refining of outputs

Author: torri

.github/workflows/check_licenses.yml

@@ -131,7 +131,7 @@ jobs:
                 echo "modules_results.json file exists, trying to fetch the license..."
                 ml Python-bundle-PyPI/2023.06-GCCcore-12.3.0 BeautifulSoup/4.12.2-GCCcore-12.3.0 PyYAML/6.0-GCCcore-12.3.0
                 python licenses/parse_licenses.py modules_results.json licenses/licenses.yml
-                cat licenses_aux.yaml
+                cat temporal_print.yaml
               else
                 echo "modules_results.json file does not exist, skipping license fetch."
               fi

licenses/licenses_transformer.py

@@ -0,0 +1,39 @@
+#!/usr/bin/env python3
+
+import yaml
+from collections import defaultdict
+
+infile = "licenses.yaml"
+outfile = "licenses.yml"
+
+with open(infile, "r") as f:
+    data = yaml.safe_load(f)
+
+# Nested dict: {software: {version: info}}
+result = {}
+
+for full_key, info in data.items():
+    # Example full_key: "ALL/0.9.2-foss-2023a"
+    try:
+        software, rest = full_key.split("/", 1)
+    except ValueError:
+        # If no '/', skip or handle differently
+        # For now just continue
+        print(f"Skipping key without '/': {full_key}")
+        continue
+
+    # Split the version from the toolchain suffix
+    # "0.9.2-foss-2023a" -> version="0.9.2"
+    version = rest.split("-", 1)[0]
+
+    # Ensure nested structure exists
+    if software not in result:
+        result[software] = {}
+    if version in result[software]:
+        # Optional: warn if we’re overwriting same software/version
+        print(f"Warning: duplicate entry for {software} {version}, overwriting.")
+
+    result[software][version] = info
+
+with open(outfile, "w") as f:
+    yaml.safe_dump(result, f, sort_keys=True)

licenses/parse_licenses.py

@@ -241,6 +241,9 @@ def process_modules_for_licenses(modules_file):
 
 def save_license_results(results, output_file="licenses_aux.yaml",licenses_original = os.sys.argv[2]):
     """Saves license information to a JSON file."""
+    with open("temporal_print.yaml", "w") as f:
+        yaml.dump(results, f, default_flow_style=False, sort_keys=True)  #Fast dump of what we have to print in the workflow
+
     full_data = {}
     with open(licenses_original, 'r') as f:
         full_data = yaml.safe_load(f)
@@ -250,7 +253,8 @@ def save_license_results(results, output_file="licenses_aux.yaml",licenses_origi
             full_data[software_name] = {}
 
         for version, details in versions_data.items():      
-            full_data[software_name][version] = details         #Add/replace the details of modules found in the new licenses data dictionary
+            if version not in full_data[software_name]: 
+                full_data[software_name][version] = details         #Add/replace the details of modules found in the new licenses data dictionary
 
     with open(output_file, "w") as f:
         yaml.dump(full_data, f, default_flow_style=False, sort_keys=True)  #Export data dictionary as licenses_aux.yaml file