diff --git a/src/graph.py b/src/graph.py
index e0e5dc58f0801f96967db05001c26b96cdd01766..2fd13eca88c66faaf91fc7066b7b3a0847a956a0 100644
--- a/src/graph.py
+++ b/src/graph.py
@@ -471,6 +471,27 @@ class Graph():
return False
else:
return False
+
+ #This method returns a list of processes
+ def get_edges_that_create_cycle(self):
+ links_flattened = initia_link_dico_rec(get_flatten_dico(self.get_process_dependency_graph()))
+ not_source_2_sink = []
+ node_2_sink = []
+ for node in links_flattened:
+ if(links_flattened[node]==[]):
+ node_2_sink.append(node)
+ else:
+ not_source_2_sink+=links_flattened[node]
+ not_source_2_sink = set(not_source_2_sink)
+ source_2_node = list(set(links_flattened.keys()).difference(not_source_2_sink))
+ links_flattened_source_sink = links_flattened.copy()
+ links_flattened_source_sink["source"], links_flattened_source_sink["sink"] = source_2_node, []
+ for node in node_2_sink:
+ links_flattened_source_sink[node].append("sink")
+
+ #The simple loops are included in this
+ _, edges_create_cycles = get_number_cycles(links_flattened_source_sink)
+ return edges_create_cycles
#============================
#METADATA FROM GRAPH
diff --git a/src/process.py b/src/process.py
index 14c7b635518f5bd660707708d350061bc15d34a7..75cbb1f7de71f1746aa73c82eca1784ebb452df4 100644
--- a/src/process.py
+++ b/src/process.py
@@ -105,6 +105,11 @@ class Process(Nextflow_Building_Blocks):
def get_script_code(self):
+ code = " "+self.script_code+" "
+ if(self.script_code.count('"""')==2):
+ return self.script_code.split('"""')[1]
+ if(self.script_code.count("'''")==2):
+ return self.script_code.split("'''")[1]
return self.script_code
def get_name(self):
@@ -446,8 +451,9 @@ class Process(Nextflow_Building_Blocks):
def initialise_inputs_outputs(self):
DSL = self.nextflow_file.get_DSL()
if(DSL=="DSL1"):
- self.initialise_inputs_DSL1()
- self.initialise_outputs_DSL1()
+ if(self.origin!=None):
+ self.initialise_inputs_DSL1()
+ self.initialise_outputs_DSL1()
elif(DSL=="DSL2"):
self.initialise_inputs_DSL2()
self.initialise_outputs_DSL2()
diff --git a/src/subworkflow.py b/src/subworkflow.py
index 8c4c0576fa392bc3086218f77ccf51025c99a2db..d9c8ff016b55dfcd82538479b9aca39eaf4814ee 100644
--- a/src/subworkflow.py
+++ b/src/subworkflow.py
@@ -103,13 +103,33 @@ class Subworkflow(Main):
for o in self.emit:
code = code.replace(o.get_code(get_OG = True), o.simplify_code(return_tab = False), 1)
- #Renaming the takes in the subworkflow
+ #Renaming the takes in the subworkflow (only body)
+ code_up_to_emit, code_after_emit = code, ""
+ for match in re.finditer(constant.EMIT_SUBWORKFLOW, code):
+ start, _ = match.span(0)
+ code_up_to_emit = code[:start]
+ code_after_emit = code[start:]
name = self.get_alias()
for t in self.take:
if(len(t.get_gives())!=1):
raise Exception("This shoudn't happen")
ch = t.get_gives()[0]
- code = replace_group1(code, fr"[^\w]({re.escape(ch.get_code())})[^\w]", f"{ch.get_code()}_{name}")
+ code_up_to_emit = replace_group1(code_up_to_emit, fr"[^\w]({re.escape(ch.get_code())})[^\w]", f"{ch.get_code()}_{name}")
+
+ #Renaming the takes in the emits -> if the takes are given as emits
+ for e in self.emit:
+ channels_take = []
+ for t in self.take:
+ channels_take.append(t.get_gives()[0])
+ re_write_channel = False
+ for o in e.origins:
+ if(o in channels_take):
+ re_write_channel = True
+ if(re_write_channel):
+ ch = e.origins[0]
+ code_after_emit = replace_group1(code_after_emit, fr"[^\w]({re.escape(ch.get_code())})[^\w]", f"{ch.get_code()}_{name}")
+
