Chapter 30
double texting
NotebookPython 3 (ipykernel)17 cells
Double Texting
Seamless handling of double texting is important for handling real-world usage scenarios, especially in chat applications.
Users can send multiple messages in a row before the prior run(s) complete, and we want to ensure that we handle this gracefully.
Reject
A simple approach is to reject any new runs until the current run completes.
In [13]python · cell 2
python
%%capture --no-stderr
%pip install -U langgraph_sdkIn [2]python · cell 3
python
from langgraph_sdk import get_client
url_for_cli_deployment = "http://localhost:8123"
client = get_client(url=url_for_cli_deployment)In [18]python · cell 4
python
import httpx
from langchain_core.messages import HumanMessage
# Create a thread
thread = await client.threads.create()
# Create to dos
user_input_1 = "Add a ToDo to follow-up with DI Repairs."
user_input_2 = "Add a ToDo to mount dresser to the wall."
config = {"configurable": {"user_id": "Test-Double-Texting"}}
graph_name = "task_maistro"
run = await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_1)]},
config=config,
)
try:
await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_2)]},
config=config,
multitask_strategy="reject",
)
except httpx.HTTPStatusError as e:
print("Failed to start concurrent run", e)Output
Failed to start concurrent run Client error '409 Conflict' for url 'http://localhost:8123/threads/2b58630e-00fd-4c35-afad-a6b59e9b9104/runs' For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/409
In [19]python · cell 5
python
from langchain_core.messages import convert_to_messages
# Wait until the original run completes
await client.runs.join(thread["thread_id"], run["run_id"])
# Get the state of the thread
state = await client.threads.get_state(thread["thread_id"])
for m in convert_to_messages(state["values"]["messages"]):
m.pretty_print()Output
================================[1m Human Message [0m=================================
Add a ToDo to follow-up with DI Repairs.
==================================[1m Ai Message [0m==================================
Tool Calls:
UpdateMemory (call_6xqHubCPNufS0bg4tbUxC0FU)
Call ID: call_6xqHubCPNufS0bg4tbUxC0FU
Args:
update_type: todo
=================================[1m Tool Message [0m=================================
New ToDo created:
Content: {'task': 'Follow-up with DI Repairs', 'time_to_complete': 30, 'deadline': None, 'solutions': ['Call DI Repairs customer service', 'Email DI Repairs support', 'Check DI Repairs website for updates'], 'status': 'not started'}
==================================[1m Ai Message [0m==================================
I've added a task to follow-up with DI Repairs to your ToDo list. If there's anything else you need, feel free to let me know!
Enqueue
We can use enqueue any new runs until the current run completes.
In [20]python · cell 7
python
# Create a new thread
thread = await client.threads.create()
# Create new ToDos
user_input_1 = "Send Erik his t-shirt gift this weekend."
user_input_2 = "Get cash and pay nanny for 2 weeks. Do this by Friday."
config = {"configurable": {"user_id": "Test-Double-Texting"}}
graph_name = "task_maistro"
first_run = await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_1)]},
config=config,
)
second_run = await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_2)]},
config=config,
multitask_strategy="enqueue",
)
# Wait until the second run completes
await client.runs.join(thread["thread_id"], second_run["run_id"])
# Get the state of the thread
state = await client.threads.get_state(thread["thread_id"])
for m in convert_to_messages(state["values"]["messages"]):
m.pretty_print()Output
================================[1m Human Message [0m=================================
Send Erik his t-shirt gift this weekend.
==================================[1m Ai Message [0m==================================
Tool Calls:
UpdateMemory (call_svTeXPmWGTLY8aQ8EifjwHAa)
Call ID: call_svTeXPmWGTLY8aQ8EifjwHAa
Args:
update_type: todo
=================================[1m Tool Message [0m=================================
New ToDo created:
Content: {'task': 'Send Erik his t-shirt gift', 'time_to_complete': 30, 'deadline': '2024-11-19T23:59:00', 'solutions': ['Wrap the t-shirt', "Get Erik's address", 'Visit the post office', 'Choose a delivery service'], 'status': 'not started'}
==================================[1m Ai Message [0m==================================
I've updated your ToDo list to send Erik his t-shirt gift this weekend. If there's anything else you need, feel free to let me know!
================================[1m Human Message [0m=================================
Get cash and pay nanny for 2 weeks. Do this by Friday.
