[ST-2214] Add more Information to Sentry Error + Solve absolute Url Bug

Author: glanzel

apps/projects/utils.py

@@ -14,6 +14,7 @@
 )
 from apps.summarization.export_utils.core import generate_full_export
 from apps.summarization.pydantic_models import ProjectSummaryResponse
+from apps.summarization.sentry_tags import set_sentry_project_tags
 from apps.summarization.services import AIService
 
 logger = logging.getLogger(__name__)
@@ -36,6 +37,8 @@ def generate_project_summary(
     Raises:
         Exception: Re-raises any exception from the AI service (caller handles display/retry).
     """
+    set_sentry_project_tags(project)
+
     export_data = generate_full_export(project)
 
     if request is not None or base_url:
@@ -46,7 +49,8 @@ def generate_project_summary(
             try:
                 service = AIService()
                 document_response = service.request_vision_dict(
-                    documents_dict=documents_dict
+                    documents_dict=documents_dict,
+                    project=project,
                 )
                 integrate_document_summaries(
                     export_data,

apps/summarization/export_utils/attachments/handlers.py

@@ -1,5 +1,11 @@
 def _make_absolute_url(attachment_url, request=None, base_url=None):
     """Build absolute URL from attachment_url using request or base_url."""
+    if not attachment_url:
+        return None
+
+    if attachment_url.startswith(("http://", "https://")):
+        return attachment_url
+
     if request is not None:
         return request.build_absolute_uri(attachment_url)
     if base_url:

apps/summarization/providers.py

@@ -18,6 +18,7 @@
 
 from .llm_json import parse_structured_llm_json
 from .pydantic_models import DocumentSummaryResponse
+from .sentry_tags import ensure_sentry_project_tags
 
 logger = logging.getLogger(__name__)
 
@@ -196,6 +197,7 @@ def text_request(
         )
 
         try:
+            ensure_sentry_project_tags()
             result = agent.run_sync(request.prompt())
             response = result.output
 
@@ -301,6 +303,7 @@ def multimodal_request(
             user_content.append(ImageUrl(url=url))
 
         try:
+            ensure_sentry_project_tags()
             result = agent.run_sync(user_content)
             response = result.output
 

apps/summarization/sentry_tags.py

@@ -0,0 +1,24 @@
+"""Sentry tags for project-scoped summarization calls."""
+
+from __future__ import annotations
+
+from contextvars import ContextVar
+
+from sentry_sdk import set_tag
+
+_current_project: ContextVar = ContextVar("sentry_project", default=None)
+
+
+def set_sentry_project_tags(project) -> None:
+    """Bind project tags for the current context and Sentry scope."""
+    set_tag("project_id", project.id)
+    set_tag("project_slug", project.slug)
+    _current_project.set(project)
+
+
+def ensure_sentry_project_tags() -> None:
+    """Re-apply project tags before nested SDK calls (e.g. OpenAI client)."""
+    project = _current_project.get()
+    if project is not None:
+        set_tag("project_id", project.id)
+        set_tag("project_slug", project.slug)

apps/summarization/services.py

@@ -18,6 +18,7 @@
 from .pydantic_models import DocumentSummaryResponse
 from .pydantic_models import ProjectSummaryResponse
 from .pydantic_models import SummaryItem
+from .sentry_tags import set_sentry_project_tags
 from .utils import extract_text_from_document
 
 logger = logging.getLogger(__name__)
@@ -78,6 +79,7 @@ def project_summarize(
         - If allow_regeneration is False and a summary exists, always return the latest summary
           without generating a new one, even when the hash changed.
         """
+        set_sentry_project_tags(project)
         request = SummaryRequest(text=text, prompt=prompt)
         latest = self._get_latest_summary(project)
         text_hash = ProjectSummary.compute_hash(text)
@@ -185,16 +187,24 @@ def _save_to_cache(
             logger.info(f"Cached summary for project {project.id}")
 
     def request_vision_dict(
-        self, documents_dict: dict[str, str], prompt: str | None = None
+        self,
+        documents_dict: dict[str, str],
+        prompt: str | None = None,
+        project=None,
     ) -> DocumentSummaryResponse:
         """Process documents from dictionary format."""
         items = [DocumentInputItem(handle=h, url=u) for h, u in documents_dict.items()]
-        return self.request_vision(items, prompt)
+        return self.request_vision(items, prompt, project=project)
 
     def request_vision(
-        self, documents: list[DocumentInputItem], prompt: str | None = None
+        self,
+        documents: list[DocumentInputItem],
+        prompt: str | None = None,
+        project=None,
     ) -> DocumentSummaryResponse:
         """Process documents and images, return combined summaries."""
+        if project is not None:
+            set_sentry_project_tags(project)
         docs, images = self._split_documents(documents)
 
         results = []

tests/summarization/test_attachment_urls.py

@@ -0,0 +1,56 @@
+"""Tests for document attachment URL handling in summarization exports."""
+
+from apps.summarization.export_utils.attachments.handlers import _make_absolute_url
+from apps.summarization.export_utils.attachments.handlers import (
+    collect_document_attachments,
+)
+
+
+def test_make_absolute_url_keeps_existing_absolute_url_with_base_url():
+    media_url = (
+        "https://roots-media-prod.liqd.net/uploads/Vanita/2025/08/04/"
+        "Innovationswerkstatt%201_Ablaufplanung_Stand%2001-08-2025.pdf"
+    )
+
+    assert (
+        _make_absolute_url(
+            media_url,
+            base_url="https://beteiligung-roots.org",
+        )
+        == media_url
+    )
+
+
+def test_make_absolute_url_prefixes_relative_path_with_base_url():
+    assert (
+        _make_absolute_url(
+            "/uploads/example.pdf",
+            base_url="https://beteiligung-roots.org",
+        )
+        == "https://beteiligung-roots.org/uploads/example.pdf"
+    )
+
+
+def test_collect_document_attachments_keeps_absolute_media_urls():
+    export_data = {
+        "project": {
+            "information_attachments": [
+                "https://roots-media-prod.liqd.net/uploads/example.pdf",
+            ],
+            "result_attachments": [],
+        }
+    }
+
+    documents_dict, handle_to_source = collect_document_attachments(
+        export_data,
+        base_url="https://beteiligung-roots.org",
+    )
+
+    assert documents_dict == {
+        "project_information_attachment_0": (
+            "https://roots-media-prod.liqd.net/uploads/example.pdf"
+        )
+    }
+    assert handle_to_source == {
+        "project_information_attachment_0": "project_information",
+    }