Mercurial > repos > bimib > cobraxy

comparison COBRAxy/utils/reaction_parsing.py @ 4:41f35c2f0c7b draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author luca_milaz
date Wed, 18 Sep 2024 10:59:10 +0000
parents
children
comparison
equal deleted inserted replaced
3:1f3ac6fd9867 4:41f35c2f0c7b
1 from enum import Enum
2 import utils.general_utils as utils
3 from typing import Dict
4 import csv
5 import re
6
7 # Reaction direction encoding:
8 class ReactionDir(Enum):
9 """
10 A reaction can go forwards, backwards or be reversible (able to proceed in both directions).
11 Models created / managed with cobrapy encode this information within the reaction's
12 formula using the arrows this enum keeps as values.
13 """
14 FORWARD = "-->"
15 BACKWARD = "<--"
16 REVERSIBLE = "<=>"
17
18 @classmethod
19 def fromReaction(cls, reaction :str) -> 'ReactionDir':
20 """
21 Takes a whole reaction formula string and looks for one of the arrows, returning the
22 corresponding reaction direction.
23
24 Args:
25 reaction : the reaction's formula.
26
27 Raises:
28 ValueError : if no valid arrow is found.
29
30 Returns:
31 ReactionDir : the corresponding reaction direction.
32 """
33 for member in cls:
34 if member.value in reaction: return member
35
36 raise ValueError("No valid arrow found within reaction string.")
37
38 ReactionsDict = Dict[str, Dict[str, float]]
39
40
41 def add_custom_reaction(reactionsDict :ReactionsDict, rId :str, reaction :str) -> None:
42 """
43 Adds an entry to the given reactionsDict. Each entry consists of a given unique reaction id
44 (key) and a :dict (value) matching each substrate in the reaction to its stoichiometric coefficient.
45 Keys and values are both obtained from the reaction's formula: if a substrate (custom metabolite id)
46 appears without an explicit coeff, the value 1.0 will be used instead.
47
48 Args:
49 reactionsDict : dictionary encoding custom reactions information.
50 rId : unique reaction id.
51 reaction : the reaction's formula.
52
53 Returns:
54 None
55
56 Side effects:
57 reactionsDict : mut
58 """
59 reaction = reaction.strip()
60 if not reaction: return
61
62 reactionsDict[rId] = {}
63 # We assume the '+' separating consecutive metabs in a reaction is spaced from them,
64 # to avoid confusing it for electrical charge:
65 for word in reaction.split(" + "):
66 metabId, stoichCoeff = word, 1.0
67 # Implicit stoichiometric coeff is equal to 1, some coeffs are floats.
68
69 # Accepted coeffs can be integer or floats with a dot (.) decimal separator
70 # and must be separated from the metab with a space:
71 foundCoeff = re.search(r"\d+(\.\d+)? ", word)
72 if foundCoeff:
73 wholeMatch = foundCoeff.group(0)
74 metabId = word[len(wholeMatch):].strip()
75 stoichCoeff = float(wholeMatch.strip())
76
77 reactionsDict[rId][metabId] = stoichCoeff
78
79 if not reactionsDict[rId]: del reactionsDict[rId] # Empty reactions are removed.
80
81
82 def create_reaction_dict(unparsed_reactions: Dict[str, str]) -> ReactionsDict:
83 """
84 Parses the given dictionary into the correct format.
85
86 Args:
87 unparsed_reactions (Dict[str, str]): A dictionary where keys are reaction IDs and values are unparsed reaction strings.
88
89 Returns:
90 ReactionsDict: The correctly parsed dict.
91 """
92 reactionsDict :ReactionsDict = {}
93 for rId, reaction in unparsed_reactions.items():
94 reactionDir = ReactionDir.fromReaction(reaction)
95 left, right = reaction.split(f" {reactionDir.value} ")
96
97 # Reversible reactions are split into distinct reactions, one for each direction.
98 # In general we only care about substrates, the product information is lost.
99 reactionIsReversible = reactionDir is ReactionDir.REVERSIBLE
100 if reactionDir is not ReactionDir.BACKWARD:
101 add_custom_reaction(reactionsDict, rId + "_F" * reactionIsReversible, left)
102
103 if reactionDir is not ReactionDir.FORWARD:
104 add_custom_reaction(reactionsDict, rId + "_B" * reactionIsReversible, right)
105
106 # ^^^ to further clarify: if a reaction is NOT reversible it will not be marked as _F or _B
107 # and whichever direction we DO keep (forward if --> and backward if <--) loses this information.
108 # This IS a small problem when coloring the map in marea.py because the arrow IDs in the map follow
109 # through with a similar convention on ALL reactions and correctly encode direction based on their
110 # model of origin. TODO: a proposed solution is to unify the standard in RPS to fully mimic the maps,
111 # which involves re-writing the "reactions" dictionary.
112
113 return reactionsDict
114
115
116 def parse_custom_reactions(customReactionsPath :str) -> ReactionsDict:
117 """
118 Creates a custom dictionary encoding reactions information from a csv file containing
119 data about these reactions, the path of which is given as input.
120
121 Args:
122 customReactionsPath : path to the reactions information file.
123
124 Returns:
125 ReactionsDict : dictionary encoding custom reactions information.
126 """
127 reactionsData :Dict[str, str] = {row[0]: row[1] for row in utils.readCsv(utils.FilePath.fromStrPath(customReactionsPath))}
128
129 return create_reaction_dict(reactionsData)
130