github/langchain
1
0
Fork 0
mirror of https://github.com/hwchase17/langchain.git synced 2025-06-25 08:03:39 +00:00
Code Issues Packages Projects Releases Wiki Activity

oai assistant multiple actions (#13068)

Browse Source
This commit is contained in:
Bagatur 2023-11-08 08:25:37 -08:00 committed by GitHub
parent a9b70baef9
commit 55aeff6777
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 78 additions and 90 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

121
cookbook/openai_v1_cookbook.ipynb
View File

@ -17,12 +17,12 @@
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"!pip install -U openai \"langchain>=0.0.331rc1\" langchain-experimental" "!pip install -U openai \"langchain>=0.0.331rc2\" langchain-experimental"
] ]
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 25, "execution_count": 1,
"id": "c3e067ce-7a43-47a7-bc89-41f1de4cf136", "id": "c3e067ce-7a43-47a7-bc89-41f1de4cf136",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
@ -43,17 +43,17 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 26, "execution_count": 2,
"id": "1c8c3965-d3c9-4186-b5f3-5e67855ef916", "id": "1c8c3965-d3c9-4186-b5f3-5e67855ef916",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [
{ {
"data": { "data": {
"text/plain": [ "text/plain": [
"AIMessage(content='This image appears to be a diagram representing the architecture or components of a software system or platform named \"LangChain.\" The diagram is organized into various blocks that represent different layers or aspects of the system. Here\\'s a breakdown of the labeled parts:\\n\\n1. **LangSmith**: This seems to be a tool or component related to testing, evaluation, monitoring, feedback, and annotation, as well as debugging.\\n\\n2. **LangServe**: This component chains as REST API, suggesting it may serve as an interface for the system\\'s functionality to be accessed over the web using RESTful API calls.\\n\\n3. **Templates**: Reference applications are mentioned here, which might be pre-built applications or use cases provided as starting points for users of the system.\\n\\n4. **Chains, agents, agent executors**: This part refers to common application logic, indicating modular components that can be combined or chained together to create complex workflows.\\n\\n5. **Model I/O**: This includes elements such as prompt, example selector, model, and output parser, which may deal with the input and output processes of a machine learning model or similar.\\n\\n6. **Retrieval**: Components under this section, like document loader, text splitter, embedding model, vector store, and retriever, imply functionality related to')" "AIMessage(content='The image appears to be a diagram representing the architecture or components of a software system or framework related to language processing, possibly named LangChain or associated with a project or product called LangChain, based on the prominent appearance of that term. The diagram is organized into several layers or aspects, each containing various elements or modules:\\n\\n1. **Protocol**: This may be the foundational layer, which includes \"LCEL\" and terms like parallelization, fallbacks, tracing, batching, streaming, async, and composition. These seem related to communication and execution protocols for the system.\\n\\n2. **Integrations Components**: This layer includes \"Model I/O\" with elements such as the model, output parser, prompt, and example selector. It also has a \"Retrieval\" section with a document loader, retriever, embedding model, vector store, and text splitter. Lastly, there\\'s an \"Agent Tooling\" section. These components likely deal with the interaction with external data, models, and tools.\\n\\n3. **Application**: The application layer features \"LangChain\" with chains, agents, agent executors, and common application logic. This suggests that the system uses a modular approach with chains and agents to process language tasks.\\n\\n4. **Deployment**: This contains \"Lang')"
] ]
}, },
"execution_count": 26, "execution_count": 2,
"metadata": {}, "metadata": {},
"output_type": "execute_result" "output_type": "execute_result"
} }
@ -103,27 +103,27 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 4, "execution_count": 1,
"id": "a9064bbe-d9f7-4a29-a7b3-73933b3197e7", "id": "a9064bbe-d9f7-4a29-a7b3-73933b3197e7",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"from langchain_experimental.openai_assistant import OpenAIAssistantRunnable\n" "from langchain_experimental.openai_assistant import OpenAIAssistantRunnable"
] ]
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 5, "execution_count": 2,
