feat: add ai/ml model and embedding components (#2781)
This commit is contained in:
Jordan Frazier
GitHub
parent
495c95f036
commit
5ed8d1a63a
14 changed files with 781 additions and 7 deletions
77
src/backend/base/langflow/base/models/aiml_constants.py
Normal file
77
src/backend/base/langflow/base/models/aiml_constants.py
Normal file
|
|
@ -0,0 +1,77 @@
|
|||
CHAT_MODELS = [
|
||||
"zero-one-ai/Yi-34B-Chat",
|
||||
"allenai/OLMo-7B-Instruct",
|
||||
"allenai/OLMo-7B-Twin-2T",
|
||||
"allenai/OLMo-7B",
|
||||
"Austism/chronos-hermes-13b",
|
||||
"cognitivecomputations/dolphin-2.5-mixtral-8x7b",
|
||||
"deepseek-ai/deepseek-coder-33b-instruct",
|
||||
"deepseek-ai/deepseek-llm-67b-chat",
|
||||
"garage-bAInd/Platypus2-70B-instruct",
|
||||
"google/gemma-2b-it",
|
||||
"google/gemma-7b-it",
|
||||
"Gryphe/MythoMax-L2-13b",
|
||||
"lmsys/vicuna-13b-v1.5",
|
||||
"lmsys/vicuna-7b-v1.5",
|
||||
"codellama/CodeLlama-13b-Instruct-hf",
|
||||
"codellama/CodeLlama-34b-Instruct-hf",
|
||||
"codellama/CodeLlama-70b-Instruct-hf",
|
||||
"codellama/CodeLlama-7b-Instruct-hf",
|
||||
"meta-llama/Llama-2-70b-chat-hf",
|
||||
"meta-llama/Llama-2-13b-chat-hf",
|
||||
"meta-llama/Llama-2-7b-chat-hf",
|
||||
"mistralai/Mistral-7B-Instruct-v0.1",
|
||||
"mistralai/Mistral-7B-Instruct-v0.2",
|
||||
"mistralai/Mixtral-8x7B-Instruct-v0.1",
|
||||
"NousResearch/Nous-Capybara-7B-V1p9",
|
||||
"NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
|
||||
"NousResearch/Nous-Hermes-2-Mixtral-8x7B-SFT",
|
||||
"NousResearch/Nous-Hermes-llama-2-7b",
|
||||
"NousResearch/Nous-Hermes-Llama2-13b",
|
||||
"NousResearch/Nous-Hermes-2-Yi-34B",
|
||||
"openchat/openchat-3.5-1210",
|
||||
"Open-Orca/Mistral-7B-OpenOrca",
|
||||
"togethercomputer/Qwen-7B-Chat",
|
||||
"Qwen/Qwen1.5-0.5B-Chat",
|
||||
"Qwen/Qwen1.5-1.8B-Chat",
|
||||
"Qwen/Qwen1.5-4B-Chat",
|
||||
"Qwen/Qwen1.5-7B-Chat",
|
||||
"Qwen/Qwen1.5-14B-Chat",
|
||||
"Qwen/Qwen1.5-72B-Chat",
|
||||
"snorkelai/Snorkel-Mistral-PairRM-DPO",
|
||||
"togethercomputer/alpaca-7b",
|
||||
"teknium/OpenHermes-2-Mistral-7B",
|
||||
"teknium/OpenHermes-2p5-Mistral-7B",
|
||||
"togethercomputer/falcon-40b-instruct",
|
||||
"togethercomputer/falcon-7b-instruct",
|
||||
"togethercomputer/Llama-2-7B-32K-Instruct",
|
||||
"togethercomputer/RedPajama-INCITE-Chat-3B-v1",
|
||||
"togethercomputer/RedPajama-INCITE-7B-Chat",
|
||||
"togethercomputer/StripedHyena-Nous-7B",
|
||||
"Undi95/ReMM-SLERP-L2-13B",
|
||||
"Undi95/Toppy-M-7B",
|
||||
"WizardLM/WizardLM-13B-V1.2",
|
||||
"upstage/SOLAR-10.7B-Instruct-v1.0",
|
||||
"gpt-4",
|
||||
"gpt-4-turbo",
|
||||
"gpt-4-0613",
|
||||
"gpt-4-32k",
|
||||
"gpt-4-32k-0613",
|
||||
"gpt-3.5-turbo-0125",
|
||||
"gpt-3.5-turbo",
|
||||
"gpt-3.5-turbo-1106",
|
||||
"gpt-3.5-turbo-instruct",
|
||||
"gpt-3.5-turbo-16k",
|
||||
"gpt-3.5-turbo-0613",
|
||||
"gpt-3.5-turbo-16k-0613",
|
||||
"gpt-4o",
|
||||
"claude-3-opus-20240229",
|
||||
"claude-3-sonnet-20240229",
|
||||
"claude-3-haiku-20240307",
|
||||
]
|
||||
|
||||
EMBEDDING_MODELS = [
|
||||
"text-embedding-3-small",
