diff --git a/src/outils.py b/src/outils.py
index 9415f31d5b8be5dd27653e8d602d00c71260106d..bff466056e8fff5deb2cbd2b64607697fdb157ab 100644
--- a/src/outils.py
+++ b/src/outils.py
@@ -1187,4 +1187,9 @@ def replace_thing_by_call(tab):
for r in to_remove:
tab.remove(r)
tab+=to_add
- return tab
\ No newline at end of file
+ return tab
+
+def replace_group1(text, pattern, replacement):
+ def replacer(match):
+ return match.group(0).replace(match.group(1), replacement)
+ return re.sub(pattern, replacer, text)
\ No newline at end of file
diff --git a/src/workflow.py b/src/workflow.py
index 18a10ad62734b68e3009086ef40ed48872c93191..cf5f359a8afc793ae22f7efb28869bbc83d31f3a 100644
--- a/src/workflow.py
+++ b/src/workflow.py
@@ -3,7 +3,7 @@
from .nextflow_file import Nextflow_File
from .ro_crate import RO_Crate
from . import constant
-from .outils import is_git_directory, format_with_tabs, replace_thing_by_call
+from .outils import is_git_directory, format_with_tabs, replace_thing_by_call, replace_group1
from .outils_graph import flatten_dico, initia_link_dico_rec, get_number_cycles
from .outils_annotate import get_tools_commands_from_user_for_process
from .bioflowinsighterror import BioFlowInsightError
@@ -849,13 +849,18 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
#Creating the subworkflows from clusters
calls_in_operations = []
non_relevant_name = 1
+ channels_to_replace_outside_of_cluster = []
+ subworkflow_clusters_to_add, subworkflow_cluster_calls_to_add = [], []
+ index_cluster = 0
for elements in clusters:
#Only create the subworkflows for clusters with more than one element
processes_added = []
things_added_in_cluster = []
if(len(elements)>1):
- name, body, take, emit = "", "", "", ""
+ name, body, take, emit = "", "main:\n", "", ""
+ first_element = True
for ele in elements:
+
if(ele.get_type()=="Process"):
#Determine the name of the created subworkflow cluster
@@ -872,6 +877,15 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
raise Exception("This shoudn't happen since duplicate mode is activated")
call = call[0]
+
+ #If first element -> add marker for the subworkflow call
+ if(first_element):
+ code = code.replace(call.get_code(get_OG = True), f"//Anker_cluster{index_cluster}")
+ first_element = False
+ else:
+ code = code.replace(call.get_code(get_OG = True), "")
+
+
processes_added.append(call.get_first_element_called())
printed_condition = " && ".join(call.get_condition().get_conditions())
if(printed_condition!=""):
@@ -880,7 +894,14 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
body+=f"\n{call.get_code()}\n"
things_added_in_cluster.append(call)
elif(ele.get_type()=="Operation"):
-
+
+ #If first element -> add marker for the subworkflow call
+ if(first_element):
+ code = code.replace(ele.get_code(get_OG = True), f"//Anker_cluster{index_cluster}")
+ first_element = False
+ else:
+ code = code.replace(ele.get_code(get_OG = True), "")
+
#Ignore these cases
if(ele.get_code()[:4] not in ["emit", "take"]):
origins = ele.get_origins()
@@ -893,6 +914,8 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
else:
body+=f"\n{ele.get_code()}\n"
things_added_in_cluster.append(ele)
+ else:
+ raise Exception("This shoudn't happen")
#TODO check this part of the code is never seen
#Here we removing the Call_12313 thing
@@ -912,15 +935,93 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
name = f"non_relevant_cluster_{non_relevant_name}"
non_relevant_name+=1
- subworkflow_code = f"subworkflow {name} {{\n{take}\n{body}\n{emit}\n}}"
