From b0dec6bea34ec2577c36057a74f34793ba08dac5 Mon Sep 17 00:00:00 2001
From: George Marchment <georgemarchment@yahoo.fr>
Date: Mon, 21 Oct 2024 16:45:47 +0200
Subject: [PATCH] update
---
src/graph.py | 88 ++++++++++++++++++++++++++++++++++++++++++++
src/nextflow_file.py | 3 ++
src/outils_graph.py | 24 ++++++++++--
src/workflow.py | 5 ++-
4 files changed, 115 insertions(+), 5 deletions(-)
diff --git a/src/graph.py b/src/graph.py
index 0f26b79..7c02415 100644
--- a/src/graph.py
+++ b/src/graph.py
@@ -3,6 +3,7 @@ import json
import networkx as nx
import numpy as np
import copy
+import re
from .outils_graph import *
@@ -358,6 +359,93 @@ class Graph():
new_node_2_subworkflows[node] = node_2_subworkflows[id]
return new_node_2_subworkflows
+ #==========================================================
+ #Check if fake dependency is created when created user view
+ #==========================================================
+ #Here to check if a fake dependency is created, I'm gonna compare the edges
+ #of the level graphs between the user view and the process dependency
+ #Each of the user view edges (with subworkflo) should be in the process dependency edges
+ def check_fake_dependency_user_view(self):
+ #This function removes the "<>" from the node name
+ #And the same for the subworkflows
+ def clean_node(node):
+ #Case the node is a process
+ if(node[0]=="<"):
+ #We just remove the '<>' around the name
+ node = node[1:-1]
+ else:#it's a subworkflow
+ for match in re.finditer(r"id_\d+\.\d+\_(.+)", node):
+ node = match.group(1)
+ return node
+
+ #First by checking if the node_2_subworkflows are the same, if it's the case i don't need to compare
+ if(self.node_2_subworkflows_process_dependency_graph!=self.node_2_subworkflows_user_view):
+ dico_process_dependency_graph = self.dico_process_dependency_graph
+ user_view_with_subworkflows = self.user_view_with_subworkflows
+ user_view_subworkflows = get_subworkflows_names(user_view_with_subworkflows)
+ #Get the level workflows for the process dependency graph
+ max_level = get_max_level(dico_process_dependency_graph)
+ dependency_levels = []
+ for l in range(max_level+1):
+ new_dico = get_graph_level_l(dico_process_dependency_graph, l)
+ dependency_levels.append(new_dico)
+ #Get the level workflows for the user view
+ max_level = get_max_level(user_view_with_subworkflows)
+ user_view_levels = []
+ for l in range(max_level+1):
+ new_dico = get_graph_level_l(user_view_with_subworkflows, l)
+ user_view_levels.append(new_dico)
+ #For each level, i'm gonna check the edges
+ for i in range(len(user_view_levels)):
+ user_view_level = user_view_levels[i]
+ dependency_level = dependency_levels[i]
+ for sub in user_view_subworkflows:
+ for edge_user in user_view_level["edges"]:
+ if(f"_{sub}" in edge_user["A"] or f"_{sub}" in edge_user["B"]):
+ if(edge_user["A"]!="input" and edge_user["A"]!="output" and edge_user["B"]!="input" and edge_user["B"]!="output"):
+ #This boolean if is to check if the edge 'edge_user' has equivalence in the process dependency graph
+ has_matching_user_dependency = False
+
+ for edge_process in get_edges(dependency_level):
+ if(f"_{sub}" in edge_process["A"] or f"_{sub}" in edge_process["B"]):
+ node = ""
+ side = ""
+ #Determine if it's A or B
+ if(f"_{sub}" in edge_process["A"]):
+ node = edge_process["B"]
+ side = "B"
+ if(f"_{sub}" in edge_process["B"]):
+ node = edge_process["A"]
+ side = "A"
+ node = clean_node(node)
+ if(node in edge_user[side]):
+ has_matching_user_dependency = True
+
+ if(not has_matching_user_dependency):
+ #Check if there is an indirect path that exist
+ node_A = clean_node(edge_user["A"])
+ node_B = clean_node(edge_user["B"])
+ nodes_level = get_nodes_from_edges(get_edges(dependency_level))
+ node_A_temp, node_B_temp = "", ""
