Examples (#146)

* examples

* examples

agentops/__init__.py

@@ -59,7 +59,7 @@ def start_session(tags: Optional[List[str]] = None, config: Optional[Configurati
     Client().start_session(tags, config)
 
 
-def record(event: Event):
+def record(event: Event | ErrorEvent):
     Client().record(event)
 
 

examples/README.md

@@ -0,0 +1,11 @@
+# AgentOps Examples
+
+## Quickstart
+- [Single LLM](./openai-gpt.ipynb)
+- [Recording Events](./recording-events.ipynb)
+- [Multi-Agent](./multi_agent_example.ipynb)
+
+## Integrations
+- [Using Langchain](./langchain_examples.ipynb)
+- [Crew.ai](https://github.com/joaomdmoura/crewAI-examples/tree/main/markdown_validator)
+  - Crew is a framework for developing agents, a number of their example projects use AgentOps

examples/openai-gpt.ipynb

@@ -7,7 +7,7 @@
     "collapsed": false
    },
    "source": [
-    "# Monitoring \n",
+    "# AgentOps Basic Monitoring \n",
     "This is an example of how to use the AgentOps library for basic Agent monitoring with OpenAI's GPT"
    ]
   },
@@ -70,7 +70,7 @@
    "outputs": [],
    "source": [
     "openai = OpenAI(api_key=OPENAI_API_KEY)\n",
-    "agentops.init(AGENTOPS_API_KEY)\n"
+    "agentops.init(AGENTOPS_API_KEY)"
    ]
   },
   {

examples/recording-events.ipynb

@@ -0,0 +1,179 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "source": [
+    "# Recording Events\n",
+    "AgentOps has a number of different [Event Types](https://docs.agentops.ai/v1/details/events)"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "dc8cfd2cfa8a594b"
+  },
+  {
+   "cell_type": "code",
+   "outputs": [],
+   "source": [
+    "# Create new session\n",
+    "agentops.start_session()\n",
+    "\n",
+    "# Optionally, we can add tags to the session\n",
+    "# agentops.start_session(['Hello Tracker'])"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "is_executing": true
+   },
+   "id": "168ecd05cc123de0",
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "The easiest way to record actions is through the use of AgentOp's decorators"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "c6d06ee8c66dad17"
+  },
+  {
+   "cell_type": "code",
+   "outputs": [],
+   "source": [
+    "from agentops import record_function\n",
+    "\n",
+    "@record_function(\"add numbers\")\n",
+    "def add(x, y):\n",
+    "    return x + y\n",
+    "\n",
+    "add(2,4)"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "b460318317adc624",
+   "execution_count": 0
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "We can also manually craft an event exactly the way we want"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "9068a4cdd328f652"
+  },
+  {
+   "cell_type": "code",
+   "outputs": [],
+   "source": [
+    "from agentops import ActionEvent\n",
+    "\n",
+    "message = {\"role\": \"user\", \"content\": \"Hello\"},\n",
+    "response = openai.chat.completions.create(\n",
+    "    model='gpt-3.5-turbo', messages=message, temperature=0.5)\n",
+    "\n",
+    "if \"hello\" in str(response.choices[0].message.content).lower():\n",
+    "    agentops.record(ActionEvent(action_type=\"Agent says hello\", params=str(message), returns=str(response.choices[0].message.content) ))"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "b62ad88921ff26f2",
+   "execution_count": 0
+  },
+  {
+   "cell_type": "code",
+   "outputs": [],
+   "source": [
+    "agentops.end_session('Success')"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "e10a89e06fe6b2be",
+   "execution_count": null
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "## Tool Event\n",
+    "Agents use tools. These tools are useful to track with information such as name, end status, runtime, etc. To record tool usage, you can create and record a `ToolEvent` similar to above."
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "f7c947d815f581e7"
+  },
+  {
+   "cell_type": "code",
+   "outputs": [],
+   "source": [
+    "from agentops import ToolEvent, record\n",
+    "def scrape_website(url: str):\n",
+    "    tool_event = ToolEvent(name='scrape_website', params={'url':url}) # the start timestamp is set when the obj is created\n",
+    "    result = integration.scrape_website(data) # perform tool logic\n",
+    "    tool_event.returns = result\n",
+    "    record(tool_event)"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "5d387a071a1c70cf"
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "## Error Events\n",
+    "Error events can be used alone or in reference to another event. Lets add a catch block to the code above"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "968d1503dd0aae9a"
+  },
+  {
+   "cell_type": "code",
+   "outputs": [],
+   "source": [
+    "from agentops import ToolEvent, record, ErrorEvent\n",
+    "def scrape_website(url: str):\n",
+    "    tool_event = ToolEvent(name='scrape_website', params={'url':url}) # the start timestamp is set when the obj is created\n",
+    "    try:\n",
+    "        result = integration.scrape_website(data) # perform tool logic\n",
+    "        tool_event.returns = result\n",
+    "    except Error as e:\n",
+    "        record(ErrorEvent(message=e, trigger_event=tool_event))\n",
+    "    record(tool_event)"
+   ],
+   "metadata": {
+    "collapsed": false
+   },
+   "id": "eb23c1325298e22f"
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 2
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython2",
+   "version": "2.7.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}