fixes

benchmarks/visualizer/server.py

@@ -1,10 +1,11 @@
 """Benchmark Visualizer Web Service.
 
 A standalone web service for visualizing benchmark results.
-Currently supports LoComo benchmark visualization.
+Supports LoComo and LongMemEval benchmark visualization.
 """
 
 import json
+import logging
 from pathlib import Path
 from typing import Any
 
@@ -63,10 +64,41 @@ async def get_locomo_results(mode: str = "search") -> dict[str, Any]:
         raise HTTPException(status_code=500, detail=str(e))
 
 
+@app.get("/api/longmemeval")
+async def get_longmemeval_results() -> dict[str, Any]:
+    """Get LongMemEval benchmark results.
+
+    Returns pre-computed benchmark results from the longmemeval directory.
+    """
+    try:
+        results_path = BENCHMARKS_DIR / "longmemeval" / "results" / "benchmark_results.json"
+        print(f"chcking path {results_path}")
+        logging.info(f"chcking path {results_path}")
+
+        if not results_path.exists():
+            raise HTTPException(
+                status_code=404,
+                detail="Benchmark results not found. Please run the benchmark first."
+            )
+
+        with open(results_path) as f:
+            results = json.load(f)
+
+        return results
+    except json.JSONDecodeError as e:
+        raise HTTPException(
+            status_code=500,
+            detail=f"Failed to parse benchmark results: {str(e)}"
+        )
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+
 # Mount static files
 app.mount("/static", StaticFiles(directory=Path(__file__).parent / "static"), name="static")
 
 
 if __name__ == "__main__":
     import uvicorn
+    logging.basicConfig(level=logging.INFO)
     uvicorn.run(app, host="127.0.0.1", port=8001)

benchmarks/visualizer/static/index.html

@@ -17,6 +17,7 @@
             <option value="">-- Select a benchmark --</option>
             <option value="locomo-search">LoComo (search)</option>
             <option value="locomo-think">LoComo (think)</option>
+            <option value="longmemeval">LongMemEval</option>
         </select>
     </div>
 

benchmarks/visualizer/static/js/app.js

@@ -22,6 +22,8 @@ function selectBenchmark() {
         loadLocomoResults('search');
     } else if (currentBenchmark === 'locomo-think') {
         loadLocomoResults('think');
+    } else if (currentBenchmark === 'longmemeval') {
+        loadLongMemEvalResults();
     }
 }
 
@@ -369,3 +371,259 @@ function filterAnswers() {
         }
     });
 }
+
+// LongMemEval functions
+async function loadLongMemEvalResults() {
+    try {
+        const response = await fetch('/api/longmemeval');
+
+        if (!response.ok) {
+            const errorData = await response.json();
+            document.getElementById('benchmark-content').innerHTML = `
+                <div class="error-message">
+                    <h3>⚠️ Benchmark Results Not Found</h3>
+                    <p>${errorData.detail || 'The requested benchmark results are not available.'}</p>
+                    <p><strong>To generate results:</strong></p>
+                    <pre style="background: #f5f5f5; padding: 10px; border-radius: 4px; overflow-x: auto;">cd benchmarks/longmemeval
+uv run python longmemeval_benchmark.py</pre>
+                    <p style="margin-top: 15px; font-size: 14px; color: #666;">
+                        Once the benchmark completes, refresh this page and select "LongMemEval" again.
+                    </p>
+                </div>
+            `;
+            return;
+        }
+
+        benchmarkData = await response.json();
+        console.log('Loaded longmemeval data:', benchmarkData);
+        renderLongMemEvalResults();
+    } catch (e) {
+        console.error('Error loading longmemeval results:', e);
+        document.getElementById('benchmark-content').innerHTML = `
+            <div class="error-message">
+                <h3>❌ Error Loading Results</h3>
+                <p>${e.message}</p>
+                <p style="font-size: 12px; color: #666; margin-top: 10px;">Check the browser console for more details.</p>
+            </div>
+        `;
+    }
+}
+
+function renderLongMemEvalResults() {
+    if (!benchmarkData) return;
+
+    const content = document.getElementById('benchmark-content');
+
+    try {
+        const results = benchmarkData.item_results || [];
+        const numItems = benchmarkData.num_items || results.length;
+
+        console.log('Rendering longmemeval results:', { resultsCount: results.length, numItems });
+
+        // Calculate per-category statistics
+        const categoryStats = {};
+
+        // Aggregate across all items
+        let totalInvalid = 0;
+        results.forEach(item => {
+            if (item.metrics && item.metrics.category_stats) {
+                Object.entries(item.metrics.category_stats).forEach(([category, stats]) => {
+                    if (!categoryStats[category]) {
+                        categoryStats[category] = { name: category, correct: 0, total: 0, invalid: 0 };
+                    }
+                    categoryStats[category].correct += stats.correct || 0;
+                    categoryStats[category].total += stats.total || 0;
+                    categoryStats[category].invalid += stats.invalid || 0;
+                });
+            }
+            if (item.metrics && item.metrics.detailed_results) {
+                item.metrics.detailed_results.forEach(result => {
+                    if (result.is_invalid) {
+                        totalInvalid++;
+                    }
+                });
+            }
+        });
+
+        // Overall stats
+        const totalInvalidDisplay = totalInvalid > 0
+            ? `<div class="stat-item">
+                <div class="stat-label">Invalid Questions</div>
+                <div class="stat-value" style="color: #ff9800;">${totalInvalid}</div>
+            </div>`
+            : '';
+
+        const overallHtml = `
+            <div style="background: #f9f9f9; padding: 20px; border: 2px solid #333; border-radius: 8px; margin-bottom: 20px;">
+                <h3 style="margin-top: 0;">LongMemEval Benchmark - Overall Performance</h3>
+                ${totalInvalid > 0 ? `<div style="background: #fff3cd; border: 1px solid #ffc107; padding: 10px; border-radius: 4px; margin-bottom: 15px;">
+                    <strong>⚠️ Note:</strong> ${totalInvalid} question(s) marked as invalid due to errors (excluded from accuracy calculation)
+                </div>` : ''}
+                <div class="stats-grid">
+                    <div class="stat-item">
+                        <div class="stat-label">Overall Accuracy</div>
+                        <div class="stat-value">${benchmarkData.overall_accuracy.toFixed(2)}%</div>
+                        ${totalInvalid > 0 ? `<div style="font-size: 11px; color: #666; margin-top: 4px;">(${benchmarkData.total_correct} / ${benchmarkData.total_valid || (benchmarkData.total_questions - totalInvalid)})</div>` : ''}
+                    </div>
+                    <div class="stat-item">
+                        <div class="stat-label">Correct Answers</div>
+                        <div class="stat-value">${benchmarkData.total_correct} / ${benchmarkData.total_questions}</div>
+                    </div>
+                    ${totalInvalidDisplay}
+                    <div class="stat-item">
+                        <div class="stat-label">Items</div>
+                        <div class="stat-value">${numItems}</div>
+                    </div>
+                </div>
+
+                <h4 style="margin: 20px 0 10px 0; padding-top: 15px; border-top: 1px solid #ddd;">Accuracy by Category</h4>
+                <div class="stats-grid" style="grid-template-columns: repeat(auto-fit, minmax(250px, 1fr));">
+                    ${Object.values(categoryStats).map(cat => {
+                        const invalidCount = cat.invalid || 0;
+                        const validTotal = cat.total - invalidCount;
+                        const accuracy = validTotal > 0 ? ((cat.correct / validTotal) * 100).toFixed(1) : 0;
+                        const color = accuracy >= 70 ? '#43a047' : accuracy >= 50 ? '#ff9800' : '#e53935';
+                        const invalidNote = invalidCount > 0 ? ` <span style="color: #ff9800; font-size: 10px;">(${invalidCount} invalid)</span>` : '';
+                        return `
+                            <div class="stat-item">
+                                <div class="stat-label">${cat.name}</div>
+                                <div class="stat-value" style="color: ${color};">${accuracy}%</div>
+                                <div style="font-size: 11px; color: #666; margin-top: 4px;">${cat.correct} / ${cat.total}${invalidNote}</div>
+                            </div>
+                        `;
+                    }).join('')}
+                </div>
+            </div>
+        `;
+
+        // Filter controls
+        const filterHtml = `
+            <div style="margin-bottom: 20px; display: flex; gap: 10px; align-items: center;">
+                <label style="font-weight: bold;">Show:</label>
+                <label><input type="radio" name="answer-filter" value="all" checked onchange="filterAnswers()"> All Answers</label>
+                <label><input type="radio" name="answer-filter" value="incorrect" onchange="filterAnswers()"> ❌ Incorrect Only</label>
+                <label><input type="radio" name="answer-filter" value="correct" onchange="filterAnswers()"> ✅ Correct Only</label>
+                ${totalInvalid > 0 ? '<label><input type="radio" name="answer-filter" value="invalid" onchange="filterAnswers()"> ⚠️ Invalid Only</label>' : ''}
+            </div>
+        `;
+
+        // Build item sections
+        let itemsHtml = '';
+        results.forEach((item, idx) => {
+            const itemId = item.item_id || `item-${idx}`;
+            const accuracy = item.metrics.accuracy.toFixed(2);
+            const correctCount = item.metrics.correct;
+            const totalCount = item.metrics.total;
+
+            itemsHtml += `
+                <div style="margin-bottom: 30px; border: 2px solid #333; border-radius: 8px; overflow: hidden;">
+                    <div style="background: #f0f0f0; padding: 15px; border-bottom: 2px solid #333; cursor: pointer;" onclick="toggleConversation(${idx})">
+                        <h3 style="margin: 0; display: flex; justify-content: space-between; align-items: center;">
+                            <span>📊 ${itemId}</span>
+                            <span style="font-size: 18px; color: ${accuracy >= 70 ? '#43a047' : accuracy >= 50 ? '#ff9800' : '#e53935'};">
+                                ${accuracy}% (${correctCount}/${totalCount})
+                            </span>
+                        </h3>
+                    </div>
+                    <div id="conv-${idx}" style="display: none; padding: 20px;">
+                        ${renderLongMemEvalItemDetails(item)}
+                    </div>
+                </div>
+            `;
+        });
+
+        content.innerHTML = overallHtml + filterHtml + itemsHtml;
+    } catch (e) {
+        console.error('Error rendering LongMemEval results:', e);
+        content.innerHTML = `
+            <div class="error-message">
+                <strong>Error rendering results:</strong> ${e.message}<br>
+                <pre style="margin-top: 10px; font-size: 11px; overflow: auto;">${e.stack}</pre>
+            </div>
+        `;
+    }
+}
+
+function renderLongMemEvalItemDetails(item) {
+    if (!item || !item.metrics) {
+        return '<div style="padding: 20px; color: #666;">No metrics available</div>';
+    }
+
+    const results = item.metrics.detailed_results;
+    if (!results || !Array.isArray(results) || results.length === 0) {
+        return '<div style="padding: 20px; color: #666;">No detailed results available</div>';
+    }
+
+    let html = '<div class="qa-results">';
+
+    results.forEach((result, idx) => {
+        const isInvalid = result.is_invalid || false;
+        const isCorrect = result.is_correct;
+        const bgColor = isInvalid ? '#fff3cd' : (isCorrect ? '#e8f5e9' : '#ffebee');
+        const icon = isInvalid ? '⚠️' : (isCorrect ? '✅' : '❌');
+        const category = result.category || 'Unknown';
+
+        html += `
+            <div class="qa-item" data-correct="${isCorrect}" data-invalid="${isInvalid}" style="background: ${bgColor}; padding: 15px; margin-bottom: 15px; border: 1px solid #ddd; border-radius: 8px;">
+                <div style="display: flex; justify-content: space-between; align-items: flex-start; margin-bottom: 10px;">
+                    <div style="flex: 1;">
+                        <div style="font-weight: bold; font-size: 16px; margin-bottom: 8px;">
+                            ${icon} Question ${idx + 1} ${isInvalid ? '<span style="font-size: 12px; background: #ff9800; color: white; padding: 2px 8px; border-radius: 4px; margin-left: 8px;">INVALID</span>' : ''} <span style="font-size: 12px; background: #666; color: white; padding: 2px 8px; border-radius: 4px; margin-left: 8px;">${category}</span>
+                        </div>
+                        <div style="margin-bottom: 8px;">
+                            <b>Q:</b> ${result.question}
+                        </div>
+                    </div>
+                </div>
+
+                <div style="display: grid; grid-template-columns: 1fr 1fr; gap: 15px; margin-bottom: 10px;">
+                    <div>
+                        <div style="font-weight: bold; color: #43a047; margin-bottom: 4px;">✓ Correct Answer:</div>
+                        <div style="background: white; padding: 8px; border-radius: 4px; border: 1px solid #ccc;">
+                            ${result.correct_answer}
+                        </div>
+                    </div>
+                    <div>
+                        <div style="font-weight: bold; color: ${isCorrect ? '#43a047' : '#e53935'}; margin-bottom: 4px;">
+                            ${isCorrect ? '✓' : '✗'} Predicted Answer:
+                        </div>
+                        <div style="background: white; padding: 8px; border-radius: 4px; border: 1px solid #ccc;">
+                            ${result.predicted_answer}
+                        </div>
+                    </div>
+                </div>
+
+                <details style="margin-top: 10px;" ${isInvalid ? 'open' : ''}>
+                    <summary style="cursor: pointer; font-weight: bold; padding: 5px; background: rgba(255,255,255,0.5); border-radius: 4px;">
+                        📝 Show Reasoning & Retrieved Memories
+                    </summary>
+                    <div style="margin-top: 10px; padding: 10px; background: white; border-radius: 4px;">
+                        ${isInvalid ? `<div style="margin-bottom: 10px; padding: 10px; background: #ffebee; border-left: 4px solid #e53935; border-radius: 4px;">
+                            <b style="color: #c62828;">⚠️ Error:</b>
+                            <div style="margin-top: 4px; color: #333;">${result.error || 'Question marked as invalid'}</div>
+                        </div>` : ''}
+                        <div style="margin-bottom: 10px;">
+                            <b>System Reasoning:</b>
+                            <div style="padding: 8px; background: #f5f5f5; border-radius: 4px; margin-top: 4px;">
+                                ${result.reasoning || 'N/A'}
+                            </div>
+                        </div>
+                        <div style="margin-bottom: 10px;">
+                            <b>Judge Reasoning:</b>
+                            <div style="padding: 8px; background: #f5f5f5; border-radius: 4px; margin-top: 4px;">
+                                ${result.correctness_reasoning || 'N/A'}
+                            </div>
+                        </div>
+                        <div>
+                            <b>Retrieved Memories (${result.retrieved_memories ? result.retrieved_memories.length : 0}):</b>
+                            ${renderRetrievedMemories(result.retrieved_memories)}
+                        </div>
+                    </div>
+                </details>
+            </div>
+        `;
+    });
+
+    html += '</div>';
+    return html;
+}

memora/fact_extraction.py

@@ -12,6 +12,7 @@ from datetime import datetime
 from typing import List, Dict, Optional, Literal
 from openai import AsyncOpenAI
 from pydantic import BaseModel, Field
+from .llm_wrapper import OutputTooLongError
 
 
 class Entity(BaseModel):
@@ -439,6 +440,112 @@ Remember:
     raise last_error
 
 
+async def _extract_facts_with_auto_split(
+    chunk: str,
+    chunk_index: int,
+    total_chunks: int,
+    event_date: datetime,
+    context: str,
+    llm_config: 'LLMConfig'
+) -> List[Dict[str, str]]:
+    """
+    Extract facts from a chunk with automatic splitting if output exceeds token limits.
+
+    If the LLM output is too long (OutputTooLongError), this function automatically
+    splits the chunk in half and processes each half recursively.
+
+    Args:
+        chunk: Text chunk to process
+        chunk_index: Index of this chunk in the original list
+        total_chunks: Total number of original chunks
+        event_date: Reference date for temporal information
+        context: Context about the conversation/document
+        llm_config: LLM configuration to use
+
+    Returns:
+        List of fact dictionaries extracted from the chunk (possibly from sub-chunks)
+    """
+    import logging
+    logger = logging.getLogger(__name__)
+
+    try:
+        # Try to extract facts from the full chunk
+        return await _extract_facts_from_chunk(
+            chunk=chunk,
+            chunk_index=chunk_index,
+            total_chunks=total_chunks,
+            event_date=event_date,
+            context=context,
+            llm_config=llm_config
+        )
+    except OutputTooLongError as e:
+        # Output exceeded token limits - split the chunk in half and retry
+        logger.warning(
+            f"Output too long for chunk {chunk_index + 1}/{total_chunks} "
+            f"({len(chunk)} chars). Splitting in half and retrying..."
+        )
+
+        # Split at the midpoint, preferring sentence boundaries
+        mid_point = len(chunk) // 2
+
+        # Try to find a sentence boundary near the midpoint
+        # Look for ". ", "! ", "? " within 20% of midpoint
+        search_range = int(len(chunk) * 0.2)
+        search_start = max(0, mid_point - search_range)
+        search_end = min(len(chunk), mid_point + search_range)
+
+        sentence_endings = ['. ', '! ', '? ', '\n\n']
+        best_split = mid_point
+
+        for ending in sentence_endings:
+            pos = chunk.rfind(ending, search_start, search_end)
+            if pos != -1:
+                best_split = pos + len(ending)
+                break
+
+        # Split the chunk
+        first_half = chunk[:best_split].strip()
+        second_half = chunk[best_split:].strip()
+
+        logger.info(
+            f"Split chunk {chunk_index + 1} into two sub-chunks: "
+            f"{len(first_half)} chars and {len(second_half)} chars"
+        )
+
+        # Process both halves recursively (in parallel)
+        sub_tasks = [
+            _extract_facts_with_auto_split(
+                chunk=first_half,
+                chunk_index=chunk_index,
+                total_chunks=total_chunks,
+                event_date=event_date,
+                context=context,
+                llm_config=llm_config
+            ),
+            _extract_facts_with_auto_split(
+                chunk=second_half,
+                chunk_index=chunk_index,
+                total_chunks=total_chunks,
+                event_date=event_date,
+                context=context,
+                llm_config=llm_config
+            )
+        ]
+
+        sub_results = await asyncio.gather(*sub_tasks)
+
+        # Combine results from both halves
+        all_facts = []
+        for sub_result in sub_results:
+            all_facts.extend(sub_result)
+
+        logger.info(
+            f"Successfully extracted {len(all_facts)} facts from split chunk {chunk_index + 1}"
+        )
+
+        return all_facts
+
+
 async def extract_facts_from_text(
     text: str,
     event_date: datetime,
@@ -452,6 +559,9 @@ async def extract_facts_from_text(
     For large texts (>chunk_size chars), automatically chunks at sentence boundaries
     to avoid hitting output token limits. Processes ALL chunks in PARALLEL for speed.
 
+    If a chunk produces output that exceeds token limits (OutputTooLongError), it is
+    automatically split in half and retried recursively until successful.
+
     Args:
         text: Input text (conversation, article, etc.)
         event_date: Reference date for resolving relative times
@@ -473,7 +583,7 @@ async def extract_facts_from_text(
 
     chunks = chunk_text(text, max_chars=chunk_size)
     tasks = [
-        _extract_facts_from_chunk(
+        _extract_facts_with_auto_split(
             chunk=chunk,
             chunk_index=i,
             total_chunks=len(chunks),

memora/llm_wrapper.py

@@ -5,7 +5,7 @@ import os
 import time
 import asyncio
 from typing import Optional, Any, Dict, List
-from openai import AsyncOpenAI, RateLimitError, APIError, APIStatusError
+from openai import AsyncOpenAI, RateLimitError, APIError, APIStatusError, LengthFinishReasonError
 import logging
 
 logger = logging.getLogger(__name__)
@@ -14,6 +14,17 @@ logger = logging.getLogger(__name__)
 logging.getLogger("httpx").setLevel(logging.WARNING)
 
 
+class OutputTooLongError(Exception):
+    """
+    Bridge exception raised when LLM output exceeds token limits.
+
+    This wraps provider-specific errors (e.g., OpenAI's LengthFinishReasonError)
+    to allow callers to handle output length issues without depending on
+    provider-specific implementations.
+    """
+    pass
+
+
 class LLMConfig:
     """Configuration for an LLM provider."""
 
@@ -134,6 +145,13 @@ class LLMConfig:
 
                 return result
 
+            except LengthFinishReasonError as e:
+                # Output exceeded token limits - raise bridge exception for caller to handle
+                logger.warning(f"LLM output exceeded token limits: {str(e)}")
+                raise OutputTooLongError(
+                    f"LLM output exceeded token limits. Input may need to be split into smaller chunks."
+                ) from e
+
             except APIStatusError as e:
                 last_exception = e
                 if attempt < max_retries:

memora/temporal_semantic_memory.py

@@ -551,8 +551,7 @@ class TemporalSemanticMemory(
         # Calculate total character count
         total_chars = sum(len(item.get("content", "")) for item in contents)
 
-        # Threshold: 50,000 characters per sub-batch (roughly 12k tokens with average 4 chars/token)
+        CHARS_PER_BATCH = 500_000
-        CHARS_PER_BATCH = 50_000
 
         if total_chars > CHARS_PER_BATCH:
             # Split into smaller batches based on character count

memora/web/__init__.py

@@ -3,6 +3,7 @@ Web interface for memory system.
 
 Provides FastAPI app and visualization interface.
 """
-from .server import app, create_app
+from memora.api import create_app
+from .server import app
 
 __all__ = ["app", "create_app"]

memora/web/server.py

@@ -4,758 +4,12 @@ FastAPI server for memory graph visualization and API.
 Provides REST API endpoints for memory operations and serves
 the interactive visualization interface.
 """
-import asyncio
-from fastapi import FastAPI, HTTPException
-from fastapi.staticfiles import StaticFiles
-from fastapi.responses import FileResponse
-from pydantic import BaseModel
-from dotenv import load_dotenv
-import os
-import sys
-from pathlib import Path
-from typing import Optional, List, Dict, Any
-from datetime import datetime
-
-# Import from parent memora package
-from memora import TemporalSemanticMemory
-from memora.embeddings import Embeddings
-
 import logging
+import os
+import argparse
 
-
+from memora import TemporalSemanticMemory
-
+from memora.api import create_app
-# Environment variables are loaded by the shell script that calls this module
-# No need to load .env files here as they're sourced by start-server.sh
-
-
-def create_app(memory: TemporalSemanticMemory) -> FastAPI:
-    """
-    Create and configure the FastAPI application.
-
-    Args:
-        memory: TemporalSemanticMemory instance (already initialized with required parameters)
-
-    Returns:
-        Configured FastAPI application
-    """
-    app = FastAPI(
-        title="Agent Memory API",
-        version="1.0.0",
-        description="""
-A temporal-semantic memory system for AI agents that stores, retrieves, and reasons over memories.
-
-## Features
-
-* **Batch Memory Storage**: Store multiple memories efficiently with automatic fact extraction
-* **Semantic Search**: Find relevant memories using natural language queries
-* **Fact Type Filtering**: Search across world facts, agent actions, and opinions separately
-* **Think Endpoint**: Generate contextual answers based on agent identity and memories
-* **Graph Visualization**: Interactive memory graph visualization
-* **Document Tracking**: Track and manage memory documents with upsert support
-
-## Architecture
-
-The system uses:
-- **Temporal Links**: Connect memories that are close in time
-- **Semantic Links**: Connect semantically similar memories
-- **Entity Links**: Connect memories that mention the same entities
-- **Spreading Activation**: Intelligent traversal for memory retrieval
-        """,
-        contact={
-            "name": "Memory System",
-        },
-        license_info={
-            "name": "Apache 2.0",
-            "url": "https://www.apache.org/licenses/LICENSE-2.0.html",
-        }
-    )
-
-    # Mount static files
-    app.mount("/static", StaticFiles(directory=str(Path(__file__).parent / "static")), name="static")
-
-    @app.on_event("startup")
-    async def startup_event():
-        """Initialize memory system on startup."""
-        await memory.initialize()
-        logging.info("Memory system initialized")
-
-    @app.on_event("shutdown")
-    async def shutdown_event():
-        """Cleanup memory system on shutdown."""
-        await memory.close()
-        logging.info("Memory system closed")
-
-    # Store memory instance on app for route handlers to access
-    app.state.memory = memory
-
-    # Register all routes
-    _register_routes(app)
-
-    return app
-
-
-class SearchRequest(BaseModel):
-    """Request model for search endpoint."""
-    query: str
-    agent_id: str = "default"
-    thinking_budget: int = 100
-    max_tokens: int = 4096
-    reranker: str = "heuristic"
-    trace: bool = False
-    fact_type: Optional[str] = None
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "query": "What did Alice say about machine learning?",
-                "agent_id": "user123",
-                "thinking_budget": 100,
-                "max_tokens": 4096,
-                "reranker": "heuristic",
-                "trace": True,
-                "fact_type": "world"
-            }
-        }
-
-
-class SearchResponse(BaseModel):
-    """Response model for search endpoints."""
-    results: List[Dict[str, Any]]
-    trace: Optional[Dict[str, Any]] = None
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "results": [
-                    {
-                        "text": "Alice works at Google on the AI team",
-                        "score": 0.95,
-                        "id": "123e4567-e89b-12d3-a456-426614174000"
-                    }
-                ],
-                "trace": {
-                    "query": "What did Alice say about machine learning?",
-                    "num_results": 1,
-                    "time_seconds": 0.123
-                }
-            }
-        }
-
-
-class MemoryItem(BaseModel):
-    """Single memory item for batch put."""
-    content: str
-    event_date: Optional[datetime] = None
-    context: Optional[str] = None
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "content": "Alice mentioned she's working on a new ML model",
-                "event_date": "2024-01-15T10:30:00Z",
-                "context": "team meeting"
-            }
-        }
-
-
-class BatchPutRequest(BaseModel):
-    """Request model for batch put endpoint."""
-    agent_id: str
-    items: List[MemoryItem]
-    document_id: Optional[str] = None
-    document_metadata: Optional[Dict[str, Any]] = None
-    upsert: bool = False
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "agent_id": "user123",
-                "items": [
-                    {
-                        "content": "Alice works at Google",
-                        "context": "work"
-                    },
-                    {
-                        "content": "Bob went hiking yesterday",
-                        "event_date": "2024-01-15T10:00:00Z"
-                    }
-                ],
-                "document_id": "conversation_123",
-                "upsert": False
-            }
-        }
-
-
-class BatchPutResponse(BaseModel):
-    """Response model for batch put endpoint."""
-    success: bool
-    message: str
-    agent_id: str
-    document_id: Optional[str] = None
-    items_count: int
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "success": True,
-                "message": "Successfully stored 2 memory items",
-                "agent_id": "user123",
-                "document_id": "conversation_123",
-                "items_count": 2
-            }
-        }
-
-
-class BatchPutAsyncResponse(BaseModel):
-    """Response model for async batch put endpoint."""
-    success: bool
-    message: str
-    agent_id: str
-    document_id: Optional[str] = None
-    items_count: int
-    queued: bool
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "success": True,
-                "message": "Batch put task queued for background processing",
-                "agent_id": "user123",
-                "document_id": "conversation_123",
-                "items_count": 2,
-                "queued": True
-            }
-        }
-
-
-class ThinkRequest(BaseModel):
-    """Request model for think endpoint."""
-    query: str
-    agent_id: str = "default"
-    thinking_budget: int = 50
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "query": "What do you think about artificial intelligence?",
-                "agent_id": "user123",
-                "thinking_budget": 50
-            }
-        }
-
-
-class OpinionItem(BaseModel):
-    """Model for an opinion with confidence score."""
-    text: str
-    confidence: float
-
-
-class ThinkResponse(BaseModel):
-    """Response model for think endpoint."""
-    text: str
-    based_on: Dict[str, List[Dict[str, Any]]]  # {"world": [...], "agent": [...], "opinion": [...]}
-    new_opinions: List[OpinionItem] = []  # List of newly formed opinions with confidence
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "text": "Based on my understanding, AI is a transformative technology...",
-                "based_on": {
-                    "world": [{"text": "AI is used in healthcare", "score": 0.9}],
-                    "agent": [{"text": "I discussed AI applications last week", "score": 0.85}],
-                    "opinion": [{"text": "I believe AI should be used ethically", "score": 0.8}]
-                },
-                "new_opinions": [
-                    {"text": "AI has great potential when used responsibly", "confidence": 0.95}
-                ]
-            }
-        }
-
-
-class AgentsResponse(BaseModel):
-    """Response model for agents list endpoint."""
-    agents: List[str]
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "agents": ["user123", "agent_alice", "agent_bob"]
-            }
-        }
-
-
-class GraphDataResponse(BaseModel):
-    """Response model for graph data endpoint."""
-    nodes: List[Dict[str, Any]]
-    edges: List[Dict[str, Any]]
-    table_rows: List[Dict[str, Any]]
-    total_units: int
-
-    class Config:
-        json_schema_extra = {
-            "example": {
-                "nodes": [
-                    {"id": "1", "label": "Alice works at Google", "type": "world"},
-                    {"id": "2", "label": "Bob went hiking", "type": "world"}
-                ],
-                "edges": [
-                    {"from": "1", "to": "2", "type": "semantic", "weight": 0.8}
-                ],
-                "table_rows": [
-                    {"id": "abc12345...", "text": "Alice works at Google", "context": "Work info", "date": "2024-01-15 10:30", "entities": "Alice (PERSON), Google (ORGANIZATION)"}
-                ],
-                "total_units": 2
-            }
-        }
-
-
-def _register_routes(app: FastAPI):
-    """Register all API routes on the given app instance."""
-
-    @app.get("/", include_in_schema=False)
-    async def index():
-        """Serve the visualization page."""
-        return FileResponse(str(Path(__file__).parent / "templates" / "index.html"))
-
-
-    @app.get(
-        "/api/graph",
-        response_model=GraphDataResponse,
-        tags=["Visualization"],
-        summary="Get memory graph data",
-        description="Retrieve graph data for visualization, optionally filtered by agent_id and fact_type (world/agent/opinion)"
-    )
-    async def api_graph(
-        agent_id: Optional[str] = None,
-        fact_type: Optional[str] = None
-    ):
-        """Get graph data from database, optionally filtered by agent_id and fact_type."""
-        try:
-            data = await app.state.memory.get_graph_data(agent_id, fact_type)
-            return data
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/graph: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.post(
-        "/api/search",
-        response_model=SearchResponse,
-        tags=["Search"],
-        summary="Search memory",
-        description="Search memory using semantic similarity and spreading activation. Optionally filter by fact_type (world, agent, opinion)"
-    )
-    async def api_search(request: SearchRequest):
-        """Run a search and return results with trace."""
-        try:
-            # Run search with tracing
-            results, trace = await app.state.memory.search_async(
-                agent_id=request.agent_id,
-                query=request.query,
-                thinking_budget=request.thinking_budget,
-                max_tokens=request.max_tokens,
-                enable_trace=request.trace,
-                reranker=request.reranker,
-                fact_type=request.fact_type
-            )
-
-            # Convert trace to dict
-            trace_dict = trace.to_dict() if trace else None
-
-            return SearchResponse(
-                results=results,
-                trace=trace_dict
-            )
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/search: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.post(
-        "/api/world_search",
-        response_model=SearchResponse,
-        tags=["Search"],
-        summary="Search world facts",
-        description="Search only world facts - general knowledge about people, places, events, and things that happen"
-    )
-    async def api_world_search(request: SearchRequest):
-        """Search only world facts (general knowledge about the world)."""
-        try:
-            # Run search with fact_type filter for 'world'
-            results, trace = await app.state.memory.search_async(
-                agent_id=request.agent_id,
-                query=request.query,
-                thinking_budget=request.thinking_budget,
-                max_tokens=request.max_tokens,
-                enable_trace=request.trace,
-                reranker=request.reranker,
-                fact_type='world'
-            )
-
-            # Convert trace to dict
-            trace_dict = trace.to_dict() if trace else None
-
-            return SearchResponse(
-                results=results,
-                trace=trace_dict
-            )
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/world_search: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.post(
-        "/api/agent_search",
-        response_model=SearchResponse,
-        tags=["Search"],
-        summary="Search agent action facts",
-        description="Search only agent facts - memories about what the AI agent did, actions taken, and tasks performed"
-    )
-    async def api_agent_search(request: SearchRequest):
-        """Search only agent facts (facts about what the agent did)."""
-        try:
-            # Run search with fact_type filter for 'agent'
-            results, trace = await app.state.memory.search_async(
-                agent_id=request.agent_id,
-                query=request.query,
-                thinking_budget=request.thinking_budget,
-                max_tokens=request.max_tokens,
-                enable_trace=request.trace,
-                reranker=request.reranker,
-                fact_type='agent'
-            )
-
-            # Convert trace to dict
-            trace_dict = trace.to_dict() if trace else None
-
-            return SearchResponse(
-                results=results,
-                trace=trace_dict
-            )
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/agent_search: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.post(
-        "/api/opinion_search",
-        response_model=SearchResponse,
-        tags=["Search"],
-        summary="Search agent opinions",
-        description="Search only opinion facts - the agent's formed beliefs, perspectives, and viewpoints"
-    )
-    async def api_opinion_search(request: SearchRequest):
-        """Search only opinion facts (agent's formed opinions and perspectives)."""
-        try:
-            # Run search with fact_type filter for 'opinion'
-            results, trace = await app.state.memory.search_async(
-                agent_id=request.agent_id,
-                query=request.query,
-                thinking_budget=request.thinking_budget,
-                max_tokens=request.max_tokens,
-                enable_trace=request.trace,
-                reranker=request.reranker,
-                fact_type='opinion'
-            )
-
-            # Convert trace to dict
-            trace_dict = trace.to_dict() if trace else None
-
-            return SearchResponse(
-                results=results,
-                trace=trace_dict
-            )
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/opinion_search: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.post(
-        "/api/think",
-        response_model=ThinkResponse,
-        tags=["Reasoning"],
-        summary="Think and generate answer",
-        description="""
-    Think and formulate an answer using agent identity, world facts, and opinions.
-
-    This endpoint:
-    1. Retrieves agent facts (agent's identity)
-    2. Retrieves world facts relevant to the query
-    3. Retrieves existing opinions (agent's perspectives)
-    4. Uses LLM to formulate a contextual answer
-    5. Extracts and stores any new opinions formed
-    6. Returns plain text answer, the facts used, and new opinions
-        """
-    )
-    async def api_think(request: ThinkRequest):
-        try:
-            # Use the memory system's think_async method
-            result = await app.state.memory.think_async(
-                agent_id=request.agent_id,
-                query=request.query,
-                thinking_budget=request.thinking_budget
-            )
-
-            return ThinkResponse(
-                text=result["text"],
-                based_on=result["based_on"],
-                new_opinions=result.get("new_opinions", [])
-            )
-
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/think: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.get(
-        "/api/agents",
-        response_model=AgentsResponse,
-        tags=["Management"],
-        summary="List all agents",
-        description="Get a list of all agent IDs that have stored memories in the system"
-    )
-    async def api_agents():
-        """Get list of available agents from database."""
-        try:
-            agent_list = await app.state.memory.list_agents()
-            return AgentsResponse(agents=agent_list)
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/agents: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-    @app.get(
-        "/api/stats/{agent_id}",
-        tags=["Memory Statistics"],
-        summary="Get memory statistics for an agent",
-        description="Get statistics about nodes and links for a specific agent"
-    )
-    async def api_stats(agent_id: str):
-        """Get statistics about memory nodes and links for an agent."""
-        try:
-            pool = await app.state.memory._get_pool()
-            async with pool.acquire() as conn:
-                # Get node counts by fact_type
-                node_stats = await conn.fetch(
-                    """
-                    SELECT fact_type, COUNT(*) as count
-                    FROM memory_units
-                    WHERE agent_id = $1
-                    GROUP BY fact_type
-                    """,
-                    agent_id
-                )
-
-                # Get link counts by link_type
-                link_stats = await conn.fetch(
-                    """
-                    SELECT ml.link_type, COUNT(*) as count
-                    FROM memory_links ml
-                    JOIN memory_units mu ON ml.from_unit_id = mu.id
-                    WHERE mu.agent_id = $1
-                    GROUP BY ml.link_type
-                    """,
-                    agent_id
-                )
-
-                # Format results
-                nodes_by_type = {row['fact_type']: row['count'] for row in node_stats}
-                links_by_type = {row['link_type']: row['count'] for row in link_stats}
-
-                total_nodes = sum(nodes_by_type.values())
-                total_links = sum(links_by_type.values())
-
-                return {
-                    "agent_id": agent_id,
-                    "total_nodes": total_nodes,
-                    "total_links": total_links,
-                    "nodes_by_type": nodes_by_type,
-                    "links_by_type": links_by_type
-                }
-
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/stats/{agent_id}: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-    @app.post(
-        "/api/memories/batch",
-        response_model=BatchPutResponse,
-        tags=["Memory Storage"],
-        summary="Store multiple memories",
-        description="""
-    Store multiple memory items in batch with automatic fact extraction.
-
-    Features:
-    - Efficient batch processing
-    - Automatic fact extraction from natural language
-    - Entity recognition and linking
-    - Document tracking with optional upsert
-    - Temporal and semantic linking
-
-    The system automatically:
-    1. Extracts semantic facts from the content
-    2. Generates embeddings
-    3. Deduplicates similar facts
-    4. Creates temporal, semantic, and entity links
-    5. Tracks document metadata
-        """
-    )
-    async def api_batch_put(request: BatchPutRequest):
-        try:
-            # Validate agent_id - prevent writing to reserved agents
-            RESERVED_AGENT_IDS = {"locomo"}
-            if request.agent_id in RESERVED_AGENT_IDS:
-                raise HTTPException(
-                    status_code=403,
-                    detail=f"Cannot write to reserved agent_id '{request.agent_id}'. Reserved agents: {', '.join(RESERVED_AGENT_IDS)}"
-                )
-
-            # Prepare contents for put_batch_async
-            contents = []
-            for item in request.items:
-                content_dict = {"content": item.content}
-                if item.event_date:
-                    content_dict["event_date"] = item.event_date
-                if item.context:
-                    content_dict["context"] = item.context
-                contents.append(content_dict)
-
-            # Call put_batch_async
-            result = await app.state.memory.put_batch_async(
-                agent_id=request.agent_id,
-                contents=contents,
-                document_id=request.document_id,
-                document_metadata=request.document_metadata,
-                upsert=request.upsert
-            )
-            logging.info(f"Batch put result: {result}")
-
-            return BatchPutResponse(
-                success=True,
-                message=f"Successfully stored {len(contents)} memory items",
-                agent_id=request.agent_id,
-                document_id=request.document_id,
-                items_count=len(contents)
-            )
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/memories/batch: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.post(
-        "/api/memories/batch_async",
-        response_model=BatchPutAsyncResponse,
-        tags=["Memory Storage"],
-        summary="Store multiple memories asynchronously",
-        description="""
-    Store multiple memory items in batch asynchronously using the task backend.
-
-    This endpoint returns immediately after queuing the task, without waiting for completion.
-    The actual processing happens in the background.
-
-    Features:
-    - Immediate response (non-blocking)
-    - Background processing via task queue
-    - Efficient batch processing
-    - Automatic fact extraction from natural language
-    - Entity recognition and linking
-    - Document tracking with optional upsert
-    - Temporal and semantic linking
-
-    The system automatically:
-    1. Queues the batch put task
-    2. Returns immediately with success=True, queued=True
-    3. Processes in background: extracts facts, generates embeddings, creates links
-        """
-    )
-    async def api_batch_put_async(request: BatchPutRequest):
-        try:
-            # Validate agent_id - prevent writing to reserved agents
-            RESERVED_AGENT_IDS = {"locomo"}
-            if request.agent_id in RESERVED_AGENT_IDS:
-                raise HTTPException(
-                    status_code=403,
-                    detail=f"Cannot write to reserved agent_id '{request.agent_id}'. Reserved agents: {', '.join(RESERVED_AGENT_IDS)}"
-                )
-
-            # Prepare contents for put_batch_async
-            contents = []
-            for item in request.items:
-                content_dict = {"content": item.content}
-                if item.event_date:
-                    content_dict["event_date"] = item.event_date
-                if item.context:
-                    content_dict["context"] = item.context
-                contents.append(content_dict)
-
-            # Submit task to background queue
-            await app.state.memory._task_backend.submit_task({
-                'type': 'batch_put',
-                'agent_id': request.agent_id,
-                'contents': contents,
-                'document_id': request.document_id,
-                'document_metadata': request.document_metadata,
-                'upsert': request.upsert
-            })
-
-            logging.info(f"Batch put task queued for agent_id={request.agent_id}, {len(contents)} items")
-
-            return BatchPutAsyncResponse(
-                success=True,
-                message=f"Batch put task queued for background processing ({len(contents)} items)",
-                agent_id=request.agent_id,
-                document_id=request.document_id,
-                items_count=len(contents),
-                queued=True
-            )
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/memories/batch_async: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
-    @app.delete(
-        "/api/memory/{unit_id}",
-        tags=["Memory Storage"],
-        summary="Delete a memory unit",
-        description="Delete a single memory unit and all its associated links (temporal, semantic, and entity links)"
-    )
-    async def api_delete_memory_unit(unit_id: str):
-        """Delete a memory unit and all its links."""
-        try:
-            result = await app.state.memory.delete_memory_unit(unit_id)
-
-            if not result["success"]:
-                raise HTTPException(status_code=404, detail=result["message"])
-
-            return result
-        except HTTPException:
-            raise
-        except Exception as e:
-            import traceback
-            error_detail = f"{str(e)}\n\nTraceback:\n{traceback.format_exc()}"
-            print(f"Error in /api/memory/{unit_id}: {error_detail}")
-            raise HTTPException(status_code=500, detail=str(e))
-
-
 
 
 # Create app at module level (required for uvicorn import string)
@@ -771,7 +25,7 @@ app = create_app(_memory)
 
 if __name__ == "__main__":
     import uvicorn
-    import argparse
+
     logging.basicConfig(level=logging.INFO)
 
     # Parse CLI arguments

scripts/benchmarks/start-visualizer.sh

@@ -9,5 +9,4 @@ echo "Server will be available at: http://localhost:8001"
 echo ""
 
 cd benchmarks/visualizer
-open http://localhost:8001
 uv run uvicorn server:app --reload --host 0.0.0.0 --port 8001

scripts/start-server.sh

@@ -82,5 +82,4 @@ if [[ ${#SERVER_ARGS[@]} -eq 0 ]]; then
   SERVER_ARGS=(--reload --host 0.0.0.0 --port 8080)
 fi
 
-open "http://localhost:${PORT}"
 uv run python -m memora.web.server "${SERVER_ARGS[@]}"