fix: preserve source task sections and explicit empty notes

Assistant: codex
Assistant-Model: gpt-6-astra
Assistant-Session: 01a07ff8-19d0-7820-b4d0-1353833cb7fc

api/services/forge_projection.py

@@ -30,7 +30,7 @@ from api.services.task_record_id_backfill import qualify_task_id
 
 _WORK_RECORD_NAMESPACE = uuid.UUID("a4058507-5c4a-5a00-ab06-fffa4fb46009")
 _TASK_BLOCK_RE = re.compile(r"```task\s*\n(.*?)\n```", re.DOTALL)
-_HEADING_RE = re.compile(r"^(#{1,4})\s+(.+?)$", re.MULTILINE)
+_HEADING_RE = re.compile(r"^(#{1,6})\s+(.+?)$", re.MULTILINE)
 DEFAULT_FORGE_BASE = "https://forgejo.coulomb.social/coulomb"
 
 
@@ -272,12 +272,23 @@ def _split_frontmatter(text: str) -> tuple[dict, str]:
     return (meta if isinstance(meta, dict) else {}), body
 
 
+def _clean_task_description(raw: str) -> str | None:
+    """Match the file-authoritative consistency parser's section boundaries."""
+    lines = raw.splitlines()
+    while lines and (not lines[0].strip() or lines[0].strip() in {"---", "***", "___"}):
+        lines.pop(0)
+    while lines and (not lines[-1].strip() or lines[-1].strip() in {"---", "***", "___"}):
+        lines.pop()
+    return "\n".join(lines).strip() or None
+
+
 def _parse_tasks(body: str, workplan_id: str) -> list[DerivedTask]:
     headings = [
-        (m.start(), m.group(2).strip()) for m in _HEADING_RE.finditer(body)
+        (m.start(), len(m.group(1)), m.group(2).strip()) for m in _HEADING_RE.finditer(body)
     ]
     out: list[DerivedTask] = []
-    for m in _TASK_BLOCK_RE.finditer(body):
+    task_matches = list(_TASK_BLOCK_RE.finditer(body))
+    for index, m in enumerate(task_matches):
         try:
             block = yaml.safe_load(m.group(1).strip()) or {}
         except yaml.YAMLError:
@@ -287,15 +298,21 @@ def _parse_tasks(body: str, workplan_id: str) -> list[DerivedTask]:
         rid = str(block.get("id") or "").strip()
         if not rid:
             continue
-        title = block.get("title")
-        if not title:
-            prev = [t for pos, t in headings if pos < m.start()]
-            title = prev[-1] if prev else None
-        following_headings = [pos for pos, _title in headings if pos > m.end()]
-        description_end = min(following_headings) if following_headings else len(body)
-        description = str(block.get("description") or "").strip()
-        if not description:
-            description = body[m.end() : description_end].strip()
+        prev = [(pos, level, text) for pos, level, text in headings if pos < m.start()]
+        heading = prev[-1] if prev else None
+        title = block.get("title") or (heading[2] if heading else None)
+        # Nested evidence belongs to this task. Stop at the next peer/ancestor
+        # heading or task block, as the ordinary consistency path does.
+        description_end = task_matches[index + 1].start() if index + 1 < len(task_matches) else len(body)
+        following_heading = next((
+            pos for pos, level, _text in headings
+            if pos > m.end() and (heading is None or level <= heading[1])
+        ), None)
+        if following_heading is not None:
+            description_end = min(description_end, following_heading)
+        description = _clean_task_description(body[m.end():description_end])
+        if description is None and block.get("description") is not None:
+            description = str(block["description"])
         out.append(
             DerivedTask(
                 # A bare `T01` is not an identifier: it is unique only within
@@ -315,16 +332,16 @@ def _parse_tasks(body: str, workplan_id: str) -> list[DerivedTask]:
                 title=title,
                 status=(str(block["status"]).strip() if block.get("status") else None),
                 priority=(str(block["priority"]).strip() if block.get("priority") else None),
-                description=description or None,
+                description=description,
                 needs_human=bool(block.get("needs_human", False)),
                 intervention_note=(
-                    str(block["intervention_note"]).strip()
-                    if block.get("intervention_note")
+                    str(block["intervention_note"])
+                    if block.get("intervention_note") is not None
                     else None
                 ),
                 blocking_reason=(
-                    str(block["blocking_reason"]).strip()
-                    if block.get("blocking_reason")
+                    str(block["blocking_reason"])
+                    if block.get("blocking_reason") is not None
                     else None
                 ),
             )