From a0af3c1a17455409f50e10d7bebff9b47e7ab704 Mon Sep 17 00:00:00 2001
From: George Marchment <georgemarchment@yahoo.fr>
Date: Tue, 29 Apr 2025 18:11:29 +0200
Subject: [PATCH] Added support for multi ternary operator -> this will need to
be checked
---
src/code_.py | 124 ++++++++++++++++++++++++++++++++++++++----------
src/outils.py | 60 +++++++++++++++++++++++
src/workflow.py | 52 ++++++++++----------
3 files changed, 186 insertions(+), 50 deletions(-)
diff --git a/src/code_.py b/src/code_.py
index c59e14d..3c9fefd 100644
--- a/src/code_.py
+++ b/src/code_.py
@@ -1,4 +1,4 @@
-from .outils import remove_comments, get_parenthese_count, get_curly_count, get_code_until_parenthese_count, extract_curly, get_next_element_caracter
+from .outils import remove_comments, get_parenthese_count, get_curly_count, get_code_until_parenthese_count, extract_curly, get_next_element_caracter, get_code_until_character
from .bioflowinsighterror import BioFlowInsightError
import re
from . import constant
@@ -201,42 +201,114 @@ class Code:
def add_map_element(self, old, new):
self.origin.add_map_element(old, new)
+ ##This methods rewrite ternary operation into "normal" conditions
+ ##variable = (condition) ? Expression2 : Expression3;
+ #def rewrite_ternary_operation_to_normal_condition(self, code):
+ # pattern = r"(def)? *(\w+) *\= *([^?\n]+) *\? *([^:\n]+) *\: *([^\n]+)\n"
+ # to_replace = []
+ # checked = []
+ # for match in re.finditer(pattern, code):
+ # def_variable = ""
+ # if(match.group(1)!=None):
+ # def_variable = match.group(1)
+ #
+ #
+ # variable = match.group(2)
+ # condition = match.group(3).strip()
+ # exp1, exp2 = match.group(4).strip(), match.group(5).strip()
+ # old = match.group(0)
+ # #print(exp1)
+ # #print(exp2)
+ # #print()
+ # new = f"if ({condition}) {{\n{def_variable} {variable} = {exp1}\n}}\n"
+ # new += f"if (!({condition})) {{\n{def_variable} {variable} = {exp2}\n}}\n\n"
+ # print(new)
+ # #else {{\n{variable} = {exp2}\n}}\n"
+ # #Here we check that it's worked correctly -> that we have done a good parsing
+ # if(get_parenthese_count(condition)==0 and get_parenthese_count(exp1)==0 and get_parenthese_count(exp2)==0 and get_curly_count(condition)==0 and get_curly_count(exp1)==0 and get_curly_count(exp2)==0):
+ # to_replace.append((old, new))
+ # else:
+ # checked.append(match.group(0))
+ # for r in to_replace:
+ # old, new = r
+ # self.add_to_ternary_operation_dico(old, new)
+ # tmp = code
+ # code = code.replace(old, new, 1)
+ # if(old!=new and tmp==code):
+ # raise Exception("This shouldn't happen -> the code wasn't replaced")
+ # #Check if there is still a ternary operation in this case we cannot analyse it
+ # #Cause it is a complexe/multiple ternanry operation
+ # for match in re.finditer(pattern, code):
+ # #print(match.group(0))
+ # if(match.group(0) not in checked):
+ # raise BioFlowInsightError(f"Detected a multi ternary operation (a ternary operation in a ternary operation) in the file '{self.origin.get_file_address()}'. BioFlow-Insight does not support this, try defining it in a different way.", type="Multi ternary operation")
+ # return code
+
#This methods rewrite ternary operation into "normal" conditions
#variable = (condition) ? Expression2 : Expression3;
def rewrite_ternary_operation_to_normal_condition(self, code):
- pattern = r"(def)? *(\w+) *\= *([^?\n]+) *\? *([^:\n]+) *\: *([^\n]+)\n"
+ pattern = r"\n *(def)? *(\w+) *\= *(([^?\n]+) *\? *([^:\n]+) *\: *([^\n]+))\n"
to_replace = []
- checked = []
- for match in re.finditer(pattern, code):
- def_variable = ""
- if(match.group(1)!=None):
- def_variable = match.group(1)
-
+
+ searching = True
+
+ while(searching):
+ searching = False
+ for match in re.finditer(pattern, code):
+ def_variable = ""
+ if(match.group(1)!=None):
+ def_variable = match.group(1)
- variable = match.group(2)
- condition = match.group(3).strip()
- exp1, exp2 = match.group(4).strip(), match.group(5).strip()
- old = match.group(0)
- new = f"if ({condition}) {{\n{def_variable} {variable} = {exp1}\n}}\n"
- new += f"if (!({condition})) {{\n{def_variable} {variable} = {exp2}\n}}\n\n"
- #else {{\n{variable} = {exp2}\n}}\n"
- #Here we check that it's worked correctly -> that we have done a good parsing
- if(get_parenthese_count(condition)==0 and get_parenthese_count(exp1)==0 and get_parenthese_count(exp2)==0 and get_curly_count(condition)==0 and get_curly_count(exp1)==0 and get_curly_count(exp2)==0):
+
+ variable = match.group(2)
+ exp = match.group(3).strip()
+ old = match.group(0)
+ old = old[1:]
+ dico_conditions = {}
+
+ def rewrite_ternary(exp, dico_conditions):
+ exp = exp.strip()
+ if(exp[0]=="(" and exp[-1]==")"):
+ exp = exp[1:-1].strip()
+ try:
+ expression, end_condition = get_code_until_character(exp, "?")