- print(format_with_tabs(subworkflow_code))
- for t in self.get_takes(things_added_in_cluster):
- print("*", t.get_code())
-
- for t in self.get_emits(things_added_in_cluster):
- print("-", t.get_code())
- print("-----------")
+ #TAKE
+ #Adding take parameters on the inside of the subworkflow
+ takes_param = self.get_takes(things_added_in_cluster)
+ new_param_names, index, old_param_names = [], 1, []
+ for param in takes_param:
+ param_name = f"param_{name}_{index}"
+ new_param_names.append(param_name)
+ old_param_names.append(param.get_code())
+ index += 1
+ if(len(new_param_names)>0):
+ temp = '\n'.join(new_param_names)
+ take = f"\ntake:\n{temp}\n"
+
+ #EMIT
+ #Adding the emitted outputs
+ emitted_outputs = self.get_emits(things_added_in_cluster)
+ new_output_names, index, old_output_names = [], 0, []
+ for output in emitted_outputs:
+ output_name = f"{name}.out[{index}]"
+ new_output_names.append(output_name)
+ old_output_names.append(output.get_code())
+ index += 1
+
+ if(len(old_output_names)>0):
+ temp = '\n'.join(old_output_names)
+ emit = f"\nemit:\n{temp}\n"
+
+
+ #Replace names inside subworkflow
+ subworkflow_code = f"subworkflow {name} {{\n{take}\n{body}\n{emit}\n}}"
+ params = ", ".join(old_param_names)
+ subworkfow_call = f"{name}({params})"
+
+ for i in range(len(new_param_names)):
+ pattern = fr"[\=\,\(] *({re.escape(takes_param[i].get_code())})[\s\,\)]"
+ subworkflow_code = replace_group1(subworkflow_code, pattern, new_param_names[i])
+ #subworkflow_code = subworkflow_code.replace(takes_param[i].get_code(), new_param_names[i])
+
+ for i in range(len(new_output_names)):
+ #In the case of channels, we just add chanel = subworkflow.out[i]
+ if(not bool(re.findall("\.\s*out", old_output_names[i]))):
+ subworkfow_call+=f"\n{old_output_names[i]} = {new_output_names[i]}"
+ #In the case of emitted values we need to replace the code on the outside
+ else:
+ channels_to_replace_outside_of_cluster.append((old_output_names[i], new_output_names[i]))
+
+
+ subworkflow_clusters_to_add.append(subworkflow_code)
+ subworkflow_cluster_calls_to_add.append(subworkfow_call)
+ index_cluster+=1
+ #print(format_with_tabs(subworkflow_code))
+ #print(subworkfow_call)
+ #
+ #for t in self.get_takes(things_added_in_cluster):
+ # print("*", t.get_code())
+ #
+ #for t in self.get_emits(things_added_in_cluster):
+ # print("-", t.get_code())
+ #print("-----------")
+
+
+ #TODO -> rmoving the conditions which are problematic
+ #This might not be the probleme -> when rerunnung the analysis isn't totally robust
+ #still_empty_conditions = True
+ #while(still_empty_conditions):
+ # still_empty_conditions = False
+ # for index_cluster in range(len(subworkflow_clusters_to_add)):
+ # #Replace empty if and else by nothing
+ # def replace_by_empty(match):
+ # still_empty_conditions = True
+ # return f"//Anker_cluster{index_cluster}"
+ # pattern = r"if\s*\([^\{]+\{\s*\/\/Anker_cluster"+str(index_cluster)+r"\s*\}\s*else\{\s*\}"
+ # code = re.sub(pattern, replace_by_empty, code)
+ #
+ # pattern = r"if\s*\([^\{]+\{\s*\}\s*else\{\s*\/\/Anker_cluster"+str(index_cluster)+r"\s*\}"
+ # code = re.sub(pattern, replace_by_empty, code)
+
+
+ for i in range(len(subworkflow_clusters_to_add)):
+ #print(f"//Anker_cluster{i}", subworkflow_cluster_calls_to_add[i])
+ code = code.replace(f"//Anker_cluster{i}", subworkflow_cluster_calls_to_add[i])
+
+ for old, new in channels_to_replace_outside_of_cluster:
+ code = code.replace(old, new)
+
+ print(code)
#So basically when retriving a thing (process or subworkflow)