fix(reindex): use get_context_type_for_uri for correct memory/skill classification

[evaldass]

Problem

During reindex (via /api/v1/content/reindex), all memory files under viking://user/memories/ and viking://agent/*/memories/ are incorrectly tagged with context_type = "resource" instead of "memory".

This causes downstream consumers (e.g. OpenClaw's auto-recall plugin) that filter by context_type == "memory" to miss all reindexed memories.

Root Cause

Two bugs in the reindex code path:

Bug 1: summarizer.py (line 61)

# Broken — 'viking://memory/' never matches real URIs
if uri.startswith("viking://memory/"):
    context_type = "memory"

Real memory URIs look like viking://user/memories/entities/mem_xxx.md or viking://agent/<id>/memories/cases/mem_xxx.md — none start with viking://memory/.

Bug 2: semantic_dag.py (lines 443, 563)

SemanticDagExecutor receives context_type from the root URI and propagates it to all child nodes via self._context_type. When reindexing from viking:// (root), the root has no /memories substring, so context_type defaults to "resource" for everything — including actual memories.

Fix

Both files now use the existing get_context_type_for_uri() from core/directories.py, which already handles all URI patterns correctly:

def get_context_type_for_uri(uri: str) -> str:
    if "/memories" in uri:
        return "memory"
    elif "/resources" in uri:
        return "resource"
    elif "/skills" in uri:
        return "skill"
    elif uri.startswith("viking://session"):
        return "memory"
    return "resource"

This is consistent with how DirectoryInitializer (line 269 in directories.py) already determines context types.

Changes

File	Change
summarizer.py	Replace broken startswith checks with get_context_type_for_uri(uri)
semantic_dag.py	Replace self._context_type with per-node get_context_type_for_uri(file_path/dir_uri)

Impact

• Auto-capture memories are unaffected (different code path via memory_updater.py)
• Reindex operations will now correctly classify memories, skills, and resources
• No breaking changes — reindex simply becomes correct

Fixes #1060

[CLAassistant]

All committers have signed the CLA.

[github-actions]

Failed to generate code suggestions for PR

[evaldass]

Closing as duplicate — PR #1061 by @muxiaomu001 already addresses the same two bugs (summarizer.py + semantic_dag.py context_type misclassification). Both PRs use the same approach (get_context_type_for_uri). Apologies for the overlap!