From 9f1bffdc1af720c2561942db7bd42c4086c4d0b0 Mon Sep 17 00:00:00 2001
From: George Marchment <georgemarchment@yahoo.fr>
Date: Sun, 4 Feb 2024 14:30:07 +0100
Subject: [PATCH] Starting to add Ro-crate support
---
src/constant.py | 23 +++++++++++++++++++++++
src/process.py | 8 ++++++++
src/ro_crate.py | 14 ++++++++++++++
src/workflow.py | 25 +++++++++++++++++++++++++
4 files changed, 70 insertions(+)
create mode 100644 src/ro_crate.py
create mode 100644 src/workflow.py
diff --git a/src/constant.py b/src/constant.py
index 9badf13..f41e6ea 100644
--- a/src/constant.py
+++ b/src/constant.py
@@ -21,6 +21,27 @@ LIST_OPERATORS = ["distinct", "filter", "first", "last", "randomSample", "take",
"close","dump","ifEmpty","print","println","set","view",
"empty", "of", "fromPath", "fromList", "subscribe", "value", "from"]#This last line is added by me:)
+TOOLS = [
+ "samtools",
+ "bcftools",
+ "fastqc",
+ "bedtools",
+ "multiqc",
+ "gatk",
+ "bwa",
+ "minimap2",
+ "tabix",
+ "vcf",
+ "wget",
+ "bgzip",
+ "hmmsearch",
+ "pigz",
+ "picard",
+ "star",
+ "iqtree",
+ "idxstats",
+]
+
#==========================
# PATTERNS
@@ -122,3 +143,5 @@ SUBWORKFLOW_HEADER = r'workflow +(\w+|\'[\w ]+\'|\"[\w ]+\") *{'
WORKFLOW_HEADER = r"workflow\s*\{"
WORKFLOW_HEADER_2 = r'[^\w](workflow\s*{)'
+
+
diff --git a/src/process.py b/src/process.py
index 8a94f27..7aee198 100644
--- a/src/process.py
+++ b/src/process.py
@@ -19,6 +19,7 @@ class Process(Nextflow_Building_Blocks):
self.output_code = ""
self.when_code = ""
self.script_code = ""
+ self.tools = []
self.initialise()
self.initialised = True
@@ -71,6 +72,12 @@ class Process(Nextflow_Building_Blocks):
def get_nb_outputs(self):
return len(self.outputs)
+ def extract_tools(self):
+ script = self.script_code.lower()
+ for tool in constant.TOOLS:
+ if tool in script:
+ self.tools.append(tool)
+
def initialise_parts(self):
code = self.get_code()
@@ -126,6 +133,7 @@ class Process(Nextflow_Building_Blocks):
self.when_code = temp_code
elif(variables_index[i]=='script'):
self.script_code = temp_code
+ self.extract_tools()
else:
raise Exception("This shoudn't happen!")
diff --git a/src/ro_crate.py b/src/ro_crate.py
new file mode 100644
index 0000000..e987fb7
--- /dev/null
+++ b/src/ro_crate.py
@@ -0,0 +1,14 @@
+
+
+class RO_Crate:
+ def __init__(self, workflow):
+ self.nextflow_file = workflow
+ self.directroy = '/'.join(workflow.get_file_address().split('/')[:-1])
+ self.dico = {}
+
+ def initialise_dico(self):
+ None
+
+ def initialise(self):
+ self.initialise_dico()
+ print("i'm initialised")
\ No newline at end of file
diff --git a/src/workflow.py b/src/workflow.py
new file mode 100644
index 0000000..5f700ae
--- /dev/null
+++ b/src/workflow.py
@@ -0,0 +1,25 @@
+
+from .nextflow_file import Nextflow_File
+from .ro_crate import RO_Crate
+
+
+class Workflow:
+ def __init__(self, file, duplicate=False, display_info=True):
+ self.nextflow_file = Nextflow_File(file, duplicate = duplicate, display_info = display_info)
+ self.rocrate = None
+
+
+ def get_file_address(self):
+ return self.nextflow_file.get_file_address()
+
+ def initialise_rocrate(self):
+ self.rocrate = RO_Crate(self)
+ self.rocrate.initialise()
+
+
+ def initialise(self):
+ self.nextflow_file.initialise()
+ self.initialise_rocrate()
+
+ def generate_all_graphs(self, render_graphs = True):
+ self.nextflow_file.generate_all_graphs(render_graphs = render_graphs)
--
GitLab