Skip to content

Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.

fix(docreader): preserve all modern XMind topic groups (#4266)

Rudy Mizrahi Celekli81aa65d3b9cf
diff --git a/docreader/parser/xmind_parser.py b/docreader/parser/xmind_parser.py
index 90ae3c00b..74d690d69 100644
--- a/docreader/parser/xmind_parser.py
+++ b/docreader/parser/xmind_parser.py
@@ -45,9 +45,13 @@ def _topic_from_json(value: object) -> _Topic | None:
topics: list[_Topic] = []
children = value.get("children")
if isinstance(children, dict):
- attached = children.get("attached")
- if isinstance(attached, list):
- for child in attached:
+ # All four groups contain recursive topics in the XMind JSON schema.
+ # Keep their text and notes, as the classic XML parser already does.
+ for group_name in ("attached", "detached", "summary", "callout"):
+ group = children.get(group_name)
+ if not isinstance(group, list):
+ continue
+ for child in group:
topic = _topic_from_json(child)
if topic is not None:
topics.append(topic)
diff --git a/docreader/tests/test_xmind_parser.py b/docreader/tests/test_xmind_parser.py
index 1d92e63e5..ea654329c 100644
--- a/docreader/tests/test_xmind_parser.py
+++ b/docreader/tests/test_xmind_parser.py
@@ -35,6 +35,160 @@ def _mark_entry_encrypted(payload: bytes) -> bytes:
class XMindParserModernTests(unittest.TestCase):
+ def test_preserves_all_json_topic_groups_matching_classic_archive(self):
+ # The official XMind topic schema defines attached, detached, summary,
+ # and callout arrays of recursive topics. Every fixture topic has the
+ # schema-required id; both supported archive formats carry this outline.
+ # https://github.com/xmindltd/xmind-model/blob/master/schemas/topic.json
+ groups = {
+ "attached": [
+ {
+ "id": "main",
+ "title": "Main branch",
+ "notes": {"plain": {"content": "Main note"}},
+ "children": {
+ "attached": [{"id": "main-task", "title": "Main task"}]
+ },
+ }
+ ],
+ "detached": [
+ {
+ "id": "idea",
+ "title": "Independent idea",
+ "notes": {"plain": {"content": "Detached note"}},
+ "children": {
+ "summary": [
+ {"id": "decision", "title": "Detached decision"}
+ ]
+ },
+ }
+ ],
+ "summary": [
+ {
+ "id": "summary",
+ "title": "Overall decision",
+ "notes": {"plain": {"content": "Summary note"}},
+ "children": {
+ "callout": [{"id": "detail", "title": "Summary detail"}]
+ },
+ }
+ ],
+ "callout": [
+ {
+ "id": "reminder",
+ "title": "Reminder",
+ "notes": {"plain": {"content": "Callout note"}},
+ "children": {
+ "detached": [{"id": "related", "title": "Related detail"}]
+ },
+ }
+ ],
+ }
+ modern = _modern_xmind_bytes(
+ [
+ {
+ "id": "sheet",
+ "title": "Plan",
+ "rootTopic": {
+ "id": "root",
+ "title": "Plan & Delivery",
+ "notes": {"plain": {"content": "Root note"}},
+ "children": groups,
+ },
+ }
+ ]
+ )
+ classic = _classic_xmind_bytes(
+ """<xmap-content xmlns="urn:xmind:xmap:xmlns:content:2.0">
+<sheet id="sheet"><title>Plan</title><topic id="root">
+<title>Plan &amp; Delivery</title><notes><plain>Root note</plain></notes>
+<children>
+<topics type="attached"><topic id="main"><title>Main branch</title>
+<notes><plain>Main note</plain></notes><children><topics type="attached">
+<topic id="main-task"><title>Main task</title></topic>
+</topics></children></topic></topics>
+<topics type="detached"><topic id="idea"><title>Independent idea</title>
+<notes><plain>Detached note</plain></notes><children><topics type="summary">
+<topic id="decision"><title>Detached decision</title></topic>
+</topics></children></topic></topics>
+<topics type="summary"><topic id="summary"><title>Overall decision</title>
+<notes><plain>Summary note</plain></notes><children><topics type="callout">
+<topic id="detail"><title>Summary detail</title></topic>
+</topics></children></topic></topics>
+<topics type="callout"><topic id="reminder"><title>Reminder</title>
+<notes><plain>Callout note</plain></notes><children><topics type="detached">
+<topic id="related"><title>Related detail</title></topic>
+</topics></children></topic></topics>
+</children></topic></sheet></xmap-content>"""
+ )
+ expected = (
+ "# Plan\n\n"
+ "- Plan & Delivery\n"
+ " > Root note\n"
+ " - Main branch\n"
+ " > Main note\n"
+ " - Main task\n"
+ " - Independent idea\n"
+ " > Detached note\n"
+ " - Detached decision\n"
+ " - Overall decision\n"
+ " > Summary note\n"
+ " - Summary detail\n"
+ " - Reminder\n"
+ " > Callout note\n"
+ " - Related detail"
+ )
+
+ for source_format, payload in (("json", modern), ("xml", classic)):
+ with self.subTest(source_format=source_format):
+ document = XMindParser(file_name="plan.xmind").parse(payload)
+ self.assertEqual(expected, document.content)
+ self.assertEqual(
+ source_format, document.metadata["xmind_content_format"]
+ )
+ self.assertEqual(1, document.metadata["sheet_count"])
+ self.assertEqual(9, document.metadata["topic_count"])
+ self.assertEqual(5, document.metadata["note_count"])
+
+ def test_promotes_each_nonattached_group_from_blank_root(self):
+ for group_name in ("detached", "summary", "callout"):
+ with self.subTest(group_name=group_name):
+ payload = _modern_xmind_bytes(
+ [
+ {
+ "id": "sheet",
+ "title": "Outline",
+ "rootTopic": {
+ "id": "root",
+ "title": " ",
+ "children": {
+ group_name: [
+ {
+ "id": "visible",
+ "title": "Visible",
+ "notes": {
+ "plain": {"content": "Retained note"}
+ },
+ "children": {
+ "attached": [
+ {"id": "nested", "title": "Nested"}
+ ]
+ },
+ }
+ ]
+ },
+ },
+ }
+ ]
+ )
+ document = XMindParser(file_name="outline.xmind").parse(payload)
+ self.assertEqual(
+ "# Outline\n\n- Visible\n > Retained note\n - Nested",
+ document.content,
+ )
+ self.assertEqual(2, document.metadata["topic_count"])
+ self.assertEqual(1, document.metadata["note_count"])
+
def test_finds_content_without_enumerating_unrelated_entries(self):
payload = _xmind_zip(
{