|
||||
"text-embedding-3-large",
|
||||
"text-embedding-ada-002",
|
||||
]
|
||||
|
|
@ -0,0 +1,34 @@
|
|||
from langflow.base.embeddings.model import LCEmbeddingsModel
|
||||
from langflow.base.models.aiml_constants import EMBEDDING_MODELS
|
||||
from langflow.components.embeddings.util.AIMLEmbeddingsImpl import AIMLEmbeddingsImpl
|
||||
from langflow.field_typing import Embeddings
|
||||
from langflow.inputs.inputs import DropdownInput
|
||||
from langflow.io import SecretStrInput
|
||||
|
||||
|
||||
class AIMLEmbeddingsComponent(LCEmbeddingsModel):
|
||||
display_name = "AI/ML Embeddings"
|
||||
description = "Generate embeddings using the AI/ML API."
|
||||
icon = "AI/ML"
|
||||
name = "AIMLEmbeddings"
|
||||
|
||||
inputs = [
|
||||
DropdownInput(
|
||||
name="model_name",
|
||||
display_name="Model Name",
|
||||
options=EMBEDDING_MODELS,
|
||||
required=True,
|
||||
),
|
||||
SecretStrInput(
|
||||
name="aiml_api_key",
|
||||
display_name="AI/ML API Key",
|
||||
value="AIML_API_KEY",
|
||||
required=True,
|
||||
),
|
||||
]
|
||||
|
||||
def build_embeddings(self) -> Embeddings:
|
||||
return AIMLEmbeddingsImpl(
|
||||
api_key=self.aiml_api_key,
|
||||
model=self.model_name,
|
||||
)
|
||||
|
|
@ -1,3 +1,4 @@
|
|||
from .AIMLEmbeddings import AIMLEmbeddingsComponent
|
||||
from .AmazonBedrockEmbeddings import AmazonBedrockEmbeddingsComponent
|
||||
from .AstraVectorize import AstraVectorizeComponent
|
||||
from .AzureOpenAIEmbeddings import AzureOpenAIEmbeddingsComponent
|
||||
|
|
@ -9,6 +10,7 @@ from .OpenAIEmbeddings import OpenAIEmbeddingsComponent
|
|||
from .VertexAIEmbeddings import VertexAIEmbeddingsComponent
|
||||
|
||||
__all__ = [
|
||||
"AIMLEmbeddingsComponent",
|
||||
"AmazonBedrockEmbeddingsComponent",
|
||||
"AstraVectorizeComponent",
|
||||
"AzureOpenAIEmbeddingsComponent",
|
||||
|
|
|
|||
|
|
@ -0,0 +1,70 @@
|
|||
import json
|
||||
from typing import List
|
||||
|
||||
import httpx
|
||||
from langflow.field_typing import Embeddings
|
||||
from langchain_core.runnables.config import run_in_executor
|
||||
from langchain_core.pydantic_v1 import BaseModel, SecretStr
|
||||
from loguru import logger
|
||||
|
||||
|
||||
class AIMLEmbeddingsImpl(BaseModel, Embeddings):
|
||||
embeddings_completion_url: str = "https://api.aimlapi.com/v1/embeddings"
|
||||
|
||||
api_key: SecretStr
|
||||
model: str
|
||||
|
||||
def embed_documents(self, texts: List[str]) -> List[List[float]]:
|
||||
result_vectors = []
|
||||
for text in texts:
|
||||
vector = self.embed_query(text)
|
||||
result_vectors.append(vector)
|
||||
|
||||
return result_vectors
|
||||
|
||||
def embed_query(self, text: str) -> List[float]:
|
||||
headers = {
|
||||
"Content-Type": "application/json",
|
||||
"Authorization": f"Bearer {self.api_key.get_secret_value()}",
|
||||
}
|
||||
|
||||
payload = {
|
||||
"model": self.model,
|
||||
"input": text,
|
||||
}
|
||||
vector = []
|
||||
try:
|
||||
response = httpx.post(
|
||||
self.embeddings_completion_url,
|
||||
headers=headers,
|
||||
json=payload,
|
||||
)
|
