Merge pull request #2597 from IFRCGo/feature/fix-surge-alert-csv-export

Surge Alert CSV fix – same columns across paginated requests

Author: szabozoltan69

notifications/drf_views.py

@@ -13,6 +13,7 @@
 from .models import Subscription, SurgeAlert
 from .serializers import (  # UnauthenticatedSurgeAlertSerializer,
     SubscriptionSerializer,
+    SurgeAlertCsvSerializer,
     SurgeAlertSerializer,
 )
 
@@ -80,6 +81,9 @@ class SurgeAlertViewset(viewsets.ReadOnlyModelViewSet):
     )  # for /docs
 
     def get_serializer_class(self):
+        # Use CSV-friendly serializer for CSV format to ensure consistent column count across pages
+        if self.request.query_params.get("format") == "csv":
+            return SurgeAlertCsvSerializer
         # if self.request.user.is_authenticated:
         #     return SurgeAlertSerializer
         # return UnauthenticatedSurgeAlertSerializer

notifications/serializers.py

@@ -46,6 +46,185 @@ class Meta:
         )
 
 
+class SurgeAlertCsvSerializer(ModelSerializer):
+    """CSV-friendly serializer with flattened fields for consistent column count across pages"""
+
+    atype_display = serializers.CharField(source="get_atype_display", read_only=True)
+    molnix_status_display = serializers.CharField(source="get_molnix_status_display", read_only=True)
+    category_display = serializers.CharField(source="get_category_display", read_only=True)
+
+    # Flattened country fields
+    country_id = serializers.IntegerField(source="country.id", read_only=True)
+    country_name = serializers.CharField(source="country.name", read_only=True)
+    country_iso = serializers.CharField(source="country.iso", read_only=True)
+    country_iso3 = serializers.CharField(source="country.iso3", read_only=True)
+    country_society_name = serializers.CharField(source="country.society_name", read_only=True)
+    country_region = serializers.IntegerField(source="country.region.id", read_only=True)
+
+    # Flattened event fields
+    event_id = serializers.IntegerField(source="event.id", read_only=True)
+    event_name = serializers.CharField(source="event.name", read_only=True)
+    event_dtype_id = serializers.IntegerField(source="event.dtype.id", read_only=True)
+    event_dtype_name = serializers.CharField(source="event.dtype.name", read_only=True)
+    event_glide = serializers.CharField(source="event.glide", read_only=True)
+    event_disaster_start_date = serializers.DateTimeField(source="event.disaster_start_date", read_only=True)
+    event_auto_generated = serializers.BooleanField(source="event.auto_generated", read_only=True)
+    event_ifrc_severity_level = serializers.IntegerField(source="event.ifrc_severity_level", read_only=True)
+    event_num_affected = serializers.IntegerField(source="event.num_affected", read_only=True)
+    event_parent_event = serializers.IntegerField(source="event.parent_event.id", read_only=True)
+    event_summary = serializers.SerializerMethodField()
+    event_tab_one_title = serializers.CharField(source="event.tab_one_title", read_only=True)
+    event_tab_two_title = serializers.CharField(source="event.tab_two_title", read_only=True)
+    event_tab_three_title = serializers.CharField(source="event.tab_three_title", read_only=True)
+    event_translation_module_original_language = serializers.CharField(
+        source="event.translation_module_original_language", read_only=True
+    )
+    event_updated_at = serializers.DateTimeField(source="event.updated_at", read_only=True)
+
+    # Comma-separated lists for many-to-many relationships
+    event_countries = serializers.SerializerMethodField()
+    event_appeals = serializers.SerializerMethodField()
+    molnix_tags_names = serializers.SerializerMethodField()
+    molnix_tags_ids = serializers.SerializerMethodField()
+    molnix_tags_types = serializers.SerializerMethodField()
+
+    class Meta:
+        model = SurgeAlert
+        fields = (
+            "id",
+            "molnix_id",
+            "created_at",
+            "operation",
+            "message",
+            "deployment_needed",
+            "is_private",
+            "atype",
+            "atype_display",
+            "category",
+            "category_display",
+            "molnix_status",
+            "molnix_status_display",
+            "opens",
+            "closes",
+            "start",
+            "end",
+            # Country fields (flattened)
+            "country_id",
+            "country_name",
+            "country_iso",
+            "country_iso3",
+            "country_society_name",
+            "country_region",
+            # Event fields (flattened)
+            "event_id",
+            "event_name",
+            "event_dtype_id",
+            "event_dtype_name",
+            "event_glide",
+            "event_disaster_start_date",
+            "event_auto_generated",
