refactor openai client

.gitignore

@@ -0,0 +1,4 @@
+__pycache__/
+dist/
+build/
+.vscode/

cookbook/agent.py

@@ -0,0 +1,38 @@
+import os
+import nest_asyncio
+from lyzr_experimental_automata import Agent, Task, State
+
+os.environ["OPENAI_API_KEY"] = "sk-"
+
+nest_asyncio.apply()
+
+agent1 = Agent(persona="Marketing Consultant")
+agent2 = Agent(persona="Tweet Generator")
+agent3 = Agent(persona="Linkedin Post Creator")
+
+task1 = Task(
+    instructions="Do research and pull out interesting marketing tips for SaaS companies. The research articles should not be more than 500 words.",
+    desired_output="Ensure that you bring the best content from the likes of HBS, YCombinator",
+    agent=agent1,
+    display_output="yes",
+)
+
+task2 = Task(
+    instructions="Use the research material provided and write 5 engaging tweets. Display only the tweets. No explanation or additional comments required.",
+    desired_output="Ensure that the tweets are as engaging as if it was written by the best influencer in the world",
+    agent=agent2,
+    display_output="yes",
+    dependencies=[task1],
+)
+
+task3 = Task(
+    instructions="Use the research material provided and write 1 short form LinkedIn post. Display only the LinkedIn post. No explanation or additional comments required.",
+    desired_output="Ensure that the post is as if it was written by the best influencer in the world",
+    agent=agent3,
+    display_output="yes",
+    dependencies=[task1],
+)
+
+output = State([task1, task2, task3])
+
+print(output)

lyzr_experimental_automata.egg-info/PKG-INFO

@@ -0,0 +1,74 @@
+Metadata-Version: 2.1
+Name: lyzr-experimental-automata
+Version: 0.2.0
+Summary: A prompt based agent workflow that integrates with other Lyzr agents
+Home-page: https://github.com/lyzrcore/lyzr-experimental-automata
+Author: Siva Surendira
+Author-email: siva@lyzr.ai
+License: MIT
+Description-Content-Type: check Readme file
+License-File: LICENSE
+Requires-Dist: openai
+Requires-Dist: asyncio
+
+# lyzr-experimental-automata
+The version 0.2 is a prompt based agent workflow that integrates with other Lyzr agents
+
+# Lyzr Automata - Parallel-Processing Agent Automation Framework
+
+Lyzr Automata is a sophisticated parallel-processing agent automation framework designed to enhance workflow efficiency and effectiveness. It enables a seamless transition between states, leverages recursive checks for continual improvement, and ensures the achievement of end goals in line with assigned objectives.
+
+## How to Install
+
+Get started with Lyzr Automata by installing the experimental package using pip:
+
+pip install lyzr-experimental-automata
+
+## Configuring Agents
+
+Begin by configuring your agents and assigning them unique personas:
+
+```python
+agent1 = Agent(persona="enter the persona of agent1")
+agent2 = Agent(persona="enter the persona of agent2")
+agent3 = Agent(persona="enter the persona of agent3")
+
+Example Configuration:
+
+agent1 = Agent(persona="Marketing Consultant")
+agent2 = Agent(persona="Tweet Generator")
+agent3 = Agent(persona="Linkedin Post Creator")
+
+Creating Tasks
+Create tasks by providing specific instructions and desired outcomes. Assign these tasks to your agents:
+
+task1 = Task("enter the instructions", "enter the desired outcome", agent1, display_output='no')
+
+Task1 is the initial task in the workflow. You can control the visibility of its output by setting display_output to either 'yes' or 'no'.
+
+Example Task 1:
+
+task1 = Task("Do research and pull out interesting marketing tips for SaaS companies. The research articles should not be more than 500 words.", "Ensure that you bring the best content from the likes of HBS, YCombinator", agent1, display_output='no')
+
+Setting Up Dependencies
+Leverage the multi-thread, parallel-processing capabilities of Lyzr Automata by specifying task dependencies:
+
+task2 = Task("enter the instructions", "enter the desired outcome", agent1, display_output='yes', dependencies=[task1])
+
+Example Task 2:
+
+task2 = Task("Use the research material provided and write 5 engaging tweets. Display only the tweets. No explanation or additional comments required.", "Ensure that the tweets are as engaging as if it was written by the best influencer in the world", agent2, display_output='yes', dependencies=[task1])
+
+Continue adding tasks as required, defining their dependencies to optimize parallel processing.
+
+Further Tasks
+Add additional tasks (like Task 3) and define their dependencies for continued workflow.
+
+task3 = Task("enter the instructions", "enter the desired outcome", agent1, display_output='yes', dependencies=[task1])
+
+Example Task 3:
+
+task3 = Task("Use the research material provided and write 1 short form LinkedIn post. Display only the LinkedIn post. No explanation or additional comments required.", "Ensure that the post is as if it was written by the best influencer in the world", agent3, display_output='yes', dependencies=[task1])
+
+Lyzr Automata is your go-to solution for creating an efficient, automated workflow powered by the capabilities of OpenAI.
+

lyzr_experimental_automata.egg-info/SOURCES.txt

@@ -0,0 +1,10 @@
+LICENSE
+README.md
+setup.py
+lyzr_experimental_automata/__init__.py
+lyzr_experimental_automata/automata.py
+lyzr_experimental_automata.egg-info/PKG-INFO
+lyzr_experimental_automata.egg-info/SOURCES.txt
+lyzr_experimental_automata.egg-info/dependency_links.txt
+lyzr_experimental_automata.egg-info/requires.txt
+lyzr_experimental_automata.egg-info/top_level.txt

lyzr_experimental_automata.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

lyzr_experimental_automata.egg-info/requires.txt

@@ -0,0 +1,2 @@
+openai
+asyncio

lyzr_experimental_automata.egg-info/top_level.txt

@@ -0,0 +1 @@
+lyzr_experimental_automata

lyzr_experimental_automata/__init__.py

@@ -1 +1 @@
-from .automata import Agent, Task, run_state
+from lyzr_experimental_automata.automata import Agent, Task, State

lyzr_experimental_automata/automata.py

@@ -1,66 +1,91 @@
+import os
 import asyncio
-import openai
+from openai import OpenAI
+
 
 class Agent:
     def __init__(self, persona):
         self.persona = persona
 
+
 class Task:
-    def __init__(self, instructions, desired_output, agent, display_output='no', dependencies=None,
-                 model="gpt-4-turbo-preview", temperature=1, max_tokens=1000, top_p=1,
-                 frequency_penalty=0, presence_penalty=0):
+    def __init__(
+        self,
+        instructions,
+        desired_output,
+        agent,
+        display_output="no",
+        dependencies=None,
+        model="gpt-4-turbo-preview",
+        temperature=1,
+        max_tokens=1000,
+        top_p=1,
+        frequency_penalty=0,
+        presence_penalty=0,
+    ):
         self.instructions = instructions
         self.desired_output = desired_output
         self.agent = agent
-        self.display_output = display_output.lower() == 'yes'
+        self.display_output = display_output.lower() == "yes"
         self.dependencies = dependencies if dependencies else []
         self.model = model
         self.temperature = temperature
         self.max_tokens = max_tokens
         self.top_p = top_p
         self.frequency_penalty = frequency_penalty
         self.presence_penalty = presence_penalty
+        self.client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
 
     async def execute(self, input_from_previous_task=""):
         persona_with_prefix = f"Your role is a {self.agent.persona}"
-        prompt = (f"Instructions {self.instructions} Desired outcome: {self.desired_output} Input: {input_from_previous_task}")
+        prompt = f"Instructions {self.instructions} Desired outcome: {self.desired_output} Input: {input_from_previous_task}"
 
         loop = asyncio.get_event_loop()
-        response = await loop.run_in_executor(None, lambda: client.chat.completions.create(
-            model=self.model,
-            messages=[
-                {"role": "system", "content": persona_with_prefix},
-                {"role": "user", "content": prompt}
-            ],
-            temperature=self.temperature,
-            max_tokens=self.max_tokens,
-            top_p=self.top_p,
-            frequency_penalty=self.frequency_penalty,
-            presence_penalty=self.presence_penalty
-        ))
+        response = await loop.run_in_executor(
+            None,
+            lambda: self.client.chat.completions.create(
+                model=self.model,
+                messages=[
+                    {"role": "system", "content": persona_with_prefix},
+                    {"role": "user", "content": prompt},
+                ],
+                temperature=self.temperature,
+                max_tokens=self.max_tokens,
+                top_p=self.top_p,
+                frequency_penalty=self.frequency_penalty,
+                presence_penalty=self.presence_penalty,
+            ),
+        )
         return response.choices[0].message.content
 
+
 async def _execute_task(task, task_to_future, outputs):
     for dependency in task.dependencies:
         await task_to_future[dependency]
 
-    input_from_previous_task = " ".join(outputs.get(dep, "") for dep in task.dependencies)
+    input_from_previous_task = " ".join(
+        outputs.get(dep, "") for dep in task.dependencies
+    )
     output = await task.execute(input_from_previous_task)
 
     outputs[task] = output
     if task.display_output:
         print("Task Output:", output)
 
+
 async def _run_tasks(tasks):
     outputs = {}
     task_to_future = {}
 
     for task in tasks:
-        task_to_future[task] = asyncio.create_task(_execute_task(task, task_to_future, outputs))
+        task_to_future[task] = asyncio.create_task(
+            _execute_task(task, task_to_future, outputs)
+        )
 
     await asyncio.gather(*task_to_future.values())
 
     return " ".join(outputs[task] for task in tasks)
 
+
 def State(tasks):
     return asyncio.run(_run_tasks(tasks))