Fimeg/letta-server
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

feat: add testing for LLM + embedding endpoints (#1308)

Browse Source 
Co-authored-by: cpacker <packercharles@gmail.com>
This commit is contained in:
Sarah Wooders
2024-04-29 22:23:42 -07:00
committed by GitHub
parent 877dd89f3c
commit e27a5eccdc
10 changed files with 136 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
configs/embedding_model_configs/memgpt-hosted.json Normal file
View File

@@ -0,0 +1,6 @@
{
"embedding_endpoint": "https://embeddings.memgpt.ai",
"embedding_model": "BAAI/bge-large-en-v1.5",
"embedding_dim": 1024,
"embedding_chunk_size": 300
}

7
configs/embedding_model_configs/text-embedding-ada-002.json Normal file
View File

@@ -0,0 +1,7 @@
{
"embedding_endpoint_type": "openai",
"embedding_endpoint": "https://api.openai.com/v1",
"embedding_model": "text-embedding-ada-002",
"embedding_dim": 1536,
"embedding_chunk_size": 300
}

7
configs/llm_model_configs/gpt-4.json Normal file
View File

@@ -0,0 +1,7 @@
{
"context_window": 8192,
"model": "gpt-4",
"model_endpoint_type": "openai",
"model_endpoint": "https://api.openai.com/v1",
"model_wrapper": null
}

6
configs/llm_model_configs/memgpt-hosted.json Normal file
View File

@@ -0,0 +1,6 @@
{
"context_window": 16384,
"model_endpoint_type": "openai",
"model_endpoint": "https://inference.memgpt.ai",
"model": "memgpt-openai"
}

1
configs/memgpt_hosted.json
View File

@@ -2,6 +2,7 @@
"context_window": 16384,
"model_endpoint_type": "openai",
"model_endpoint": "https://inference.memgpt.ai",
"model": "memgpt-openai",
"embedding_endpoint_type": "hugging-face",
"embedding_endpoint": "https://embeddings.memgpt.ai",
"embedding_model": "BAAI/bge-large-en-v1.5",

4
memgpt/data_types.py
View File

@@ -78,10 +78,10 @@ class Message(Record):
def __init__(
self,
user_id: uuid.UUID,
agent_id: uuid.UUID,
role: str,
text: str,
user_id: Optional[uuid.UUID] = None,
agent_id: Optional[uuid.UUID] = None,
model: Optional[str] = None, # model used to make function call
name: Optional[str] = None, # optional participant name
created_at: Optional[datetime] = None,

3
memgpt/llm_api/openai.py
View File

@@ -363,7 +363,8 @@ def openai_chat_completions_request(
printd(f"Sending request to {url}")
try:
response = requests.post(url, headers=headers, json=data)
printd(f"response = {response}")
# printd(f"response = {response}, response.text = {response.text}")
print(f"response = {response}, response.text = {response.text}")
response.raise_for_status() # Raises HTTPError for 4XX/5XX status
response = response.json() # convert to dict from string

1
memgpt/models/chat_completion_request.py
View File

@@ -42,6 +42,7 @@ class ToolMessage(BaseModel):
ChatMessage = Union[SystemMessage, UserMessage, AssistantMessage, ToolMessage]
# TODO: this might not be necessary with the validator
def cast_message_to_subtype(m_dict: dict) -> ChatMessage:
"""Cast a dictionary to one of the individual message types"""
role = m_dict.get("role")

48
memgpt/presets/presets.py
View File

@@ -65,6 +65,25 @@ def create_preset_from_file(filename: str, name: str, user_id: uuid.UUID, ms: Me
return preset
def load_preset(preset_name: str, user_id: uuid.UUID):
preset_config = available_presets[preset_name]
preset_system_prompt = preset_config["system_prompt"]
preset_function_set_names = preset_config["functions"]
functions_schema = generate_functions_json(preset_function_set_names)
preset = Preset(
user_id=user_id,
name=preset_name,
system=gpt_system.get_system_text(preset_system_prompt),
persona=get_persona_text(DEFAULT_PERSONA),
persona_name=DEFAULT_PERSONA,
human=get_human_text(DEFAULT_HUMAN),
human_name=DEFAULT_HUMAN,
functions_schema=functions_schema,
)
return preset
def add_default_presets(user_id: uuid.UUID, ms: MetadataStore):
"""Add the default presets to the metadata store"""
# make sure humans/personas added
@@ -72,25 +91,26 @@ def add_default_presets(user_id: uuid.UUID, ms: MetadataStore):
# add default presets
for preset_name in preset_options:
preset_config = available_presets[preset_name]
preset_system_prompt = preset_config["system_prompt"]
preset_function_set_names = preset_config["functions"]
functions_schema = generate_functions_json(preset_function_set_names)
# preset_config = available_presets[preset_name]
# preset_system_prompt = preset_config["system_prompt"]
# preset_function_set_names = preset_config["functions"]
# functions_schema = generate_functions_json(preset_function_set_names)
if ms.get_preset(user_id=user_id, name=preset_name) is not None:
printd(f"Preset '{preset_name}' already exists for user '{user_id}'")
continue
preset = Preset(
user_id=user_id,
name=preset_name,
system=gpt_system.get_system_text(preset_system_prompt),
persona=get_persona_text(DEFAULT_PERSONA),
persona_name=DEFAULT_PERSONA,
human=get_human_text(DEFAULT_HUMAN),
human_name=DEFAULT_HUMAN,
functions_schema=functions_schema,
)
preset = load_preset(preset_name, user_id)
# preset = Preset(
# user_id=user_id,
# name=preset_name,
# system=gpt_system.get_system_text(preset_system_prompt),
# persona=get_persona_text(DEFAULT_PERSONA),
# persona_name=DEFAULT_PERSONA,
# human=get_human_text(DEFAULT_HUMAN),
# human_name=DEFAULT_HUMAN,
# functions_schema=functions_schema,
# )
ms.create_preset(preset)

70
tests/test_endpoints.py Normal file
View File

@@ -0,0 +1,70 @@
import json
import os
import uuid
from memgpt.agent import Agent
from memgpt.data_types import Message
from memgpt.embeddings import embedding_model
from memgpt.llm_api.llm_api_tools import create
from memgpt.models.pydantic_models import EmbeddingConfigModel, LLMConfigModel
from memgpt.presets.presets import load_preset
from memgpt.prompts import gpt_system
messages = [Message(role="system", text=gpt_system.get_system_text("memgpt_chat")), Message(role="user", text="How are you?")]
embedding_config_path = "configs/embedding_model_configs/memgpt-hosted.json"
llm_config_path = "configs/llm_model_configs/memgpt-hosted.json"
def test_embedding_endpoints():
embedding_config_dir = "configs/embedding_model_configs"
# list JSON files in directory
for file in os.listdir(embedding_config_dir):
if file.endswith(".json"):
# load JSON file
print("testing", file)
config_data = json.load(open(os.path.join(embedding_config_dir, file)))
embedding_config = EmbeddingConfigModel(**config_data)
# model = embedding_model(embedding_config, user_id=uuid.UUID(int=1))
model = embedding_model(embedding_config)
query_text = "hello"
query_vec = model.get_text_embedding(query_text)
print("vector dim", len(query_vec))
def test_llm_endpoints():
llm_config_dir = "configs/llm_model_configs"
# use openai default config
embedding_config = EmbeddingConfigModel(**json.load(open(embedding_config_path)))
# list JSON files in directory
for file in os.listdir(llm_config_dir):
if file.endswith(".json"):
# load JSON file
print("testing", file)
config_data = json.load(open(os.path.join(llm_config_dir, file)))
print(config_data)
llm_config = LLMConfigModel(**config_data)
agent = Agent(
interface=None,
preset=load_preset("memgpt_chat", user_id=uuid.UUID(int=1)),
name="test_agent",
created_by=uuid.UUID(int=1),
llm_config=llm_config,
embedding_config=embedding_config,
# gpt-3.5-turbo tends to omit inner monologue, relax this requirement for now
first_message_verify_mono=True,
)
response = create(
llm_config=llm_config,
user_id=uuid.UUID(int=1), # dummy user_id
# messages=agent_state.messages,
messages=agent._messages,
functions=agent.functions,
functions_python=agent.functions_python,
)
assert response is not None