"id": "7a20a008-49ac-46d2-aa26-b270118af5ea", "id": "7a20a008-49ac-46d2-aa26-b270118af5ea",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [
{ {
"data": { "data": {
"text/plain": [ "text/plain": [
"[ThreadMessage(id='msg_RGOsJ2RBYp79rILrZ0NsAX68', assistant_id='asst_9Xb1ZgefoAbp2V5ddRlDGisy', content=[MessageContentText(text=Text(annotations=[], value='\\\\( 10 - 4^{2.7} \\\\) is approximately -32.2243.'), type='text')], created_at=1699385426, file_ids=[], metadata={}, object='thread.message', role='assistant', run_id='run_E2PRoP04ryly4p5ds5ek2qxs', thread_id='thread_pu9CpsYIWWZtxemZxpbYfhm4')]" "[ThreadMessage(id='msg_g9OJv0rpPgnc3mHmocFv7OVd', assistant_id='asst_hTwZeNMMphxzSOqJ01uBMsJI', content=[MessageContentText(text=Text(annotations=[], value='The result of \\\\(10 - 4^{2.7}\\\\) is approximately \\\\(-32.224\\\\).'), type='text')], created_at=1699460600, file_ids=[], metadata={}, object='thread.message', role='assistant', run_id='run_nBIT7SiAwtUfSCTrQNSPLOfe', thread_id='thread_14n4GgXwxgNL0s30WJW5F6p0')]"
] ]
}, },
"execution_count": 5, "execution_count": 2,
"metadata": {}, "metadata": {},
"output_type": "execute_result" "output_type": "execute_result"
} }
@ -151,26 +151,36 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 7, "execution_count": null,
"id": "48681ac7-b267-48d4-972c-8a7df8393a21", "id": "ee4cc355-f2d6-4c51-bcf7-f502868357d3",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"from langchain.tools import E2BDataAnalysisTool\n", "!pip install e2b duckduckgo-search"
"\n",
"tools = [E2BDataAnalysisTool(api_key=\"...\")]"
] ]
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 8, "execution_count": 3,
"id": "48681ac7-b267-48d4-972c-8a7df8393a21",
"metadata": {},
"outputs": [],
"source": [
"from langchain.tools import E2BDataAnalysisTool, DuckDuckGoSearchRun\n",
"\n",
"tools = [E2BDataAnalysisTool(api_key=\"...\"), DuckDuckGoSearchRun()]"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "1c01dd79-dd3e-4509-a2e2-009a7f99f16a", "id": "1c01dd79-dd3e-4509-a2e2-009a7f99f16a",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
"source": [ "source": [
"agent = OpenAIAssistantRunnable.create_assistant(\n", "agent = OpenAIAssistantRunnable.create_assistant(\n",
" name=\"langchain assistant e2b tool\",\n", " name=\"langchain assistant e2b tool\",\n",
" instructions=\"You are a personal math tutor. Write and run code to answer math questions.\",\n", " instructions=\"You are a personal math tutor. Write and run code to answer math questions. You can also search the internet.\",\n",
" tools=tools,\n", " tools=tools,\n",
" model=\"gpt-4-1106-preview\",\n", " model=\"gpt-4-1106-preview\",\n",
" as_agent=True\n", " as_agent=True\n",
@ -187,18 +197,18 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 9, "execution_count": 5,
"id": "1f137f94-801f-4766-9ff5-2de9df5e8079", "id": "1f137f94-801f-4766-9ff5-2de9df5e8079",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [
{ {
"data": { "data": {
"text/plain": [ "text/plain": [
"{'content': \"What's 10 - 4 raised to the 2.7\",\n", "{'content': \"What's the weather in SF today divided by 2.7\",\n",
" 'output': 'The result of \\\\(10 - 4\\\\) raised to the power of \\\\(2.7\\\\) is approximately \\\\(126.19\\\\).'}" " 'output': \"The weather in San Francisco today is reported to have temperatures as high as 66 °F. To get the temperature divided by 2.7, we will calculate that:\\n\\n66 °F / 2.7 = 24.44 °F\\n\\nSo, when the high temperature of 66 °F is divided by 2.7, the result is approximately 24.44 °F. Please note that this doesn't have a meteorological meaning; it's purely a mathematical operation based on the given temperature.\"}"
] ]
}, },
"execution_count": 9, "execution_count": 5,
"metadata": {}, "metadata": {},
"output_type": "execute_result" "output_type": "execute_result"
} }
@ -207,7 +217,7 @@
"from langchain.agents import AgentExecutor\n", "from langchain.agents import AgentExecutor\n",
"\n", "\n",
"agent_executor = AgentExecutor(agent=agent, tools=tools)\n", "agent_executor = AgentExecutor(agent=agent, tools=tools)\n",
"agent_executor.invoke({\"content\": \"What's 10 - 4 raised to the 2.7\"})" "agent_executor.invoke({\"content\": \"What's the weather in SF today divided by 2.7\"})"
] ]
}, },
{ {
@ -220,7 +230,7 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 19, "execution_count": 6,
"id": "c0475fa7-b6c1-4331-b8e2-55407466c724", "id": "c0475fa7-b6c1-4331-b8e2-55407466c724",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
@ -236,7 +246,7 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 22, "execution_count": 7,
"id": "b76cb669-6aba-4827-868f-00aa960026f2", "id": "b76cb669-6aba-4827-868f-00aa960026f2",
"metadata": {}, "metadata": {},
"outputs": [], "outputs": [],
@ -259,7 +269,7 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 23, "execution_count": 8,
"id": "7946116a-b82f-492e-835e-ca958a8949a5", "id": "7946116a-b82f-492e-835e-ca958a8949a5",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [
@ -267,9 +277,9 @@
"name": "stdout", "name": "stdout",
"output_type": "stream", "output_type": "stream",
"text": [ "text": [
"e2b_data_analysis {'python_code': 'print((10 - 4) ** 2.7)'} {\"stdout\": \"126.18518711065899\", \"stderr\": \"\", \"artifacts\": []}\n", "e2b_data_analysis {'python_code': 'print(10 - 4 ** 2.7)'} {\"stdout\": \"-32.22425314473263\", \"stderr\": \"\", \"artifacts\": []}\n",
"\n", "\n",
"\\( 10 - 4 \\) raised to the power of 2.7 is approximately 126.185.\n" "\\( 10 - 4^{2.7} \\) is approximately \\(-32.22425314473263\\).\n"
] ]
} }
], ],
@ -280,7 +290,7 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 24, "execution_count": 9,
"id": "f2744a56-9f4f-4899-827a-fa55821c318c", "id": "f2744a56-9f4f-4899-827a-fa55821c318c",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [
@ -288,9 +298,9 @@
"name": "stdout", "name": "stdout",
"output_type": "stream", "output_type": "stream",
"text": [ "text": [
"e2b_data_analysis {'python_code': 'result = (10 - 4) ** 2.7 + 17.241\\nprint(result)'} {\"stdout\": \"143.426187110659\", \"stderr\": \"\", \"artifacts\": []}\n", "e2b_data_analysis {'python_code': 'result = 10 - 4 ** 2.7\\nprint(result + 17.241)'} {\"stdout\": \"-14.983253144732629\", \"stderr\": \"\", \"artifacts\": []}\n",
"\n", "\n",
"\\( (10 - 4)^{2.7} + 17.241 \\) is approximately 143.426.\n" "When you add \\( 17.241 \\) to \\( 10 - 4^{2.7} \\), the result is approximately \\( -14.98325314473263 \\).\n"
] ]
} }
], ],
@ -313,29 +323,10 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 16, "execution_count": null,
"id": "db6072c4-f3f3-415d-872b-71ea9f3c02bb", "id": "db6072c4-f3f3-415d-872b-71ea9f3c02bb",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [],
{
"name": "stdout",
"output_type": "stream",
"text": [
"{\n",
" \"companies\": [\n",
" {\n",
" \"name\": \"Google\",\n",
" \"origin\": \"USA\"\n",
" },\n",
" {\n",
" \"name\": \"Deepmind\",\n",
" \"origin\": \"UK\"\n",
" }\n",
" ]\n",
"}\n"
]
}
],
"source": [ "source": [
"chat = ChatOpenAI(model=\"gpt-3.5-turbo-1106\").bind(\n", "chat = ChatOpenAI(model=\"gpt-3.5-turbo-1106\").bind(\n",
" response_format={\"type\": \"json_object\"}\n", " response_format={\"type\": \"json_object\"}\n",
@ -356,22 +347,10 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 17, "execution_count": null,
"id": "08e00ccf-b991-4249-846b-9500a0ccbfa0", "id": "08e00ccf-b991-4249-846b-9500a0ccbfa0",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [],
{
"data": {
"text/plain": [
"{'companies': [{'name': 'Google', 'origin': 'USA'},\n",
" {'name': 'Deepmind', 'origin': 'UK'}]}"
]
},
"execution_count": 17,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [ "source": [
"import json\n", "import json\n",
"\n", "\n",
@ -390,18 +369,10 @@
}, },
{ {
"cell_type": "code", "cell_type": "code",
"execution_count": 18, "execution_count": null,
"id": "1281883c-bf8f-4665-89cd-4f33ccde69ab", "id": "1281883c-bf8f-4665-89cd-4f33ccde69ab",
"metadata": {}, "metadata": {},
"outputs": [ "outputs": [],
{
"name": "stdout",
"output_type": "stream",
"text": [
"{'token_usage': {'completion_tokens': 43, 'prompt_tokens': 49, 'total_tokens': 92}, 'model_name': 'gpt-3.5-turbo-1106', 'system_fingerprint': 'fp_eeff13170a'}\n"
]
}
],
"source": [ "source": [
"chat = ChatOpenAI(model=\"gpt-3.5-turbo-1106\")\n", "chat = ChatOpenAI(model=\"gpt-3.5-turbo-1106\")\n",
"output = chat.generate(\n", "output = chat.generate(\n",

47
libs/experimental/langchain_experimental/openai_assistant/base.py
View File

@ -2,7 +2,7 @@ from __future__ import annotations
import json import json
from time import sleep from time import sleep
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Sequence, Union from typing import TYPE_CHECKING, Any, Dict, List, Optional, Sequence, Tuple, Union
from langchain.pydantic_v1 import Field from langchain.pydantic_v1 import Field
from langchain.schema.agent import AgentAction, AgentFinish from langchain.schema.agent import AgentAction, AgentFinish
@ -212,8 +212,12 @@ class OpenAIAssistantRunnable(RunnableSerializable[Dict, OutputType]):
will return OpenAI types will return OpenAI types
Union[List[ThreadMessage], List[RequiredActionFunctionToolCall]]. Union[List[ThreadMessage], List[RequiredActionFunctionToolCall]].
