apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-06-20 15:28:07 +02:00
Code Issues Projects Releases Packages Wiki Activity

more changes

Browse source
This commit is contained in:
Adil Hafeez 2025-09-17 00:57:26 -07:00
parent 6cfce60501
commit c173074864
No known key found for this signature in database
GPG key ID: 9B18EF7691369645
9 changed files with 286 additions and 215 deletions
Download patch file Download diff file Expand all files Collapse all files

141
arch/tools/cli/config_generator.py
View file

@ -1,5 +1,6 @@
import json
import os
from cli.utils import convert_legacy_llm_providers
from jinja2 import Environment, FileSystemLoader
import yaml
from jsonschema import validate
@ -66,6 +67,12 @@ def validate_and_render_schema():
_ = yaml.safe_load(arch_config_schema)
inferred_clusters = {}
listeners, llm_gateway, prompt_gateway = convert_legacy_llm_providers(
config_yaml.get("listeners"), config_yaml.get("llm_providers")
)
config_yaml["listeners"] = listeners
endpoints = config_yaml.get("endpoints", {})
# Process agents section and convert to endpoints
@ -126,64 +133,62 @@ def validate_and_render_schema():
model_name_keys = set()
model_usage_name_keys = set()
# legacy listeners
# check if type is array or object
# if its dict its legacy format let's convert it to array
prompt_gateway_listener = {
"name": "ingress_traffic",
"port": 10000,
"address": "0.0.0.0",
"timeout": "30s",
"protocol": "openai",
}
llm_gateway_listener = {
"name": "egress_traffic",
"port": 12000,
"address": "0.0.0.0",
"timeout": "30s",
"llm_providers": [],
"protocol": "openai",
}
if isinstance(config_yaml["listeners"], dict):
ingress_traffic = config_yaml["listeners"].get("ingress_traffic", None)
egress_traffic = config_yaml["listeners"].get("egress_traffic", {})
config_yaml["listeners"] = []
# # legacy listeners
# # check if type is array or object
# # if its dict its legacy format let's convert it to array
# prompt_gateway_listener = {
# "name": "ingress_traffic",
# "port": 10000,
# "address": "0.0.0.0",
# "timeout": "30s",
# "protocol": "openai",
# }
# llm_gateway_listener = {
# "name": "egress_traffic",
# "port": 12000,
# "address": "0.0.0.0",
# "timeout": "30s",
# "llm_providers": [],
# "protocol": "openai",
# }
# if isinstance(config_yaml["listeners"], dict):
# ingress_traffic = config_yaml["listeners"].get("ingress_traffic", None)
# egress_traffic = config_yaml["listeners"].get("egress_traffic", {})
# config_yaml["listeners"] = []
llm_providers = []
if config_yaml.get("llm_providers"):
llm_providers = config_yaml["llm_providers"]
del config_yaml["llm_providers"]
llm_gateway_listener["port"] = egress_traffic.get(
"port", llm_gateway_listener["port"]
)
llm_gateway_listener["address"] = egress_traffic.get(
"address", llm_gateway_listener["address"]
)
llm_gateway_listener["timeout"] = egress_traffic.get(
"timeout", llm_gateway_listener["timeout"]
)
llm_gateway_listener["llm_providers"] = llm_providers
config_yaml["listeners"].append(llm_gateway_listener)
# llm_providers = []
# if config_yaml.get("llm_providers"):
# llm_providers = config_yaml["llm_providers"]
# del config_yaml["llm_providers"]
# llm_gateway_listener["port"] = egress_traffic.get(
# "port", llm_gateway_listener["port"]
# )
# llm_gateway_listener["address"] = egress_traffic.get(
# "address", llm_gateway_listener["address"]
# )
# llm_gateway_listener["timeout"] = egress_traffic.get(
# "timeout", llm_gateway_listener["timeout"]
# )
# llm_gateway_listener["llm_providers"] = llm_providers
# config_yaml["listeners"].append(llm_gateway_listener)
if ingress_traffic:
prompt_gateway_listener["port"] = ingress_traffic.get(
"port", prompt_gateway_listener["port"]
)
prompt_gateway_listener["address"] = ingress_traffic.get(
"address", prompt_gateway_listener["address"]
)
prompt_gateway_listener["timeout"] = ingress_traffic.get(
"timeout", prompt_gateway_listener["timeout"]
)
config_yaml["listeners"].append(prompt_gateway_listener)
# if ingress_traffic:
# prompt_gateway_listener["port"] = ingress_traffic.get(
# "port", prompt_gateway_listener["port"]
# )
# prompt_gateway_listener["address"] = ingress_traffic.get(
# "address", prompt_gateway_listener["address"]
# )
# prompt_gateway_listener["timeout"] = ingress_traffic.get(
# "timeout", prompt_gateway_listener["timeout"]
# )
# config_yaml["listeners"].append(prompt_gateway_listener)
for listener in config_yaml["listeners"]:
print("Processing listener: ", listener)
name = listener.get("name", None)
# TODO: for now we only support llm_providers under egress_traffic listener
if name != "egress_traffic":
for listener in listeners:
if listener.get("llm_providers") is None or listener.get("llm_providers") == []:
continue
print("Processing listener with llm_providers: ", listener)
name = listener.get("name", None)
for llm_provider in listener.get("llm_providers", []):
if llm_provider.get("usage", None):
@ -293,12 +298,18 @@ def validate_and_render_schema():
}
)
for listener in config_yaml["listeners"]:
updated_llm_providers = []
for listener in listeners:
print("Processing listener: ", listener)
if listener.get("name") == "egress_traffic":
llm_providers = listener.get("llm_providers", None)
if llm_providers is not None and llm_providers != []:
print("processing egress traffic listener")
print("updated_llm_providers: ", updated_llm_providers)
listener["llm_providers"] = deepcopy(updated_llm_providers)
if updated_llm_providers is not None and updated_llm_providers != []:
raise Exception(
"Please provide llm_providers either under listeners or at root level, not both. Currently we don't support multiple listeners with llm_providers"
)
updated_llm_providers = deepcopy(llm_providers)
config_yaml["llm_providers"] = updated_llm_providers
# Validate model aliases if present
@ -314,16 +325,6 @@ def validate_and_render_schema():
arch_config_string = yaml.dump(config_yaml)
arch_llm_config_string = yaml.dump(config_yaml)
# prompt_gateway_listener = config_yaml.get("listeners", {}).get(
# "ingress_traffic", {}
# )
# if prompt_gateway_listener.get("port") == None:
# prompt_gateway_listener["port"] = 10000 # default port for prompt gateway
# if prompt_gateway_listener.get("address") == None:
# prompt_gateway_listener["address"] = "127.0.0.1"
# if prompt_gateway_listener.get("timeout") == None:
# prompt_gateway_listener["timeout"] = "10s"
use_agent_orchestrator = config_yaml.get("overrides", {}).get(
"use_agent_orchestrator", False
)
@ -346,8 +347,8 @@ def validate_and_render_schema():
print("agent_orchestrator: ", agent_orchestrator)
data = {
"prompt_gateway_listener": prompt_gateway_listener,
"llm_gateway_listener": llm_gateway_listener,
"prompt_gateway_listener": prompt_gateway,
"llm_gateway_listener": llm_gateway,
"arch_config": arch_config_string,
"arch_llm_config": arch_llm_config_string,
"arch_clusters": inferred_clusters,
@ -355,7 +356,7 @@ def validate_and_render_schema():
"arch_tracing": arch_tracing,
"local_llms": llms_with_endpoint,
"agent_orchestrator": agent_orchestrator,
"listeners": config_yaml["listeners"].copy(),
"listeners": listeners,
}
rendered = template.render(data)