fix(semantic-query): symbols table never populated — index_size always 0

[Wolfvin]

Root Cause (verified)

semantic_search_engine.py reads from the symbols table:
SELECT COUNT() FROM symbols → 0 (always empty after scan)
SELECT COUNT() FROM graph_nodes → 652 (data lives here)

scan populates graph_nodes/graph_edges but nothing populates symbols.
semantic-query reads the wrong table → index_size: 0, returned: 0 on every query.

Fix (Option A — recommended)

Change _load_symbols_from_db in semantic_search_engine.py to read from graph_nodes:
node_id → id, name → name, file → file_path, node_type → kind, line → line_start

No schema change needed.

Definition of Done

• semantic-query "google auth session flow" returns index_size > 0 and returned > 0 on a scanned workspace
• Results ranked by TF-IDF relevance
• index_size ≈ number of functions/classes in workspace

[Wolfvin]

Mengerjakan issue ini -- 2026-07-03T14:37:02Z

[Wolfvin]

PR siap review

PR: #191 -- fix(semantic-query): read symbols from graph_nodes, not empty symbols table (closes #188)
Branch: fix/issue-188-semantic-query-graph-nodes
Summary: Changed _load_symbols_from_db to read from graph_nodes (which scan actually populates via graph_model.populate_graph_tables) instead of the legacy symbols table that the scan flow never writes to. Column mapping follows the issue spec (node_id→id, name→name, file→file_path, node_type→kind, line→line_start, extra_json→extra_json). No schema change. Verified DoD: semantic-query "google auth session flow" on a 6-node scanned workspace returns index_size=22, returned=4, top result GoogleAuthSession score 0.6988.
Tests: 31 passed (test_semantic_search_engine.py), 0 failed; 341 passed / 17 skipped / 0 failed across related modules (registry, graph, command-registry, query-graph, affected, CLI, command-count, codelens).