apunkt/SurfSense
1
0
Fork 0
mirror of https://github.com/MODSetter/SurfSense.git synced 2026-06-08 20:25:19 +02:00
Code Issues Projects Releases Packages Wiki Activity

refact: follow the structure of document returned from retriever

Browse source
This commit is contained in:
CREDO23 2025-12-24 20:12:40 +02:00
parent deec8c5c6c
commit 3660b91e63
1 changed files with 22 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

27
surfsense_backend/app/tasks/chat/stream_new_chat.py
View file

@ -59,11 +59,28 @@ def format_mentioned_documents_as_context(documents: list[Document]) -> str:
"These documents are directly relevant to the query and should be prioritized as primary sources."
)
for i, doc in enumerate(documents, 1):
context_parts.append(
f"<document index='{i}' id='{doc.id}' title='{doc.title}' type='{doc.document_type.value}'>"
)
context_parts.append(f"<![CDATA[{doc.content}]]>")
context_parts.append("</document>")
# Prepare retriever-style structure
doc_metadata = doc.document_metadata if doc.document_metadata else {}
xml_doc = f"""
<document index='{i}'>
<document_id>{doc.id}</document_id>
<content><![CDATA[{doc.content}]]></content>
<score>1.0</score>
<chunks>
<chunk>
<chunk_id>{doc.id}-full</chunk_id>
<content><![CDATA[{doc.content}]]></content>
</chunk>
</chunks>
<document_info>
<id>{doc.id}</id>
<title><![CDATA[{doc.title}]]></title>
<document_type>{doc.document_type.value}</document_type>
<metadata><![CDATA[{json.dumps(doc_metadata)}]]></metadata>
</document_info>
<source>{doc.document_type.value}</source>
</document>"""
context_parts.append(xml_doc.strip())
context_parts.append("</mentioned_documents>")
return "\n".join(context_parts)