+ except:
+ return exp
+ condition = exp[:end_condition].strip()
+ exp2, end_exp1 = get_code_until_character(expression, ":")
+ dico_true, dico_false = {}, {}
+ dico_conditions[condition] = {"True": rewrite_ternary(expression[:end_exp1], dico_true), "False": rewrite_ternary(exp2, dico_false)}
+ return dico_conditions
+
+
+ rewrite_ternary(exp, dico_conditions)
+
+
+ def rewrite_dico_2_condition(var, dico_condition, num = 0):
+ code = ''
+ if(type(dico_condition)==str):
+ return f"{var} = {dico_condition}\n"
+ for condition in dico_condition:
+ code = f"if({condition}) {{\n\t{rewrite_dico_2_condition(var, dico_condition[condition]['True'], num = num+1)}}} else {{\n\t{rewrite_dico_2_condition(var, dico_condition[condition]['False'], num = num+1)}}}\n"
+ return code
+
+ new = rewrite_dico_2_condition(f"{def_variable} {variable}", dico_conditions)+'\n'
+
to_replace.append((old, new))
- else:
- checked.append(match.group(0))
+ tmp = code
+ code = code.replace(old, new, 1)
+ if(old!=new and tmp==code):
+ raise Exception("This shouldn't happen -> the code wasn't replaced")
+ searching = True
+ break
+
for r in to_replace:
old, new = r
self.add_to_ternary_operation_dico(old, new)
- tmp = code
- code = code.replace(old, new, 1)
- if(old!=new and tmp==code):
- raise Exception("This shouldn't happen -> the code wasn't replaced")
+
#Check if there is still a ternary operation in this case we cannot analyse it
- #Cause it is a complexe/multiple ternanry operation
for match in re.finditer(pattern, code):
- if(match.group(0) not in checked):
- raise BioFlowInsightError(f"Detected a multi ternary operation (a ternary operation in a ternary operation) in the file '{self.origin.get_file_address()}'. BioFlow-Insight does not support this, try defining it in a different way.", type="Multi ternary operation")
+ raise BioFlowInsightError(f"Detected a multi ternary operation (a ternary operation in a ternary operation) in the file '{self.origin.get_file_address()}'. BioFlow-Insight does not support this, try defining it in a different way.", type="Multi ternary operation")
return code
def rewrite_jump_dot(self, code):
diff --git a/src/outils.py b/src/outils.py
index a69da0e..a80f07c 100644
--- a/src/outils.py
+++ b/src/outils.py
@@ -1666,3 +1666,63 @@ def extract_tools(script, extract_general_tools = False):
#Return the tools extarcted
return list(set(tools))
+
+
+def get_code_until_character(code, char):
+
+ start = 0
+
+ curly_count, parenthese_count = 0, 0
+ quote_single, quote_double = False, False
+ triple_single, triple_double = False, False
+
+
+ while(start<len(code)):
+ checked_triple = False
+ if(start+3<=len(code)):
+ if(code[start:start+3]=="'''" and not quote_single and not quote_double and not triple_single and not triple_double):
+ triple_single = True
+ start+=3
+ checked_triple = True
+ elif(code[start:start+3]=="'''" and not quote_single and not quote_double and triple_single and not triple_double):
+ triple_single = False
+ start+=3
+ checked_triple = True
+
+ if(code[start:start+3]=='"""' and not quote_single and not quote_double and not triple_single and not triple_double):
+ triple_double = True
+ start+=3
+ checked_triple = True
+ elif(code[start:start+3]=='"""' and not quote_single and not quote_double and not triple_single and triple_double):
+ triple_double = False
+ start+=3
+ checked_triple = True
+
+ if(not checked_triple):
+ if(code[start]=="{" and not quote_single and not quote_double and not triple_single and not triple_double):
+ curly_count+=1
+ elif(code[start]=="}" and not quote_single and not quote_double and not triple_single and not triple_double):
+ curly_count-=1
+
+ if(code[start]=="(" and not quote_single and not quote_double and not triple_single and not triple_double):
+ parenthese_count+=1
+ elif(code[start]==")" and not quote_single and not quote_double and not triple_single and not triple_double):
+ parenthese_count-=1
+
+ if(code[start]=="'" and not quote_single and not quote_double and not triple_single and not triple_double):
+ if(code[start-1]!="\\" or (code[start-1]=="\\" and code[start-2]=="\\")):
+ quote_single=True
+ elif(code[start]=="'" and quote_single and not quote_double and not triple_single and not triple_double):
+ if(code[start-1]!="\\" or (code[start-1]=="\\" and code[start-2]=="\\")):
+ quote_single=False
+
+ if(code[start]=='"' and not quote_single and not quote_double and not triple_single and not triple_double):
+ if(code[start-1]!="\\" or (code[start-1]=="\\" and code[start-2]=="\\")):
+ quote_double=True
+ elif(code[start]=='"' and not quote_single and quote_double and not triple_single and not triple_double):
+ if(code[start-1]!="\\" or (code[start-1]=="\\" and code[start-2]=="\\")):
+ quote_double=False
+ if(code[start]==char and parenthese_count==0 and curly_count ==0 and [quote_single, quote_double, triple_single, triple_double]==[False, False, False, False]):
+ return code[start+1:], start
+ start+=1
+ raise Exception("")
diff --git a/src/workflow.py b/src/workflow.py
index fe35077..790c0c9 100644
--- a/src/workflow.py
+++ b/src/workflow.py
@@ -708,31 +708,35 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
if(escape>=100):
return min_processes
already_tried.append(set(random_relevant_processes))
- _, cluster_organisation = w.convert_workflow_2_user_view(relevant_processes=random_relevant_processes, render_graphs = False)
-
- tab_nb_executors_per_cluster, tab_nb_processes_per_cluster, tab_nb_conditions_per_cluster = [], [], []
- for c in cluster_organisation:
- tab_nb_executors_per_cluster.append(cluster_organisation[c]["nb_executors"])
- tab_nb_processes_per_cluster.append(cluster_organisation[c]["nb_processes"])
- tab_nb_conditions_per_cluster.append(cluster_organisation[c]["nb_conditions"])
- nb_clusters = len(cluster_organisation)
- nb_non_relevant_clusters = 0
- for c in cluster_organisation:
- #This means it's a non relvant cluster
- if("non_relevant_cluster_" in c):
- nb_non_relevant_clusters+=1
-
- uniformity_variance = 0
- average_number_of_process_per_cluster = np.mean(tab_nb_processes_per_cluster)
- for x in tab_nb_processes_per_cluster:
- uniformity_variance += (average_number_of_process_per_cluster-x)**2/nb_clusters
-
- score = concordance_factor * np.max(np.array(tab_nb_conditions_per_cluster)/np.array(tab_nb_executors_per_cluster)) + \
- uniformity_factor * (uniformity_variance / number_processes_called) + \
- min_nb_clusters_factor * (nb_clusters / number_processes_called) + \
- min_nb_non_relevant_cluster_factor * (nb_non_relevant_clusters / nb_clusters)
-
+ def get_score_from_set_relevant_processes(random_relevant_processes):
+ _, cluster_organisation = w.convert_workflow_2_user_view(relevant_processes=random_relevant_processes, render_graphs = False)
+
+ tab_nb_executors_per_cluster, tab_nb_processes_per_cluster, tab_nb_conditions_per_cluster = [], [], []
+ for c in cluster_organisation:
+ tab_nb_executors_per_cluster.append(cluster_organisation[c]["nb_executors"])
+ tab_nb_processes_per_cluster.append(cluster_organisation[c]["nb_processes"])
+ tab_nb_conditions_per_cluster.append(cluster_organisation[c]["nb_conditions"])
+
+ nb_clusters = len(cluster_organisation)
+ nb_non_relevant_clusters = 0
+ for c in cluster_organisation:
+ #This means it's a non relvant cluster
+ if("non_relevant_cluster_" in c):
+ nb_non_relevant_clusters+=1
+
+ uniformity_variance = 0
+ average_number_of_process_per_cluster = np.mean(tab_nb_processes_per_cluster)
+ for x in tab_nb_processes_per_cluster:
+ uniformity_variance += (average_number_of_process_per_cluster-x)**2/nb_clusters
+
+ score = concordance_factor * np.max(np.array(tab_nb_conditions_per_cluster)/np.array(tab_nb_executors_per_cluster)) + \
+ uniformity_factor * (uniformity_variance / number_processes_called) + \
+ min_nb_clusters_factor * (nb_clusters / number_processes_called) + \
+ min_nb_non_relevant_cluster_factor * (nb_non_relevant_clusters / nb_clusters)
+ return score, cluster_organisation
+
+ score, cluster_organisation = get_score_from_set_relevant_processes(random_relevant_processes)
if(len(cluster_organisation)>=reduction_alpha*number_processes_called and
len(cluster_organisation)<=reduction_beta*number_processes_called and
score<min_score):
--
GitLab