Enables creation of workflows for Dria Agents
Project description
Dria Workflows
Dria Workflows enables the creation of workflows for Dria Agents.
Installation
You can install Dria Workflows using pip:
pip install dria_workflows
Usage Example
Here's a simple example of how to use Dria Workflows:
import logging
from dria_workflows import WorkflowBuilder, Operator, Write, Edge, validate_workflow_json
def main():
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
builder = WorkflowBuilder()
# Add a step to your workflow
builder.generative_step(id="write_poem", prompt="Write a poem as if you are Kahlil Gibran", operator=Operator.GENERATION, outputs=[Write.new("poem")])
# Define the flow of your workflow
flow = [Edge(source="write_poem", target="_end")]
builder.flow(flow)
# Set the return value of your workflow
builder.set_return_value("poem")
# Build your workflow
workflow = builder.build()
# Validate your workflow
validate_workflow_json(workflow.model_dump_json(indent=2, exclude_unset=True, exclude_none=True))
# Save workflow
workflow.save("poem_workflow.json")
if __name__ == "__main__":
main()
Here is a more complex workflow
import logging
from dria_workflows import WorkflowBuilder, ConditionBuilder, Operator, Write, GetAll, Read, Push, Edge, Expression, validate_workflow_json
def main():
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
# Give a starting memory as input
builder = WorkflowBuilder(memory={"topic_1":"Linear Algebra", "topic_2":"CUDA"})
# Add steps to your workflow
builder.generative_step(id="create_query", prompt="Write down a search query related to following topics: {{topic_1}} and {{topic_2}}. If any, avoid asking questions asked before: {{history}}", operator=Operator.GENERATION, inputs=[GetAll.new("history", False)], outputs=[Write.new("search_query")])
builder.generative_step(id="search", prompt="{{search_query}}", operator=Operator.FUNCTION_CALLING, outputs=[Write.new("result"), Push.new("history")])
builder.generative_step(id="evaluate", prompt="Evaluate if search result is related and high quality to given question by saying Yes or No. Question: {{search_query}} , Search Result: {{result}}. Only output Yes or No and nothing else.", operator=Operator.GENERATION, outputs=[Write.new("is_valid")])
# Define the flow of your workflow
flow = [
Edge(source="create_query", target="search"),
Edge(source="search", target="evaluate"),
Edge(source="evaluate", target="_end", condition=ConditionBuilder.build(expected="Yes", target_if_not="create_query", expression=Expression.CONTAINS, input=Read.new("is_valid", True))),
]
builder.flow(flow)
# Set the return value of your workflow
builder.set_return_value("result")
# Build your workflow
workflow = builder.build()
validate_workflow_json(workflow.model_dump_json(indent=2, exclude_unset=True, exclude_none=True))
workflow.save("search_workflow.json")
if __name__ == "__main__":
main()
Detailed docs soon. andthattoo
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
dria_workflows-0.3.0.tar.gz
(20.7 kB
view details)
Built Distribution
File details
Details for the file dria_workflows-0.3.0.tar.gz
.
File metadata
- Download URL: dria_workflows-0.3.0.tar.gz
- Upload date:
- Size: 20.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.8.3 CPython/3.12.5 Darwin/23.6.0
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | be956bce5e7f9ef47dcf6702b60fd0d0fae6ac90b4489456cfabdfe3ffab72f0 |
|
MD5 | f8e4b6e84ec9838f90c07ea388a0363d |
|
BLAKE2b-256 | b1c20fb1de7969486613128a22e4b7119e6f51634ab5623d80b1f28fefb27d31 |
File details
Details for the file dria_workflows-0.3.0-py3-none-any.whl
.
File metadata
- Download URL: dria_workflows-0.3.0-py3-none-any.whl
- Upload date:
- Size: 27.0 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.8.3 CPython/3.12.5 Darwin/23.6.0
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 411f6f3c9a0ea270af0d31242b268d4908c1604c1ee66e6970434e84d9d50409 |
|
MD5 | 75fea7f3988acaa5226c04fd701c25f3 |
|
BLAKE2b-256 | 9bd3ffded72eeec6e05e0f67f157acc707e2e72e0a8a96f6371a951bd62e4001 |