||||
try:
|
||||
response.raise_for_status()
|
||||
result_data = response.json()
|
||||
vector = result_data["data"][0]["embedding"]
|
||||
except httpx.HTTPStatusError as http_err:
|
||||
logger.error(f"HTTP error occurred: {http_err}")
|
||||
raise http_err
|
||||
except httpx.RequestError as req_err:
|
||||
logger.error(f"Request error occurred: {req_err}")
|
||||
raise req_err
|
||||
except json.JSONDecodeError:
|
||||
logger.warning(f"Failed to decode JSON, response text: {response.text}")
|
||||
except KeyError as key_err:
|
||||
logger.warning(f"Key error: {key_err}, response content: {result_data}")
|
||||
raise key_err
|
||||
except httpx.TimeoutException:
|
||||
logger.error("Request timed out.")
|
||||
raise
|
||||
except Exception as exc:
|
||||
logger.error(f"Error: {exc}")
|
||||
raise
|
||||
|
||||
return vector
|
||||
|
||||
async def aembed_documents(self, texts: List[str]) -> List[List[float]]:
|
||||
return await run_in_executor(None, self.embed_documents, texts)
|
||||
|
||||
async def aembed_query(self, text: str) -> List[float]:
|
||||
return await run_in_executor(None, self.embed_query, text)
|
||||
148
src/backend/base/langflow/components/models/AIMLModel.py
Normal file
148
src/backend/base/langflow/components/models/AIMLModel.py
Normal file
|
|
@ -0,0 +1,148 @@
|
|||
import json
|
||||
import httpx
|
||||
from langflow.base.models.aiml_constants import CHAT_MODELS
|
||||
from langflow.custom.custom_component.component import Component
|
||||
|
||||
from langflow.inputs.inputs import FloatInput, IntInput, MessageInput, SecretStrInput
|
||||
from langflow.schema.message import Message
|
||||
from langflow.template.field.base import Output
|
||||
from loguru import logger
|
||||
from pydantic.v1 import SecretStr
|
||||
|
||||
from langflow.inputs import (
|
||||
DropdownInput,
|
||||
StrInput,
|
||||
)
|
||||
|
||||
|
||||
class AIMLModelComponent(Component):
|
||||
display_name = "AI/ML API"
|
||||
description = "Generates text using the AI/ML API"
|
||||
icon = "AI/ML"
|
||||
chat_completion_url = "https://api.aimlapi.com/v1/chat/completions"
|
||||
|
||||
outputs = [
|
||||
Output(display_name="Text", name="text_output", method="make_request"),
|
||||
]
|
||||
|
||||
inputs = [
|
||||
DropdownInput(
|
||||
name="model_name",
|
||||
display_name="Model Name",
|
||||
options=CHAT_MODELS,
|
||||
required=True,
|
||||
),
|
||||
SecretStrInput(
|
||||
name="aiml_api_key",
|
||||
display_name="AI/ML API Key",
|
||||
value="AIML_API_KEY",
|
||||
),
|
||||
MessageInput(name="input_value", display_name="Input", required=True),
|
||||
IntInput(
|
||||
name="max_tokens",
|
||||
display_name="Max Tokens",
|
||||
advanced=True,
|
||||
info="The maximum number of tokens to generate. Set to 0 for unlimited tokens.",
|
||||
),
|
||||
StrInput(
|
||||
name="stop_tokens",
|
||||
display_name="Stop Tokens",
|
||||
info="Comma-separated list of tokens to signal the model to stop generating text.",
|
||||
advanced=True,
|
||||
),
|
||||
IntInput(
|
||||
name="top_k",
|
||||
display_name="Top K",
|
||||
info="Limits token selection to top K. (Default: 40)",
|
||||
advanced=True,
|
||||
),
|
||||
FloatInput(
|
||||
name="top_p",
|
||||
display_name="Top P",
|
||||
info="Works together with top-k. (Default: 0.9)",
|
||||
advanced=True,
|
||||
),
|
||||
FloatInput(
|
||||
name="repeat_penalty",
|
||||
display_name="Repeat Penalty",
|
||||
info="Penalty for repetitions in generated text. (Default: 1.1)",
|
||||
advanced=True,
|
||||
),
|
||||
FloatInput(
|
||||
name="temperature",
|
||||
display_name="Temperature",
|
||||
value=0.2,
|
||||
info="Controls the creativity of model responses.",
|
||||
),
|
||||
StrInput(
|
||||
name="system_message",
|
||||
display_name="System Message",
|
||||
info="System message to pass to the model.",
|
||||
advanced=True,
|
||||
),
|
||||
]
|
||||
|
||||
def make_request(self) -> Message:
|
||||
api_key = SecretStr(self.aiml_api_key) if self.aiml_api_key else None
|
||||
|
||||
headers = {
|
||||
"Content-Type": "application/json",
|
||||
"Authorization": f"Bearer {api_key.get_secret_value()}" if api_key else "",
|
||||
}
|
||||
|
||||
messages = []
|
||||
if self.system_message:
|
||||
messages.append({"role": "system", "content": self.system_message})
|
||||
|
||||
if self.input_value:
|
||||
if isinstance(self.input_value, Message):
|
||||
# Though we aren't using langchain here, the helper method is useful
|
||||
message = self.input_value.to_lc_message()
|
||||
if message.type == "human":
|
||||
messages.append({"role": "user", "content": message.content})
|
||||
else:
|
||||
raise ValueError(f"Expected user message, saw: {message.type}")
|
||||
else:
|
||||
raise TypeError(f"Expected Message type, saw: {type(self.input_value)}")
|
||||
else:
|
||||
raise ValueError("Please provide an input value")
|
||||
|
||||
payload = {
|
||||
"model": self.model_name,
|
||||
"messages": messages,
|
||||
"max_tokens": self.max_tokens or None,
|
||||
"temperature": self.temperature or 0.2,
|
||||
"top_k": self.top_k or 40,
|
||||
"top_p": self.top_p or 0.9,
|
||||
"repeat_penalty": self.repeat_penalty or 1.1,
|
||||
"stop_tokens": self.stop_tokens or None,
|
||||
}
|
||||
|
||||
try:
|
||||
response = httpx.post(self.chat_completion_url, headers=headers, json=payload)
|
||||
try:
|
||||
response.raise_for_status()
|
||||
result_data = response.json()
|
||||
choice = result_data["choices"][0]
|
||||
result = choice["message"]["content"]
|
||||
except httpx.HTTPStatusError as http_err:
|
||||
logger.error(f"HTTP error occurred: {http_err}")
|
||||
raise http_err
|
||||
except httpx.RequestError as req_err:
|
||||
logger.error(f"Request error occurred: {req_err}")
|
||||
raise req_err
|
||||
except json.JSONDecodeError:
|
||||
logger.warning("Failed to decode JSON, response text: {response.text}")
|
||||
result = response.text
|
||||
except KeyError as key_err:
|
||||
logger.warning(f"Key error: {key_err}, response content: {result_data}")
|
||||
raise key_err
|
||||
|
||||
self.status = result
|
||||
except httpx.TimeoutException:
|
||||
return Message(text="Request timed out.")
