From 7768c12680cd88a47336fecc335b860c50cbabd2 Mon Sep 17 00:00:00 2001
From: Zio Gabber <78922322+Gabrymi93@users.noreply.github.com>
Date: Wed, 13 May 2026 11:13:04 +0100
Subject: [PATCH] feat: source_id in DICleanDataset

---
 src/agent_context_builder/signals.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/src/agent_context_builder/signals.py b/src/agent_context_builder/signals.py
index a0834fb..809447c 100644
--- a/src/agent_context_builder/signals.py
+++ b/src/agent_context_builder/signals.py
@@ -116,6 +116,7 @@ class DICleanDataset:
     slug: str
     name: str
     stage: str
+    source_id: str = ""
     source: str = ""
     period: dict[str, Any] = field(default_factory=dict)
     location: dict[str, Any] = field(default_factory=dict)
@@ -230,6 +231,7 @@ def parse_di_clean_catalog(raw: str) -> DICleanCatalog:
                 slug=item.get("slug", ""),
                 name=item.get("name", item.get("slug", "")),
                 stage=item.get("stage", "incubating"),
+                source_id=item.get("source_id", ""),
                 source=item.get("source", ""),
                 period=item.get("period", {}),
                 location=item.get("location", {}),