diff --git a/src/graph.py b/src/graph.py
index f6c091921e52911459f9a1bee085f94a034793b6..ffff0593bf161cf927460e106a039610b6234f5f 100644
--- a/src/graph.py
+++ b/src/graph.py
@@ -147,7 +147,7 @@ class Graph():
def get_specification_graph(self, dirc = 'graphs', filename = "specification_graph", render_graphs = True):
generate_graph(self.get_output_dir()/ dirc /filename, self.full_dico, render_graphs = render_graphs)
- generate_graph(self.get_output_dir()/ dirc /(filename+"_without_artificial_nodes"), remove_artificial_nodes(self.full_dico), render_graphs = render_graphs)
+ #generate_graph(self.get_output_dir()/ dirc /(filename+"_without_artificial_nodes"), remove_artificial_nodes(self.full_dico), render_graphs = render_graphs)
def get_specification_graph_wo_labels(self, filename = "specification_graph_wo_labels", render_graphs = True):
generate_graph(self.get_output_dir()/'graphs'/filename, self.full_dico, label_edge=False, label_node=False, render_graphs = render_graphs)
@@ -434,23 +434,25 @@ class Graph():
self.initialise_flattened_dico(self.full_dico)
dico = remove_artificial_nodes(self.dico_flattened)
- self.user_view, self.new_nodes_user_view = relev_user_view_builder(dico, relevant_modules=relevant_processes)
+ self.user_view, self.new_nodes_user_view = relev_user_view_builder(dico, relevant_modules=relevant_processes, show_operations = False)
with open(self.get_output_dir()/ "graphs/user_view.json", 'w') as output_file :
json.dump(self.user_view, output_file, indent=4)
- user_view_with_subworkflows = add_subworkflows_2_dico(self.dico_process_dependency_graph, self.user_view)
- with open(self.get_output_dir()/ "graphs/user_view_with_subworkflows.json", 'w') as output_file :
- json.dump(user_view_with_subworkflows, output_file, indent=4)
+ #user_view_with_subworkflows = add_subworkflows_2_dico(self.dico_process_dependency_graph, self.user_view)
+ #with open(self.get_output_dir()/ "graphs/user_view_with_subworkflows.json", 'w') as output_file :
+ # json.dump(user_view_with_subworkflows, output_file, indent=4)
- return self.user_view, user_view_with_subworkflows
+ #return self.user_view, user_view_with_subworkflows
+ return self.user_view
def generate_user_view(self, relevant_processes = [], render_graphs = True):
- user_view, user_view_with_subworkflows = self.get_user_view_graph(relevant_processes = relevant_processes)
- self.user_view_with_subworkflows = user_view_with_subworkflows
+ #user_view, user_view_with_subworkflows = self.get_user_view_graph(relevant_processes = relevant_processes)
+ user_view = self.get_user_view_graph(relevant_processes = relevant_processes)
+ #self.user_view_with_subworkflows = user_view_with_subworkflows
generate_graph(self.get_output_dir()/'graphs'/"user_view", user_view, label_edge=True, label_node=True, render_graphs = render_graphs, root = False, relevant_nodes = copy.deepcopy(relevant_processes))
- generate_graph(self.get_output_dir()/'graphs'/"user_view_with_subworkflows", user_view_with_subworkflows, label_edge=True, label_node=True, render_graphs = render_graphs, root = False, relevant_nodes = copy.deepcopy(relevant_processes))
+ #generate_graph(self.get_output_dir()/'graphs'/"user_view_with_subworkflows", user_view_with_subworkflows, label_edge=True, label_node=True, render_graphs = render_graphs, root = False, relevant_nodes = copy.deepcopy(relevant_processes))
#This method returns the list of the clusters in topological order
diff --git a/src/outils_graph.py b/src/outils_graph.py
index 6a243f70456a8d53aca8d100b1ca9a7664e055b9..08b71a5dfb5578e6e38e131a550c054471706ca8 100644
