diff --git a/src/outils.py b/src/outils.py
index 65a91ba32c62747ae090066fcb7803e36e4d60cc..5c3e7c6c3b8aff35f8fa9286fcb5ab06ae3507c6 100644
--- a/src/outils.py
+++ b/src/outils.py
@@ -1025,7 +1025,7 @@ def extract_conditions(code):
start_else, end_else = match.span(0)
start_else+=end
end_else = extract_curly(code, end_else+end)
- conditions_dico[' && '.join(["neg({})".format(v) for v in conditions])] = (start_else,end_else)
+ conditions_dico[' && '.join(["!({})".format(v) for v in conditions])] = (start_else,end_else)
start_inside, end_inside = match.span(0)[1]+end, end_else-1
conditions_dico = adding_inside(conditions_dico, code, start_inside, end_inside)
#print(code[start_else:end_else])
@@ -1133,6 +1133,12 @@ def operation_2_DSL2(code, origin):
return code
def format_with_tabs(code):
+
+ def replace_jump(match):
+ return "\n"
+ #Removing the current "\t"s and extras " "
+ code = re.sub(r"\n[\t ]*", replace_jump, code)
+
start = 0
curly_count, parenthese_count = 0, 0
@@ -1251,4 +1257,54 @@ def group_together_ifs(code):
code = code.replace(code[big_start:big_end], f"if({condition_1}{{\n{inside_1}\n{inside_2}\n}}\n")
break
- return code
\ No newline at end of file
+ return code
+
+def remove_extra_jumps(code):
+ changed = True
+ while(changed):
+ changed = False
+ temp = code
+ def replacer(match):
+ return "\n\n"
+ code = re.sub(r"\n\s*\n\s*\n", replacer, code)
+ if(code!=temp):
+ changed = True
+
+ return code
+
+#This functions analyses the body of the subworkflow and the emitted values
+#If a channel is created in a certain condition (and not the negative) -> then we create it
+def get_channels_to_add_in_false_conditions(body, emitted_channels):
+ conditions = extract_conditions(body)
+ channels_2_conditions = {}
+ #Creating the dictionnary channels 2 conditions
+ for channel in emitted_channels:
+ channels_2_conditions[channel] = []
+ for match in re.finditer(fr"{re.escape(channel)}\s*=", body):
+ start, end = match.span(0)
+ for c in conditions:
+ start_condition, end_conditions = conditions[c]
+ if(start_condition<=start and start<=end_conditions):
+ channels_2_conditions[channel].append(c)
+ #Simplifying the list of conditions
+ #TODO -> here it's important that the input workflow doesn't have a too complexe condition systems (e.g. writting the same condition in multiple ways)
+ #Cause the converter doesn't analyse the conditions -> and basically it would create things which shloudn't
+ for channel in channels_2_conditions:
+ tab = channels_2_conditions[channel]
+ to_remove = []
+ for condition in tab:
+ #If the condition and it's neagtion are in the tab -> then we remove the condition and it's negation form the list
+ negation = f"!({condition})"
+ if(negation in tab):
+ to_remove.append(condition)
+ to_remove.append(negation)
+ for r in to_remove:
+ tab.remove(r)
+ channels_2_conditions[channel] = tab
+
+ #For the remaining condition in the list -> need to create an empty channel in the case of the negation
+ for condition in channels_2_conditions[channel]:
+ body += f"\nif(!({condition})) {{\n{channel} = Channel.empty()\n}}"
+
+ return body
+
diff --git a/src/workflow.py b/src/workflow.py
index 179595d2ccb217e6ba114edca4bc5a436e4d2ead..4aa19ba19e59e72e6eb400b4e6e37d5679e93c9f 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, replace_group1, group_together_ifs, extract_curly
+from .outils import is_git_directory, format_with_tabs, replace_thing_by_call, replace_group1, group_together_ifs, extract_curly, remove_extra_jumps, get_channels_to_add_in_false_conditions