""" """
input = self._parse_input(input) # Being run within AgentExecutor and there are tool outputs to submit.
if "thread_id" not in input: if self.as_agent and input.get("intermediate_steps"):
tool_outputs = self._parse_intermediate_steps(input["intermediate_steps"])
run = self.client.beta.threads.runs.submit_tool_outputs(**tool_outputs)
# Starting a new thread and a new run.
elif "thread_id" not in input:
thread = { thread = {
"messages": [ "messages": [
{ {
@ -226,6 +230,7 @@ class OpenAIAssistantRunnable(RunnableSerializable[Dict, OutputType]):
"metadata": input.get("thread_metadata"), "metadata": input.get("thread_metadata"),
} }
run = self._create_thread_and_run(input, thread) run = self._create_thread_and_run(input, thread)
# Starting a new run in an existing thread.
elif "run_id" not in input: elif "run_id" not in input:
_ = self.client.beta.threads.messages.create( _ = self.client.beta.threads.messages.create(
input["thread_id"], input["thread_id"],
@ -235,21 +240,31 @@ class OpenAIAssistantRunnable(RunnableSerializable[Dict, OutputType]):
metadata=input.get("message_metadata"), metadata=input.get("message_metadata"),
) )
run = self._create_run(input) run = self._create_run(input)
# Submitting tool outputs to an existing run, outside the AgentExecutor
# framework.
else: else:
run = self.client.beta.threads.runs.submit_tool_outputs(**input) run = self.client.beta.threads.runs.submit_tool_outputs(**input)
return self._get_response(run.id, run.thread_id) return self._get_response(run.id, run.thread_id)
def _parse_input(self, input: dict) -> dict: def _parse_intermediate_steps(
if self.as_agent and input.get("intermediate_steps"): self, intermediate_steps: List[Tuple[OpenAIAssistantAction, str]]
last_action, last_output = input["intermediate_steps"][-1] ) -> dict:
input = { last_action, last_output = intermediate_steps[-1]
"tool_outputs": [ run = self._wait_for_run(last_action.run_id, last_action.thread_id)
{"output": last_output, "tool_call_id": last_action.tool_call_id} required_tool_call_ids = {
], tc.id for tc in run.required_action.submit_tool_outputs.tool_calls
"run_id": last_action.run_id, }
"thread_id": last_action.thread_id, tool_outputs = [
} {"output": output, "tool_call_id": action.tool_call_id}
return input for action, output in intermediate_steps
if action.tool_call_id in required_tool_call_ids
]
submit_tool_outputs = {
"tool_outputs": tool_outputs,
"run_id": last_action.run_id,
"thread_id": last_action.thread_id,
}
return submit_tool_outputs
def _create_run(self, input: dict) -> Any: def _create_run(self, input: dict) -> Any:
params = { params = {
@ -307,6 +322,8 @@ class OpenAIAssistantRunnable(RunnableSerializable[Dict, OutputType]):
for tool_call in run.required_action.submit_tool_outputs.tool_calls: for tool_call in run.required_action.submit_tool_outputs.tool_calls:
function = tool_call.function function = tool_call.function
args = json.loads(function.arguments) args = json.loads(function.arguments)
if len(args) == 1 and "__arg1" in args:
args = args["__arg1"]
actions.append( actions.append(
OpenAIAssistantAction( OpenAIAssistantAction(
tool=function.name, tool=function.name,
@ -321,7 +338,7 @@ class OpenAIAssistantRunnable(RunnableSerializable[Dict, OutputType]):
else: else:
run_info = json.dumps(run.dict(), indent=2) run_info = json.dumps(run.dict(), indent=2)
raise ValueError( raise ValueError(
f"Unknown run status {run.status}. Full run info:\n\n{run_info})" f"Unexpected run status: {run.status}. Full run info:\n\n{run_info})"
) )
def _wait_for_run(self, run_id: str, thread_id: str) -> Any: def _wait_for_run(self, run_id: str, thread_id: str) -> Any: