No project description provided
Project description
OpenPlugin
:open_hands::electric_plug: LLM Tool Runner - Chat with your APIs - Abstraction Layer over Function Calling -
This is an open source effort to create an ecosystem around LLM enabled APIs. We make it easy to "chat with an API", that is, to send natural language as the input to the API and have it process it. Then, we offer several options to easily transform the APIs response into something better suited for human consumption like text, markdown, HTML, JSX, voice, video, etc.
Starting a Server
1. Starting an OpenPlugin Server from PyPI
pip install openplugin
openplugin --help
export OPENAI_API_KEY=<your key>
openplugin start-server
2. Starting OpenPlugin Server from Docker
# Passing environment variables in the startup script
docker run --name openplugin_container -p 8006:8006 -e "OPENAI_API_KEY=<your_key>" -e "COHERE_API_KEY=<your_key>" -e "GOOGLE_APPLICATION_CREDENTIALS=<your_key>" -d shrikant14/openplugin:latest
# Passing environment variables as a file
nano [env-filename]
Add to file
[variable1-name]=[value1]
[variable2-name]=[value2]
[variable3-name]=[value3]
docker run --name openplugin_container -p 8006:8006 --env-file my_env.env -d shrikant14/openplugin:latest
3. Starting an OpenPlugin Server from code
git clone <openplugin>
cd openplugin
poetry install
python openplugin/main.py run-plugin --openplugin manifests/sample_klarna.json --prompt sample_prompt.txt --log-level="FLOW"
Run an OpenPlugin
1. Run an OpenPlugin using PyPI
pip install openplugin
openplugin --help
export OPENAI_API_KEY=<your key>
openplugin start-servero
openplugin run-plugin --openplugin manifests/sample_klarna.json --prompt sample_prompt.txt --log-level="FLOW"
2. Run via an API call
curl --location 'https://api.imprompt.ai/openplugin/api/plugin-execution-pipeline' \
--header 'Content-Type: application/json' \
--header 'x-api-key: 'YOUR-API-KEY' \
--data '{
"prompt": "USER_PROMPT",
"conversation": [],
"openplugin_manifest_url": "MANIFEST_URL",
"header":{},
"approach": {
"base_strategy": "oai functions",
"llm": {
"frequency_penalty": 0,
"max_tokens": 2048,
"model_name": "gpt-3.5-turbo-0613",
"presence_penalty": 0,
"provider": "OpenAI",
"temperature": 0,
"top_p": 1
},
"name": "OAI functions-OpenAI",
"pre_prompt": null
},
"output_module_names":["default_cleanup_response"]
}'
3. Run via Code
pip install openplugin
from openplugin.core.plugin_runner import run_prompt_on_plugin
openplugin=""
prompt=""
response =await run_prompt_on_plugin(openplugin, prompt)
4. Run via SDK
NOTE: Learn more about openplugin-sdk at: https://github.com/ImpromptAI/openplugin-sdk
pip install openplugin-sdk
remote_server_endpoint = "...."
openplugin_api_key = "...."
svc = OpenpluginService(
remote_server_endpoint=remote_server_endpoint, api_key=openplugin_api_key
)
openplugin_manifest_url = "...."
prompt = "..."
output_module_name="..."
response = svc.run(
openplugin_manifest_url=openplugin_manifest_url,
prompt=prompt,
output_module_names=[output_module_name],
)
print(f"Response={response.value}")
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
Built Distribution
File details
Details for the file openplugin-0.1.8.tar.gz
.
File metadata
- Download URL: openplugin-0.1.8.tar.gz
- Upload date:
- Size: 59.0 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.5.0 CPython/3.10.12 Linux/6.5.0-1023-azure
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 7f492a4c2976c67ec02ddf90fcef6a0f74abc8846e1d3245ffa86399e3d058f5 |
|
MD5 | 340317557849b1bab930e092c2ff3633 |
|
BLAKE2b-256 | 729e8fc707899db4daa3b19d01286808ff7783a517467c1d03c805e761927b16 |
File details
Details for the file openplugin-0.1.8-py3-none-any.whl
.
File metadata
- Download URL: openplugin-0.1.8-py3-none-any.whl
- Upload date:
- Size: 93.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: poetry/1.5.0 CPython/3.10.12 Linux/6.5.0-1023-azure
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 1f768b41c05d95bb33a8900913c864ef7076bbbec66558d53998fc0a08601171 |
|
MD5 | d02a71e71628fb9cf69bf52b2e7dd6e8 |
|
BLAKE2b-256 | dbcb74fb44e4b7c5d3d7e2598cc9f47796a03b112cc706c77abb567bad1c05c0 |