add "section" metadata to nodes

TheDataGuild · Sep 27, 2023 · 1168cde · 1168cde
1 parent 195ce33
commit 1168cde
Showing 1 changed file with 6 additions and 1 deletion.
diff --git a/mind_palace/docs.py b/mind_palace/docs.py
@@ -28,13 +28,17 @@ def title(xml, doc_id):
     return TextNode(
         text=xml.header.title,
         id_=f"{doc_id}-title",
+        metadata={"section": "title"},
+        excluded_embed_metadata_keys=["section"],
     )
 
 
 def abstract(xml, doc_id):
     return TextNode(
         text=xml.abstract,
         id_=f"{doc_id}-abstract",
+        metadata={"section": "abstract"},
+        excluded_embed_metadata_keys=["section"],
     )
 
 
@@ -60,7 +64,8 @@ def body(xml, doc_id):
     return [
         TextNode(
             text=line,
-            metadata={"paragraph_number": index + 1},
+            metadata={"section": "body", "paragraph_number": index + 1},
+            excluded_embed_metadata_keys=["section"],
             id_=f"{doc_id}-body-paragraph-{index}",
         )
         for index, line in enumerate(xml.body.split("\n"))