else item}"
79 for item in self ]) + f"\n{indent}]"
80
81 class RuleStack:
82 """
83 FILO stack used during parsing to build nested OpLists; the top is the current level.
84 """
85 def __init__(self) -> None:
86 """
87 (Private) initializes an instance of RuleStack.
88
89 Returns:
90 None : practically, a RuleStack instance.
91 """
92 self.__stack = [OpList()] # the stack starts out with the result list already allocated
93 self.__updateCurrent()
94
95 def pop(self) -> None:
96 """
97 Removes the OpList on top of the stack, also flattening it once when possible.
98
99 Side Effects:
100 self : mut
101
102 Returns:
103 None
104 """
105 oldTop = self.__stack.pop()
106 if len(oldTop) == 1 and isinstance(oldTop[0], OpList): self.__stack[-1][-1] = oldTop[0]
107 self.__updateCurrent()
108
109 def push(self, operator = "") -> None:
110 """
111 Adds a new nesting level, in the form of a new OpList on top of the stack.
112
113 Args:
114 operator : the operator assigned to the new OpList.
115
116 Side Effects:
117 self : mut
118
119 Returns:
120 None
121 """
122 newLevel = OpList(operator)
123 self.current.append(newLevel)
124 self.__stack.append(newLevel)
125 self.__updateCurrent()
126
127 def popForward(self) -> None:
128 """
129 Moves the last "actual" item from the 2nd to last list to the beginning of the top list, as per
130 the example below:
131 stack : [list_a, list_b]
132 list_a : [item1, item2, list_b] --> [item1, list_b]
133 list_b : [item3, item4] --> [item2, item3, item4]
134
135 This is essentially a "give back as needed" operation.
136
137 Side Effects:
138 self : mut
139
140 Returns:
141 None
142 """
143 self.current.insert(0, self.__stack[-2].pop(-2))
144
145 def currentIsAnd(self) -> bool:
146 """
147 Checks if the current OpList's assigned operator is "and".
148
149 Returns:
150 bool : True if the current OpList's assigned operator is "and", False otherwise.
151 """
152 return self.current.op is RuleOp.AND
153
154 def obtain(self, err :Optional[utils.CustomErr] = None) -> Optional[OpList]:
155 """
156 Obtains the first OpList on the stack, only if it's the only element.
157
158 Args:
159 err : The error to raise if obtaining the result is not possible.
160
161 Side Effects:
162 self : mut
163
164 Raises:
165 err: If given, otherwise None is returned.
166
167 Returns:
168 Optional[OpList]: The first OpList on the stack, only if it's the only element.
169 """
170
171 if len(self.__stack) == 1: return self.__stack.pop()
172 if err: raise err
173 return None
174
175 def __updateCurrent(self) -> None:
176 """
177 (Private) Updates the current OpList to the one on top of the stack.
178
179 Side Effects:
180 self : mut
181
182 Returns:
183 None
184 """
185 self.current = self.__stack[-1]
186
187 def parseRuleToNestedList(rule :str) -> OpList:
188 """
189 Parse a rule string into an OpList, making operator precedence explicit via nesting.
190
191 Args:
192 rule: Rule string to parse (supports parentheses, 'and', 'or').
193
194 Raises:
195 RuleErr: If the rule is malformed (e.g., mismatched parentheses or misplaced operators).
196
197 Returns:
198 OpList: Parsed rule as an OpList structure.
199 """
200 source = iter(rule
201 .replace("(", "( ").replace(")", " )") # single out parentheses as words
202 .strip() # trim edges
203 .split()) # split by spaces
204
205 stack = RuleStack()
206 nestingErr = RuleErr(rule, "mismatch between open and closed parentheses")
207 try:
208 while True: # read until source ends
209 while True:
210 operand = next(source, None) # expect operand or '('
211 if operand is None: raise RuleErr(rule, "found trailing open parentheses")
212 if operand in ("and", "or", ")"): # unexpected operator position
213 raise RuleErr(rule, f"found \"{operand}\" in unexpected position")
214
215 if operand != "(": break # got a name
216
217 # found rule opening: add a new nesting level
218 stack.push()
219
220 stack.current.append(operand)
221
222 while True: # read until operator found or source ends
223 operator = next(source, None) # expect operator or ')'
224 if operator and operator != ")": break # got operator
225
226 if stack.currentIsAnd(): stack.pop() # close current AND chain
227
228 if not operator: break
229 stack.pop() # close parentheses
230
231 if not operator: break
232
233 if not RuleOp.isOperator(operator): raise RuleErr(
234 rule, f"found \"{operator}\" in unexpected position, expected operator")
235
236 operator = RuleOp(operator)
237 if operator is RuleOp.OR and stack.currentIsAnd():
238 stack.pop()
239
240 elif operator is RuleOp.AND and not stack.currentIsAnd():
241 stack.push(operator)
242 stack.popForward()
243
244 stack.current.setOpIfMissing(operator)
245
246 except RuleErr as err: raise err # bubble up proper errors
247 except: raise nestingErr # everything else is interpreted as a nesting error.
248
249 parsedRule = stack.obtain(nestingErr)
250 return parsedRule[0] if len(parsedRule) == 1 and isinstance(parsedRule[0], list) else parsedRule