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
@@ -759,9 +759,9 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
code = self.nextflow_file.get_code()
#params_section = f"//PARAMS_SECTION_{tag}"
- function_section = f"//FUNCTION_SECTION_{tag}"
- process_section = f"//PROCESS_SECTION_{tag}"
- subworkflow_section = f"//SUBWORKFLOW_SECTION_{tag}"
+ function_section = f"//FUNCTION_SECTION"
+ process_section = f"//PROCESS_SECTION"
+ subworkflow_section = f"//SUBWORKFLOW_SECTION"
ankers = function_section+ "\n"*3 + process_section+ "\n"*3 + subworkflow_section
@@ -823,14 +823,31 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
self.__init__(str(temp_file), display_info = False, duplicate=True)
self.initialise(create_rocrate=False)
- #for e in self.get_all_executors():
- # print(e.get_code(get_OG = True))
+
#Get the clusters and the code
self.nextflow_file.generate_user_view(relevant_processes = relevant_processes, processes_2_remove = [])
clusters = self.nextflow_file.graph.get_clusters_from_user_view()
+
+ #cluster_2_subworkflows = []
+ #print(clusters)
+ ##Basically here i'm checking if a subworkflow is split
+ ##TODO Check this -> i think in the case something is at the root -> it won't work
+ #for sub in self.get_subworkflows_called():
+ # clusters_in_which_elements_are_taken = []
+ # #TODO do this
+ # print(sub.get_name(), sub.get_executors())
+ # for exe in sub.get_executors():
+ # print(exe, exe.get_subworkflow_origin())
+ # #If there are multiple clusters and the number of clusters isn't equal to all the clusters -> it means that subworkflow is broken
+ # if(len(clusters_in_which_elements_are_taken)>1 and len(clusters_in_which_elements_are_taken)!=len(clusters)):
+ # print(sub.get_name())
#print(clusters)
+ #for c in clusters:
+ # for ele in c:
+ # print(ele.get_type())
+
#TODO -> need to break clusters here
#And redo analysis
@@ -975,6 +992,10 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
temp = '\n'.join(old_output_names)
emit = f"\nemit:\n{temp}\n"
+ #Adding empty channels if it doesn't exist in the case of a negative condition
+ body = get_channels_to_add_in_false_conditions(body, old_output_names)
+
+
#Replace names inside subworkflow
subworkflow_code = f"workflow {name} {{\n{take}\nmain:\n{body}\n{emit}\n}}"
@@ -1010,10 +1031,6 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
subworkflow_cluster_calls_to_add.append(subworkfow_call)
index_cluster+=1
-
- print(format_with_tabs(subworkflow_code))
- print("//-----------")
-
#TODO -> rmoving the conditions which are problematic
#This might not be the probleme -> when rerunnung the analysis isn't totally robust
@@ -1048,7 +1065,24 @@ George Marchment, Bryan Brancotte, Marie Schmit, Frédéric Lemoine, Sarah Cohen
code = replace_group1(code, pattern, new)
#code = code.replace(old, new)
- print(code)
+
+ #Add the subworkflow defintions
+ #-------------------------------------
+ #Add anker
+ subworkflow_section = f"//ANKER 4 SUBWORKFLOW DEF"
+ to_replace = ""
+ for match in re.finditer(r"workflow\s+\w*\s*\{", code):
+ to_replace = match.group(0)
+ break
+ if(to_replace==""):
+ raise Exception("No call to a workflow")
+
+ code = code.replace(to_replace, f"{subworkflow_section}\n\n{to_replace}")
+
+ for sub in subworkflow_clusters_to_add:
+ code = code.replace(f'{subworkflow_section}', f"{sub}\n\n{subworkflow_section}")
+
+ print(remove_extra_jumps(format_with_tabs(code)))
#So basically when retriving a thing (process or subworkflow)