+ for A in node_A.split("_$$_"):
+ for tmp in nodes_level:
+ if A in tmp:
+ node_A_temp = tmp
+ for B in node_B.split("_$$_"):
+ for tmp in nodes_level:
+ if B in tmp:
+ node_B_temp = tmp
+
+ if(not exist_path_dico(node_A_temp, node_B_temp, dependency_level)):
+ print("False dependency", edge_user)
+ return True
+
+
+
+ return False
+ else:
+ return False
+
#============================
#METADATA FROM GRAPH
#============================
diff --git a/src/nextflow_file.py b/src/nextflow_file.py
index 4180fd0..30e3daa 100644
--- a/src/nextflow_file.py
+++ b/src/nextflow_file.py
@@ -732,6 +732,9 @@ class Nextflow_File(Nextflow_Building_Blocks):
def node_2_subworkflows_user_view(self):
return self.graph.node_2_subworkflows_user_view()
+ def check_fake_dependency_user_view(self):
+ return self.graph.check_fake_dependency_user_view()
+
def add_main_DSL1_2_rocrate(self, dico, file_dico, file_name):
diff --git a/src/outils_graph.py b/src/outils_graph.py
index 0f5c1b1..1ef2b2d 100644
--- a/src/outils_graph.py
+++ b/src/outils_graph.py
@@ -493,6 +493,16 @@ def exist_path(A, B, edges):
visited[n] = False
return exist_path_rec(A, B, edges, visited)
+def get_edges(dico, val= []):
+ val+=dico["edges"]
+ for sub in dico["subworkflows"]:
+ val=get_edges(dico["subworkflows"][sub], val)
+ return val
+
+def exist_path_dico(A, B, dico):
+ edges = get_edges(dico)
+ return exist_path(A, B, edges)
+
def nr_path_succ(n, r, dico, R):
rest_of_R = set(R)-set([r])
@@ -589,7 +599,7 @@ def get_name_new_node(new_nodes, relevant_modules):
#Arbitrary choice of choosing the name with the longest name
longest_name = new_nodes[0][0]
for name in new_nodes:
- if(len(longest_name)<len(name[0])):
+ if(len(longest_name)>len(name[0])):
longest_name = name[0]
return longest_name
@@ -638,7 +648,7 @@ def relev_user_view_builder(dico_param, relevant_modules):
for out in outputs:
dico["edges"].append({'A':out, 'B':'output'})
#TODO remove this -> it's to replicate the one in the algortihm demo
- #dico["edges"].append({'A':get_id_from_name(dico, f"M5{tag}0")[0], 'B':'output'})
+ #dico["edges"].append({'A':get_id_from_name(dico, f"M5_0{tag}0")[0], 'B':'output'})
for input in inputs:
dico["edges"].append({'A':"input", 'B':input})
U = []
@@ -738,7 +748,7 @@ def relev_user_view_builder(dico_param, relevant_modules):
for i in range(len(new_nodes)):
new_nodes[i].sort()
new_name = get_name_new_node(get_names_tab(dico, new_nodes[i]), relevant_modules)
- node = {"id": ''.join(new_nodes[i]).replace('<', '').replace('>', ''),
+ node = {"id": '_$$_'.join(new_nodes[i]).replace('<', '').replace('>', ''),
"name": new_name.split(tag)[0],
"shape": "ellipse",
"xlabel": f"{len(new_nodes[i])}",
@@ -897,12 +907,18 @@ def get_graph_level_l(dico, level):
return new_dico
def get_number_of_subworkflows(dico, val= 0):
-
for sub in dico["subworkflows"]:
if(dico["subworkflows"][sub]["nodes"]!=[]):
val += 1
val=get_number_of_subworkflows(dico["subworkflows"][sub], val)
return val
+def get_subworkflows_names(dico, val= []):
+ for sub in dico["subworkflows"]:
+ if(dico["subworkflows"][sub]["nodes"]!=[]):
+ val.append(sub)
+ val=get_subworkflows_names(dico["subworkflows"][sub], val)
+ return val
+
diff --git a/src/workflow.py b/src/workflow.py
index 1636137..61df16e 100644
--- a/src/workflow.py
+++ b/src/workflow.py
@@ -593,4 +593,7 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
return self.nextflow_file.node_2_subworkflows_process_dependency_graph()
def node_2_subworkflows_user_view(self):
- return self.nextflow_file.node_2_subworkflows_user_view()
\ No newline at end of file
+ return self.nextflow_file.node_2_subworkflows_user_view()
+
+ def check_fake_dependency_user_view(self):
+ return self.nextflow_file.check_fake_dependency_user_view()
\ No newline at end of file
--
GitLab