mirror of
https://github.com/katanemo/plano.git
synced 2026-04-25 08:46:24 +02:00
4182879717
* add precommit check
* remove check
* Revert "remove check"
This reverts commit 9987b62b9b.
* fix checks
* fix whitespace errors
78 lines
3.2 KiB
YAML
78 lines
3.2 KiB
YAML
version: "0.1-beta"
|
|
|
|
listener:
|
|
address: 0.0.0.0 # or 127.0.0.1
|
|
port_value: 8080
|
|
messages: "hugging-face-messages-json" # Defines how Arch should parse the content from application/json or text/pain Content-type in the http request
|
|
common_tls_context: # If you configure port 443, you'll need to update the listener with your TLS certificates
|
|
tls_certificates:
|
|
- certificate_chain:
|
|
filename: "/etc/arch/certs/cert.pem"
|
|
private_key:
|
|
filename: "/etc/arch/certs/key.pem"
|
|
|
|
system_prompts:
|
|
- name: "network_assistant"
|
|
content: |
|
|
You are a network assistant that just offers facts; not advice on manufacturers or purchasing decisions.
|
|
|
|
llm_providers: #Centralized way to manage LLMs, manage keys, retry logic, failover and limits in a central way
|
|
- name: "OpenAI"
|
|
access_key: $OPENAI_API_KEY
|
|
model: gpt-4o
|
|
default: true
|
|
stream: true
|
|
rate_limit:
|
|
selector: #optional headers, to add rate limiting based on http headers like JWT tokens or API keys
|
|
http-header:
|
|
name: "Authorization"
|
|
value: "" # Empty value means each separate value has a separate limit
|
|
limit:
|
|
tokens: 100000 # Tokens per unit
|
|
unit: "minute"
|
|
- name: "Mistral"
|
|
access_key: $MISTRAL_API_KEY
|
|
model: "mistral-7B"
|
|
|
|
prompt_endpoints: #Arch creates a round-robin load balancing between different endpoints, managed via the cluster subsystem.
|
|
- "http://127.0.0.2" #assumes port 8000, unless port is specified with :5000
|
|
- "http://127.0.0.1:5000"
|
|
|
|
prompt_guards:
|
|
input_guard:
|
|
- name: "jailbreak"
|
|
on_exception:
|
|
forward_to_error_target: true
|
|
- name: "toxicity"
|
|
on_exception:
|
|
message: "Looks like you're curious about my abilities, but I can only provide assistance within my programmed parameters."
|
|
|
|
prompt_targets:
|
|
- name: "information_extraction"
|
|
type: "default"
|
|
description: "This prompt handles all scenarios that are question and answer in nature. Like summarization, information extraction, etc."
|
|
path: "/agent/summary"
|
|
auto-llm-dispatch-on-response: true #Arch uses the default LLM and treats the response from the endpoint as the prompt to send to the LLM
|
|
|
|
- name: "reboot_network_device"
|
|
path: "/agent/action"
|
|
description: "Helps network operators perform device operations like rebooting a device."
|
|
parameters:
|
|
- name: "device_id"
|
|
type: "string" # additional type options include: integer | float | list | dictionary | set
|
|
description: "Identifier of the network device to reboot."
|
|
default_value: ""
|
|
required: true
|
|
- name: "confirmation"
|
|
type: "integer" # additional type options include: integer | float | list | dictionary | set
|
|
description: "Confirmation flag to proceed with reboot."
|
|
required: true
|
|
|
|
error_target:
|
|
name: "error_handler"
|
|
path: "/errors"
|
|
|
|
tracing: 100 #sampling rate. Note by default Arch works on OpenTelemetry compatible tracing.
|
|
|
|
intent-detection-threshold-override: 0.60 # By default Arch uses an NLI + embedding approach to match an incomming prompt to a prompt target.
|
|
# The intent matching threshold is kept at 0.80, you can overide this behavior if you would like
|