--- a/src/outils_graph.py
+++ b/src/outils_graph.py
@@ -835,7 +835,7 @@ def get_color_node(node, new_nodes):
prop = 256- int(127*len(node)/max)
return rgb_to_hex(prop, prop, prop)
-def relev_user_view_builder(dico_param, relevant_modules):
+def relev_user_view_builder(dico_param, relevant_modules, show_operations = True):
import time
dico = copy.deepcopy(dico_param)
tag = str(time.time())
@@ -964,14 +964,27 @@ 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": '_$$_'.join(new_nodes[i]).replace('<', '').replace('>', ''),
- "name": new_name.split(tag)[0],
- "shape": "ellipse",
- "xlabel": f"{len(new_nodes[i])}",
- "fillcolor": get_color_node(new_nodes[i], new_nodes)}
+ name_printed = new_name.split(tag)[0]
+ shape = "ellipse"
+ if(show_operations):
+ if(name_printed==""):
+ shape = "point"
+ node = {"id": '_$$_'.join(new_nodes[i]).replace('<', '').replace('>', ''),
+ "name": name_printed,
+ "shape": shape,
+ "xlabel": f"{len(new_nodes[i])}",
+ "fillcolor": get_color_node(new_nodes[i], new_nodes)}
+ else:
+ if(name_printed!=""):
+ node = {"id": '_$$_'.join(new_nodes[i]).replace('<', '').replace('>', ''),
+ "name": name_printed,
+ "shape": shape,
+ "xlabel": f"{len(new_nodes[i])}",
+ "fillcolor": get_color_node(new_nodes[i], new_nodes)}
+
#If relevant module -> color it differently
if(new_name in relevant_modules):
- node["color"] = "yellow"
+ node["color"] = "#006903"
new_dico["nodes"].append(node)
added_edges = []
for edge in dico["edges"]:
@@ -984,12 +997,13 @@ def relev_user_view_builder(dico_param, relevant_modules):
if(edge["A"].replace('<', '').replace('>', '') in nA["id"]
and edge["B"].replace('<', '').replace('>', '') in nB["id"]
and edge_string not in added_edges):#So we don't have dupliacte edges
- new_dico["edges"].append({
- "A": nA["id"],
- "B": nB["id"],
- "label": ""
- })
- added_edges.append(edge_string)
+ if(show_operations or ((not show_operations) and nA["id"]!=nB["id"])):
+ new_dico["edges"].append({
+ "A": nA["id"],
+ "B": nB["id"],
+ "label": ""
+ })
+ added_edges.append(edge_string)
#The output nodes are the nodes which their outputs aren't connected to anything else
#TODO -> remove these comments if you want to root the graph
diff --git a/src/process.py b/src/process.py
index 6beb2873a99935e78961b729bbe0a62da8eef90b..e887430525b9685dce550c02149d1753ca9d5dea 100644
--- a/src/process.py
+++ b/src/process.py
@@ -582,6 +582,16 @@ class Process(Nextflow_Building_Blocks):
def get_tools(self, extract_general_tools = False):
return extract_tools(self.get_script_code(), extract_general_tools = extract_general_tools)
+
+ def get_all_conditions(self):
+ if(self.nextflow_file.get_DSL()=="DSL2"):
+ if(len(self.called_by)!=1):
+ raise Exception("This shouldn't happen")
+ return self.called_by[0].get_all_conditions()
+ else:
+ conditions = {}
+ return self.origin.get_all_conditions(conditions)
+
diff --git a/src/workflow.py b/src/workflow.py
index 790c0c9adbb5ddbe888c0c053748a16205467d21..371af87930f3bd56ba0e0c32a6c7ecce988586c3 100644
--- a/src/workflow.py
+++ b/src/workflow.py
@@ -686,10 +686,11 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
print('.', end='')
processes_2_tools[p.get_code()] = p.get_tools()
print("\n")
- print("Testing different combinations")
- print('-'*number_of_tries+">")
+ #print("Testing different combinations")
+ #print('-'*number_of_tries+">")
for i in range(number_of_tries):
- print('.', end='')
+ #print('.', end='')
+ print(i/number_of_tries*100)
w = copy.deepcopy(w_save)
if(process_pre_selection == "bioinfo"):
@@ -709,20 +710,81 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