|
||||
except Exception as exc:
|
||||
logger.error(f"Error: {exc}")
|
||||
raise
|
||||
|
||||
return Message(text=result)
|
||||
|
|
@ -45,7 +45,7 @@ class MistralAIModelComponent(LCModelComponent):
|
|||
),
|
||||
),
|
||||
SecretStrInput(
|
||||
name="mistral_api_key",
|
||||
name="api_key",
|
||||
display_name="Mistral API Key",
|
||||
info="The Mistral API Key to use for the Mistral model.",
|
||||
advanced=False,
|
||||
|
|
@ -67,7 +67,7 @@ class MistralAIModelComponent(LCModelComponent):
|
|||
]
|
||||
|
||||
def build_model(self) -> LanguageModel: # type: ignore[type-var]
|
||||
mistral_api_key = self.mistral_api_key
|
||||
mistral_api_key = self.api_key
|
||||
temperature = self.temperature
|
||||
model_name = self.model_name
|
||||
max_tokens = self.max_tokens
|
||||
|
|
|
|||
|
|
@ -76,7 +76,7 @@ class ChatOllamaComponent(LCModelComponent):
|
|||
value="http://localhost:11434",
|
||||
),
|
||||
DropdownInput(
|
||||
name="model",
|
||||
name="model_name",
|
||||
display_name="Model Name",
|
||||
value="llama2",
|
||||
info="Refer to https://ollama.ai/library for more models.",
|
||||
|
|
@ -107,6 +107,7 @@ class ChatOllamaComponent(LCModelComponent):
|
|||
info="Enable/disable Mirostat sampling for controlling perplexity.",
|
||||
value="Disabled",
|
||||
advanced=True,
|
||||
real_time_refresh=True,
|
||||
),
|
||||
FloatInput(
|
||||
name="mirostat_eta",
|
||||
|
|
@ -238,7 +239,7 @@ class ChatOllamaComponent(LCModelComponent):
|
|||
# Mapping system settings to their corresponding values
|
||||
llm_params = {
|
||||
"base_url": self.base_url,
|
||||
"model": self.model,
|
||||
"model": self.model_name,
|
||||
"mirostat": mirostat_value,
|
||||
"format": self.format,
|
||||
"metadata": self.metadata,
|
||||
|
|
|
|||
|
|
@ -58,7 +58,7 @@ class OpenAIModelComponent(LCModelComponent):
|
|||
info="The base URL of the OpenAI API. Defaults to https://api.openai.com/v1. You can change this to use other APIs like JinaChat, LocalAI and Prem.",
|
||||
),
|
||||
SecretStrInput(
|
||||
name="openai_api_key",
|
||||
name="api_key",
|
||||
display_name="OpenAI API Key",
|
||||
info="The OpenAI API Key to use for the OpenAI model.",
|
||||
advanced=False,
|
||||
|
|
@ -82,10 +82,10 @@ class OpenAIModelComponent(LCModelComponent):
|
|||
]
|
||||
|
||||
def build_model(self) -> LanguageModel: # type: ignore[type-var]
|
||||
# self.output_schea is a list of dictionarie s
|
||||
# self.output_schema is a list of dictionaries
|
||||
# let's convert it to a dictionary
|
||||
output_schema_dict: dict[str, str] = reduce(operator.ior, self.output_schema or {}, {})
|
||||
openai_api_key = self.openai_api_key
|
||||
openai_api_key = self.api_key
|
||||
temperature = self.temperature
|
||||
model_name: str = self.model_name
|
||||
max_tokens = self.max_tokens
|
||||
|
|
|
|||
|
|
@ -1,3 +1,4 @@
|
|||
from .AIMLModel import AIMLModelComponent
|
||||
from .AmazonBedrockModel import AmazonBedrockComponent
|
||||
from .AnthropicModel import AnthropicModelComponent
|
||||
from .AzureOpenAIModel import AzureChatOpenAIComponent
|
||||
|
|
@ -10,6 +11,7 @@ from .OpenAIModel import OpenAIModelComponent
|
|||
from .VertexAiModel import ChatVertexAIComponent
|
||||
|
||||
__all__ = [
|
||||
"AIMLModelComponent",
|
||||
"AmazonBedrockComponent",
|
||||
"AnthropicModelComponent",
|
||||
"AzureChatOpenAIComponent",
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue