[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-81987":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":8,"htmlUrl":8,"language":9,"languages":8,"totalLinesOfCode":8,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":13,"subscribersCount":13,"size":13,"stars1d":13,"stars7d":14,"stars30d":15,"stars90d":13,"forks30d":13,"starsTrendScore":13,"compositeScore":16,"rankGlobal":8,"rankLanguage":8,"license":17,"archived":18,"fork":18,"defaultBranch":19,"hasWiki":20,"hasPages":18,"topics":21,"createdAt":8,"pushedAt":8,"updatedAt":22,"readmeContent":23,"aiSummary":24,"trendingCount":13,"starSnapshotCount":13,"syncStatus":25,"lastSyncTime":26,"discoverSource":27},81987,"agentsre-langchain","Ajay150313\u002Fagentsre-langchain","Ajay150313",null,"Python",53,30,35,0,9,10,4.47,"Other",false,"main",true,[],"2026-06-12 02:04:21","# agentsre-langchain\n\n**Semantic SLI monitoring for LangChain agents** — Track Decision Quality, Tool Efficiency, Escalations, and Queue Depth in production.\n\n[![PyPI version](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpypi-v0.1.0-blue.svg)](https:\u002F\u002Fpypi.org\u002Fproject\u002Fagentsre-langchain\u002F)\n[![Python](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.9%2B-blue.svg)]()\n[![License: MIT](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-green.svg)](LICENSE)\n\n---\n\n## The Problem\n\nYou're running LangChain agents in production.\n\nYour agent returns HTTP 200. All tool calls succeed. Every health check passes.\n\n**But it's making wrong decisions 30% of the time.**\n\nYour existing monitoring won't catch this until it causes business impact.\n\n---\n\n## The Solution\n\nTrack the four semantic SLIs that matter:\n\n| SLI | What it measures | Healthy | Alert |\n|-----|------------------|---------|-------|\n| **DQR** | Decision Quality Rate | >92% | \u003C85% |\n| **TIE** | Tool Invocation Efficiency | 1.0-1.2x | >1.5x |\n| **HER** | Human Escalation Rate | \u003C2% | >5% |\n| **AQDD** | Queue Depth Drift | \u003C20 | >50 |\n\n---\n\n## Install\n\n```bash\n# Basic installation\npip install agentsre-langchain\n\n# With agentsre integration\npip install agentsre-langchain agentsre\n```\n\n---\n\n## Quick Start\n\n```python\nfrom langchain.agents import AgentExecutor\nfrom agentsre_langchain import monitor_agent, MonitorConfig\n\n@monitor_agent(\n    agent_id=\"payment-router\",\n    task_class=\"payments\",\n    config=MonitorConfig(verbose=True, track_cost=True)\n)\ndef run_agent(query: str):\n    executor = AgentExecutor(agent=agent, tools=tools)\n    return executor.invoke({\"input\": query})\n\n# Now every execution is monitored\nresult = run_agent(\"Route this payment...\")\n\n# Get metrics\nfrom agentsre_langchain import get_metrics\nmetrics = get_metrics(\"payments\")\nprint(f\"DQR: {metrics['dqr']}%\")\nprint(f\"TIE: {metrics['tie']}x\")\nprint(f\"HER: {metrics['her']}%\")\nprint(f\"Cost: ${metrics['total_cost']:.4f}\")\n```\n\n---\n\n## Examples\n\n1. **Simple Agent** - `examples\u002F1_simple_agent.py`\n   - Basic monitoring with decorator\n\n2. **Multi-Tool Routing** - `examples\u002F2_multi_tool_agent.py`\n   - Track tool selection efficiency\n\n3. **ReAct Pattern** - `examples\u002F3_react_agent.py`\n   - Monitor reasoning + acting agents\n\n4. **With Memory** - `examples\u002F4_with_memory.py`\n   - Track conversation context overhead\n\n5. **Cost Optimization** - `examples\u002F5_cost_tracking.py`\n   - Monitor reliability AND cost together\n\n---\n\n## Integration with agentsre\n\nIf you have `agentsre` installed, metrics automatically flow through:\n\n```python\nfrom agentsre_langchain.integrations import integrate_with_agentsre\n\n# Automatic integration\nmetrics = get_metrics(\"payment_routing\")\nagentsre_results = integrate_with_agentsre(\n    agent_id=\"payment-router\",\n    task_class=\"payment_routing\",\n    metrics=metrics\n)\n```\n\n---\n\n## Configuration\n\n```python\nfrom agentsre_langchain import MonitorConfig\n\nconfig = MonitorConfig(\n    track_tokens=True,           # Track input\u002Foutput tokens\n    track_decisions=True,        # Track decision quality\n    track_escalations=True,      # Track human escalations\n    track_cost=True,             # Track API costs\n    alert_on_breach=True,        # Alert when SLI breaches\n    dqr_threshold=85.0,          # DQR breach threshold\n    tie_threshold=1.5,           # TIE breach threshold\n    her_threshold=5.0,           # HER breach threshold\n    verbose=False,               # Log metrics\n)\n\n@monitor_agent(\"my-agent\", \"task_type\", config=config)\ndef my_agent(query: str):\n    ...\n```\n\n---\n\n## How It Works\n\n### 1. Decorator Captures Execution\n```python\n@monitor_agent(agent_id=\"my-agent\", task_class=\"my_tasks\")\ndef agent_function(query):\n    # Your LangChain agent code\n    return result\n```\n\n### 2. Metrics Extracted\n- Confidence score (from agent output)\n- Tool calls (how many tools invoked)\n- Tokens (input\u002Foutput tracking)\n- Cost (API call pricing)\n\n### 3. SLIs Calculated\n- **DQR**: % of high-confidence decisions\n- **TIE**: Tool calls vs baseline\n- **HER**: % of failed executions\n- **AQDD**: Pending items in queue\n\n### 4. Results Available\n```python\nmetrics = get_metrics(\"task_class\")\n# {\n#   \"executions\": 100,\n#   \"dqr\": 92.5,\n#   \"tie\": 1.2,\n#   \"her\": 2.1,\n#   \"total_cost\": 2.45,\n#   \"avg_cost_per_execution\": 0.0245\n# }\n```\n\n---\n\n## SLO Targets (Starting Points)\n\n| Environment | DQR | TIE | HER | AQDD |\n|-------------|-----|-----|-----|------|\n| Development | >75% | \u003C2.0x | \u003C10% | \u003C50 |\n| Staging | >85% | \u003C1.5x | \u003C5% | \u003C20 |\n| Production | >92% | \u003C1.2x | \u003C2% | \u003C10 |\n\n**Rule:** Run 30-day observation window before committing to SLO targets.\n\n---\n\n## Contributing\n\nWe welcome contributions! See [CONTRIBUTING.md](CONTRIBUTING.md)\n\n---\n\n## License\n\nMIT © [Ajay Devineni](https:\u002F\u002Flinkedin.com\u002Fin\u002Fajay-devineni)\n\n---\n\n**If this helps you instrument your agents, a ⭐ means a lot.**\n","agentsre-langchain 是一个用于监控 LangChain 代理在生产环境中决策质量、工具效率、升级率和队列深度的 Python 库。其核心功能包括通过四个关键指标（决策质量率 DQR、工具调用效率 TIE、人工升级率 HER 和队列深度漂移 AQDD）来评估代理的表现，并支持与 agentsre 的集成以进一步分析数据。该库适合那些已经在使用 LangChain 代理并希望提高系统可靠性和性能的企业或开发者，尤其是在需要确保 AI 决策准确性和成本效益的应用场景中。通过简单的装饰器模式即可快速上手，实现对不同复杂度任务的有效监控。",2,"2026-06-11 04:07:23","CREATED_QUERY"]