==================================[1m Ai Message [0m==================================
Tool Calls:
UpdateMemory (call_Cq0Tfn6yqccHH8n0DOucz5OQ)
Call ID: call_Cq0Tfn6yqccHH8n0DOucz5OQ
Args:
update_type: todo
=================================[1m Tool Message [0m=================================
New ToDo created:
Content: {'task': 'Get cash and pay nanny for 2 weeks', 'time_to_complete': 15, 'deadline': '2024-11-17T23:59:00', 'solutions': ['Visit the ATM', 'Calculate the total amount for 2 weeks', 'Hand over the cash to the nanny'], 'status': 'not started'}
Document af1fe011-f3c5-4c1c-b98b-181869bc2944 updated:
Plan: Update the deadline for sending Erik his t-shirt gift to this weekend, which is by 2024-11-17.
Added content: 2024-11-17T23:59:00
==================================[1m Ai Message [0m==================================
I've updated your ToDo list to ensure you get cash and pay the nanny for 2 weeks by Friday. Let me know if there's anything else you need!
Interrupt
We can use interrupt to interrupt the current run, but save all the work that has been done so far up to that point.
In [32]python · cell 9
python
import asyncio
# Create a new thread
thread = await client.threads.create()
# Create new ToDos
user_input_1 = "Give me a summary of my ToDos due tomrrow."
user_input_2 = "Never mind, create a ToDo to Order Ham for Thanksgiving by next Friday."
config = {"configurable": {"user_id": "Test-Double-Texting"}}
graph_name = "task_maistro"
interrupted_run = await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_1)]},
config=config,
)
# Wait for some of run 1 to complete so that we can see it in the thread
await asyncio.sleep(1)
second_run = await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_2)]},
config=config,
multitask_strategy="interrupt",
)
# Wait until the second run completes
await client.runs.join(thread["thread_id"], second_run["run_id"])
# Get the state of the thread
state = await client.threads.get_state(thread["thread_id"])
for m in convert_to_messages(state["values"]["messages"]):
m.pretty_print()Output
================================[1m Human Message [0m=================================
Give me a summary of my ToDos due tomrrow.
================================[1m Human Message [0m=================================
Never mind, create a ToDo to Order Ham for Thanksgiving by next Friday.
==================================[1m Ai Message [0m==================================
Tool Calls:
UpdateMemory (call_Rk80tTSJzik2oY44tyUWk8FM)
Call ID: call_Rk80tTSJzik2oY44tyUWk8FM
Args:
update_type: todo
=================================[1m Tool Message [0m=================================
New ToDo created:
Content: {'task': 'Order Ham for Thanksgiving', 'time_to_complete': 30, 'deadline': '2024-11-22T23:59:59', 'solutions': ['Check local grocery stores for availability', 'Order online from a specialty meat provider', 'Visit a local butcher shop'], 'status': 'not started'}
==================================[1m Ai Message [0m==================================
I've added the task "Order Ham for Thanksgiving" to your ToDo list with a deadline of next Friday. If you need any more help, feel free to ask!
We can see the initial run is saved, and has status interrupted.
In [33]python · cell 11
python
# Confirm that the first run was interrupted
print((await client.runs.get(thread["thread_id"], interrupted_run["run_id"]))["status"])Output
interrupted
Rollback
We can use rollback to interrupt the prior run of the graph, delete it, and start a new run with the double-texted input.
In [28]python · cell 13
python
# Create a new thread
thread = await client.threads.create()
# Create new ToDos
user_input_1 = "Add a ToDo to call to make appointment at Yoga."
user_input_2 = "Actually, add a ToDo to drop by Yoga in person on Sunday."
config = {"configurable": {"user_id": "Test-Double-Texting"}}
graph_name = "task_maistro"
rolled_back_run = await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_1)]},
config=config,
)
second_run = await client.runs.create(
thread["thread_id"],
graph_name,
input={"messages": [HumanMessage(content=user_input_2)]},
config=config,
multitask_strategy="rollback",
)
# Wait until the second run completes
await client.runs.join(thread["thread_id"], second_run["run_id"])
# Get the state of the thread
state = await client.threads.get_state(thread["thread_id"])
for m in convert_to_messages(state["values"]["messages"]):
m.pretty_print()Output
================================[1m Human Message [0m================================= Actually, add a ToDo to drop by Yoga in person on Sunday. ==================================[1m Ai Message [0m================================== It looks like the task "Drop by Yoga in person" is already on your ToDo list with a deadline of November 19, 2024. Would you like me to update the deadline to the upcoming Sunday instead?
The initial run was deleted.
In [29]python · cell 15
python
# Confirm that the original run was deleted
try:
await client.runs.get(thread["thread_id"], rolled_back_run["run_id"])
except httpx.HTTPStatusError as _:
print("Original run was correctly deleted")Output
Original run was correctly deleted
Summary
We can see all the methods summarized:
In [ ]python · cell 17
python