Add tech-enablement tag to keyword mapping and normalize related raw …

src/preprocessing/extract_geo_topic.py

@@ -40,10 +40,15 @@
     "Animal-Related",
     "Water",
     "Nature",
-    "Human/Civil Rights"  # FIX: War vorher vergessen
+    "Human/Civil Rights",  # FIX: War vorher vergessen
+    "Technology",
+    "Digital Transformation",
+    "Scientific Research and Technology Transfer",
+    "Innovation",
+    "tech-enablement"
 ]
 
-# 2. Die zentrale Mapping-Schmiede (Konsolidiert von ~24 auf 13 Hauptkategorien)
+# 2. Die zentrale Mapping-Schmiede (Konsolidiert von ~24 auf 14 Hauptkategorien inkl. tech-enablement)
 TAG_NORMALIZATION = {
     # Schreibweisen-Korrekturen
     "Arts and Culture": "Arts & Culture",
@@ -57,7 +62,14 @@
     "Employment/Workforce": "Socio-economic Development, Poverty",
     "Social/Human Services": "Socio-economic Development, Poverty",
     "Recreation, Sport & Well-being": "Health",
-    "Policy development": "Citizenship, Social Justice & Public Affairs"
+    "Policy development": "Citizenship, Social Justice & Public Affairs",
+
+    # Technology / Digitalization Mapping to tech-enablement
+    "Technology": "tech-enablement",
+    "Digital Transformation": "tech-enablement",
+    "Scientific Research and Technology Transfer": "tech-enablement",
+    "Innovation": "tech-enablement",
+    "tech-enablement": "tech-enablement"
 }
 
 # 3. Die Keyword-Kanten für den Freitext-Fallback (Exakt synchron zu den Normalisierungs-Targets)
@@ -116,6 +128,10 @@
     ],
     "Peace & Conflict Resolution": [
         r"peacebuilding", r"conflict sensitivity", r"peace work"
+    ],
+    "tech-enablement": [
+        r"tech\w*", r"technolog\w*", r"digital\w*", r"software", r"data science", r"artificial intelligence", r"ai\b",
+        r"\bit\b(?=\s+(systems?|services?|infrastructure|department|team|strategy|support))"
     ]
 }
 

src/tests/test_pipeline.py

@@ -125,6 +125,31 @@ def test_extract_tags_special_boundary(self):
         extract_tags(members)
         self.assertEqual(members[0]["tags_focus"], ["Human/Civil Rights"])
 
+    def test_extract_tags_tech_enablement(self):
+        # 1. Robust path: from official tags
+        members_robust = [
+            {
+                "name": "Org Tech Robust",
+                "philea_info": {
+                    "Programme Areas": "Digital Transformation\nEducation"
+                }
+            }
+        ]
+        extract_tags(members_robust)
+        self.assertEqual(sorted(members_robust[0]["tags_focus"]), ["Education", "tech-enablement"])
+
+        # 2. Fallback path: from free text keywords
+        members_fallback = [
+            {
+                "name": "Org Tech Fallback",
+                "philea_info": {
+                    "About": "Supporting the development of custom software and ai-driven solutions."
+                }
+            }
+        ]
+        extract_tags(members_fallback)
+        self.assertEqual(members_fallback[0]["tags_focus"], ["tech-enablement"])
+
     def test_extract_geo_taxonomy_matching(self):
         members = [
             {