+ code = code_up_to_emit+code_after_emit
return code
diff --git a/src/workflow.py b/src/workflow.py
index 10c4d0cd21ca3d14d3ab7dcdf22bb0b4359e02e7..7e496aff14ab9f6942bd8637319ffcbb499fe4e2 100644
--- a/src/workflow.py
+++ b/src/workflow.py
@@ -379,6 +379,7 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
#This methods generates a random set of processes to consider as relavant
#It's not a uniform random it's a bit of a gaussian, centered at 0.5
def generate_random_relevant_processes(self, alpha = -1):
+ edges_create_cycles = self.graph.get_edges_that_create_cycle()
import random
#Random value between 0 and 1, centered at 0.5
@@ -394,13 +395,6 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
return val
if(self.duplicate):
- if(alpha == -1):
- alpha = get_value()
- else:
- if(0<=alpha and alpha<=1):
- None
- else:
- raise BioFlowInsightError("alpha is not in the interval [0; 1]")
processes_called = []
if(self.get_DSL()=="DSL2"):
@@ -410,11 +404,33 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
processes_called.append(p)
else:
processes_called = self.get_first_file().get_processes()
- nb_2_select = int(alpha*len(processes_called))
- sampled = random.sample(set(processes_called), nb_2_select)
- name_select = []
- for p in sampled:
- name_select.append(p.get_alias())
+
+ searching = True
+ while(searching):
+ searching = False
+ if(alpha == -1):
+ alpha = get_value()
+ else:
+ if(0<=alpha and alpha<=1):
+ None
+ else:
+ raise BioFlowInsightError("alpha is not in the interval [0; 1]")
+ nb_2_select = int(alpha*len(processes_called))
+ sampled = random.sample(set(processes_called), nb_2_select)
+
+ sampled_str = []
+ for s in sampled:
+ sampled_str.append(str(s))
+ for e in edges_create_cycles:
+ if(e[0] in sampled_str and e[1] in sampled_str):
+ #So that means there are the 2 nodes which form the cycle edge in the relevant processes
+ #-> it means we need to regenerated relevant processes
+ searching = True
+ break
+
+ name_select = []
+ for p in sampled:
+ name_select.append(p.get_alias())
return name_select
else:
raise BioFlowInsightError("Trying to generate random relevant processes however option 'duplicate' is not activated.")
@@ -753,6 +769,21 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
def convert_workflow_2_user_view(self, relevant_processes = [], render_graphs = True):
self.iniatilise_tab_processes_2_remove()
self.graph.initialise(processes_2_remove = self.processes_2_remove)
+
+ def get_object(address):
+ address = int(re.findall(r"\dx\w+", address)[0], base=16)
+ return ctypes.cast(address, ctypes.py_object).value
+ #Check that there are no cycles which will break the creation of the user view:
+ edges_create_cycles = self.graph.get_edges_that_create_cycle()
+ edges_create_cycles_objects = []
+ for e in edges_create_cycles:
+ edges_create_cycles_objects.append((get_object(e[0]), get_object(e[1])))
+ for e in edges_create_cycles_objects:
+ n1 = e[0].get_alias()
+ n2 = e[1].get_alias()
+ if(n1 in relevant_processes and n2 in relevant_processes):
+ raise BioFlowInsightError(f"The processes '{n1}' and '{n2}' cannot both be relevant processes since there is a dependency apparant in the workflow between the 2")
+
ternary_operation_dico = self.ternary_operation_dico
map_element_dico = self.map_element_dico
@@ -763,6 +794,7 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
if(self.get_DSL()=="DSL2"):
code = self.simplify_workflow_code()
self.rewrite_and_initialise(code, self.processes_2_remove, render_graphs=render_graphs)
+
if(self.duplicate):