+            "event_ifrc_severity_level",
+            "event_num_affected",
+            "event_parent_event",
+            "event_summary",
+            "event_tab_one_title",
+            "event_tab_two_title",
+            "event_tab_three_title",
+            "event_translation_module_original_language",
+            "event_updated_at",
+            "event_countries",
+            "event_appeals",
+            # Molnix tags (comma-separated)
+            "molnix_tags_names",
+            "molnix_tags_ids",
+            "molnix_tags_types",
+        )
+
+    @staticmethod
+    def get_event_summary(obj):
+        """Return HTML-stripped first 300 characters of event summary with empty lines removed"""
+        if obj.event and obj.event.summary:
+            import re
+            from html import unescape
+
+            text = obj.event.summary
+
+            # Remove Microsoft Office table styles and metadata (before HTML stripping)
+            # Match content between /* Style Definitions */ and the closing brace
+            text = re.sub(r"/\*\s*Style Definitions\s*\*/.*?}", "", text, flags=re.DOTALL)
+            # Remove other common MS Office artifacts
+            text = re.sub(r"Normal\s+0\s+\d+\s+false\s+false\s+false\s+[A-Z-]+(?:\s+[A-Z-]+)*", "", text)
+            text = re.sub(r"table\.MsoNormalTable\s*{[^}]*}", "", text, flags=re.DOTALL)
+
+            # Strip HTML tags
+            text = re.sub(r"<[^>]+>", "", text)
+            # Decode HTML entities (&eacute; -> é, etc.)
+            text = unescape(text)
+            # Remove lines that contain only whitespace or are empty
+            lines = [line.strip() for line in text.split("\n") if line.strip()]
+            # Remove lines that only contain "DISASTER OVERVIEW" or "Summary"
+            lines = [line for line in lines if line.strip().upper() != "DISASTER OVERVIEW" and line.strip().upper() != "SUMMARY"]
+            # Join remaining lines with newline
+            text = "\n".join(lines)
+            # Return first 300 characters
+            return text[:300]
+        return ""
+
+    @staticmethod
+    def get_event_countries(obj):
+        """Return structured list of countries with name, fdrs, iso, iso3 (separated by |)"""
+        if obj.event and obj.event.countries.exists():
+            country_data = []
+            for country in obj.event.countries.all():
+                # Format: name;fdrs;iso;iso3
+                parts = [country.name or "", country.fdrs or "", country.iso or "", country.iso3 or ""]
+                country_data.append(";".join(parts))
+            return " | ".join(country_data)
+        return ""
+
+    @staticmethod
+    def get_event_appeals(obj):
+        """Return structured list of appeals with all details (separated by |)
+        in a format code;amount_funded;amount_requested;atype;atype_display;start_date;end_date;
+        num_beneficiaries;sector;status_display
+        """
+        if obj.event:
+            appeals = obj.event.appeals.all()
+            if appeals:
+                appeal_data = []
+                for appeal in appeals:
+                    parts = [
+                        appeal.code or "",
+                        str(appeal.amount_funded) if appeal.amount_funded is not None else "",
+                        str(appeal.amount_requested) if appeal.amount_requested is not None else "",
+                        str(appeal.atype) if appeal.atype is not None else "",
+                        appeal.get_atype_display() if hasattr(appeal, "get_atype_display") else "",
+                        str(appeal.start_date) if appeal.start_date else "",
+                        str(appeal.end_date) if appeal.end_date else "",
+                        str(appeal.num_beneficiaries) if appeal.num_beneficiaries is not None else "",
+                        appeal.sector or "",
+                        appeal.get_status_display() if hasattr(appeal, "get_status_display") else "",
+                    ]
+                    appeal_data.append(";".join(parts))
+                return " | ".join(appeal_data)
+        return ""
+
+    @staticmethod
+    def get_molnix_tags_names(obj):
+        """Return comma-separated list of molnix tag names"""
+        return ", ".join([tag.name for tag in obj.molnix_tags.all()])
+
+    @staticmethod
+    def get_molnix_tags_ids(obj):
+        """Return comma-separated list of molnix tag IDs"""
+        return ", ".join([str(tag.molnix_id) for tag in obj.molnix_tags.all()])
+
+    @staticmethod
+    def get_molnix_tags_types(obj):
+        """Return comma-separated list of molnix tag types"""
+        return ", ".join([tag.tag_type for tag in obj.molnix_tags.all()])
+
+
 # class UnauthenticatedSurgeAlertSerializer(ModelSerializer):
 #     event = MiniEventSerializer()
 #     atype_display = serializers.CharField(source='get_atype_display', read_only=True)