return min_processes
already_tried.append(set(random_relevant_processes))
- 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)
+ #Here the nb of conditions returned is the number of conditions in the clusters after the rewrite
+ def get_nb_conditions_in_clusters(clusters):
+ nb_conditions_in_clusters = []
+ for cluster in clusters:
+ all_conditions_cluster = []
+ for c in cluster:
+ conditions_for_element = c.get_all_conditions()
+ if(len(conditions_for_element)==0):
+ all_conditions_cluster.append("no value")
+ else:
+ for condition in conditions_for_element:
+ all_conditions_cluster.append(condition.get_value())
+
+ all_conditions_cluster = list(set(all_conditions_cluster))
+
+ if(len(all_conditions_cluster)==1):
+ nb_conditions_in_clusters.append(0)
+ else:
+ try:
+ all_conditions_cluster.remove("no value")
+ except:
+ None
+ nb_conditions_in_clusters.append(len(all_conditions_cluster))
+ return nb_conditions_in_clusters
+
+
+ def get_score_from_set_relevant_processes(w_save, random_relevant_processes):
+ #w = copy.deepcopy(w_save)
+ #_, cluster_organisation = w.convert_workflow_2_user_view(relevant_processes=random_relevant_processes, render_graphs = False)
+ #print(random_relevant_processes)
+
+ #tab_nb_executors_per_cluster_1, tab_nb_processes_per_cluster_1, tab_nb_conditions_per_cluster_1 = [], [], []
+ #for c in cluster_organisation:
+ # tab_nb_executors_per_cluster_1.append(cluster_organisation[c]["nb_executors"])
+ # tab_nb_processes_per_cluster_1.append(cluster_organisation[c]["nb_processes"])
+ # tab_nb_conditions_per_cluster_1.append(cluster_organisation[c]["nb_conditions"])
- 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"])
+ w = copy.deepcopy(w_save)
+ w.generate_user_view(relevant_processes = random_relevant_processes, render_graphs=False)
+ clusters = w.graph.get_clusters_from_user_view()
+ cluster_with_processes = []
+ for cluster in clusters:
+ there_is_a_process = False
+ for ele in cluster:
+ if(ele.get_type() == "Process"):
+ there_is_a_process = True
+ cluster_with_processes.append(cluster)
+
+ #Number executors per cluster
+ tab_nb_executors_per_cluster = []
+ for cluster in cluster_with_processes:
+ tab_nb_executors_per_cluster.append(len(cluster))
- nb_clusters = len(cluster_organisation)
+ #Number condtions per cluster
+ tab_nb_conditions_per_cluster = get_nb_conditions_in_clusters(cluster_with_processes)
+ #Number of processes per cluster
+ tab_nb_processes_per_cluster = []
+ for cluster in cluster_with_processes:
+ nb_processes = 0
+ for ele in cluster:
+ if(ele.get_type()=="Process"):
+ nb_processes+=1
+ tab_nb_processes_per_cluster.append(nb_processes)
+
+
+ nb_clusters = len(cluster_with_processes)
nb_non_relevant_clusters = 0
- for c in cluster_organisation:
- #This means it's a non relvant cluster
- if("non_relevant_cluster_" in c):
+ for cluster in cluster_with_processes:
+ cluster_with_relevant_process = False
+ for c in cluster:
+ if(c.get_type()=="Process"):
+ if(c.get_alias() in random_relevant_processes):
+ #This means it's relvant cluster
+ cluster_with_relevant_process = True
+ if(not cluster_with_relevant_process):
nb_non_relevant_clusters+=1
uniformity_variance = 0
@@ -734,9 +796,9 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
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
+ return score, cluster_with_processes
- score, cluster_organisation = get_score_from_set_relevant_processes(random_relevant_processes)
+ score, cluster_organisation = get_score_from_set_relevant_processes(w_save, 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):