Add RAG rewrite, 60-session experiment scripts, and analysis tools

- RAG rewrite adapter and vector preference pipeline in personalized_llm
- 60-session experiment queue scripts (reflection, rag, rag_vector, rag_rewrite)
- Vector-preference correlation analysis and visualization scripts
- Local reward model batch processing improvements
- Updated CLAUDE.md with full experiment documentation and notes

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

1 files changed, 44 insertions, 0 deletions

diff --git a/collaborativeagents/scripts/queue_next_run.sh b/collaborativeagents/scripts/queue_next_run.sh
new file mode 100755
index 0000000..524f0e2
--- /dev/null
+++ b/collaborativeagents/scripts/queue_next_run.sh
@@ -0,0 +1,44 @@
+#!/bin/bash
+# Wait for current fullrun_4methods to finish, then start real-profile experiment
+
+LOG="/workspace/personalization-user-model/collaborativeagents/results/fullrun_4methods.log"
+SCRIPTS_DIR="/workspace/personalization-user-model/collaborativeagents/scripts"
+
+echo "[$(date)] Waiting for fullrun_4methods to complete..."
+
+while true; do
+    if grep -q "EXPERIMENT COMPLETE" "$LOG" 2>/dev/null; then
+        echo "[$(date)] fullrun_4methods completed!"
+        break
+    fi
+    # Check if process died without completing
+    if ! pgrep -f "fullrun_4methods" > /dev/null 2>&1; then
+        if ! grep -q "EXPERIMENT COMPLETE" "$LOG" 2>/dev/null; then
+            echo "[$(date)] WARNING: process died before completion. Starting next run anyway."
+            break
+        fi
+    fi
+    sleep 60
+done
+
+echo "[$(date)] Starting real-profile experiment: 4 methods x 100 profiles x 20 sessions"
+
+cd "$SCRIPTS_DIR"
+nohup python3 run_experiments.py \
+    --methods vanilla,reflection,rag,rag_vector \
+    --datasets math-hard \
+    --n-profiles 100 \
+    --n-sessions 20 \
+    --max-turns 8 \
+    --use-vllm \
+    --vllm-agent-url http://localhost:8003/v1 \
+    --vllm-user-url http://localhost:8004/v1 \
+    --parallel-profiles 100 \
+    --reward-mode llm_local \
+    --reward-vllm-url http://localhost:8005/v1 \
+    --profile-path ../data/complex_profiles_v2/profiles_200.jsonl \
+    --output-dir ../results/realprofile_4methods \
+    > ../results/realprofile_4methods.log 2>&1 &
+
+echo "[$(date)] Experiment launched with PID $!"
+echo "Log: /workspace/personalization-user-model/collaborativeagents/results/realprofile_4methods.log"