issue_owner_repo listlengths 2 2 | issue_body stringlengths 0 261k ⌀ | issue_title stringlengths 1 925 | issue_comments_url stringlengths 56 81 | issue_comments_count int64 0 2.5k | issue_created_at stringlengths 20 20 | issue_updated_at stringlengths 20 20 | issue_html_url stringlengths 37 62 | issue_github_id int64 387k 2.46B | issue_number int64 1 127k |
|---|---|---|---|---|---|---|---|---|---|
[
"langchain-ai",
"langchain"
] | ### System Info
Windows 10
Python 3.10.11
I am using FAISS as the vectorstore, HuggingFaceEmbeddings, and the llm is HuggingFaceHub with google/flan-t5-small. There are no changes to any of these elements when running with and without 'with_sources'.
All libraries were installed within the last week.
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [X] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
embeddings = HuggingFaceEmbeddings()
db = FAISS.load_local("FAISS-DB", embeddings)
llm = HuggingFaceHub(repo_id="google/flan-t5-small", model_kwargs={"temperature": 0.2, "max_length": 256})
query = "what does the company sodexo provide?"
similarity_docs = db.similarity_search(query)
chain = load_qa_chain(llm, chain_type="stuff")
results = chain.run(input_documents=similarity_docs, question=query)
print(query)
print(results)
similarity_docs = db.similarity_search(query)
chain = load_qa_with_sources_chain (llm, chain_type="stuff")
results = chain.run(input_documents=similarity_docs, question=query)
print(query)
print(results)
------------------------------
Output:
what does the company sodexo provide?
integrated food, facilities management and other services
what does the company sodexo provide?
0
[similarity_docs.txt](https://github.com/langchain-ai/langchain/files/12918308/similarity_docs.txt)
### Expected behavior
The expected result was the same answer with the source, similar to {'answer': 'integrated food, facilities management and other services', 'source': 'Inputs\\bodyExtract_71756.txt'} | load_qa_chain succeeds where load_qa_with_sources_chain fails to return result | https://api.github.com/repos/langchain-ai/langchain/issues/11865/comments | 2 | 2023-10-16T14:55:08Z | 2024-02-06T16:19:46Z | https://github.com/langchain-ai/langchain/issues/11865 | 1,945,420,637 | 11,865 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
When I use the chatglm code from the official documentation, I can't connect to the endpoint URL provided there.
The official website is: https://python.langchain.com/docs/integrations/llms/chatglm
The specific code is as follows：
from langchain.llms import ChatGLM
from langchain.prompts import PromptTemplate
from langchain.chains import LLMChain
template = """{question}"""
prompt = PromptTemplate(template=template, input_variables=["question"])
endpoint_url = "http://127.0.0.1:8000"
llm = ChatGLM(
endpoint_url=endpoint_url,
max_token=80000,
history=[["我将从美国到中国来旅游，出行前希望了解中国的城市", "欢迎问我任何问题。"]],
top_p=0.9,
model_kwargs={"sample_model_args": False},
)
llm_chain = LLMChain(prompt=prompt, llm=llm)
question = "北京和上海两座城市有什么不同？"
llm_chain.run(question)
The code is encountering the following error：
ValueError: Error raised by inference endpoint: HTTPConnectionPool(host='127.0.0.1', port=8000): Max retries exceeded with
url: / (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x00000163A6691F70>: Failed to establish a new connection: [WinError 10061] 由于目标计算机积极拒绝，无法连接。'))
I have tried many methods I found online, but they didn't have much effect. I'm not sure if it's necessary to open a specific server through cmd to connect to 127.0.0.1:8000.
### Suggestion:
_No response_ | Issue: When using ChatGLM, I can't connect to the endpoint_url. | https://api.github.com/repos/langchain-ai/langchain/issues/11859/comments | 5 | 2023-10-16T11:35:37Z | 2024-02-11T16:10:51Z | https://github.com/langchain-ai/langchain/issues/11859 | 1,945,007,015 | 11,859 |
[
"langchain-ai",
"langchain"
] | This is my current code and it does run however it answers completely differently sometimes and it also writes the answer 3 or 4 times. Please fix it for me and if possible, fix my templates to make my output consistent
from dotenv import load_dotenv
import csv
import PyPDF2
from PyPDF2 import PdfReader
from langchain.document_loaders import DirectoryLoader, PyPDFLoader, PyPDFDirectoryLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.callbacks import get_openai_callback
from langchain.chat_models import ChatOpenAI
from langchain.chains import ConversationalRetrievalChain
from langchain.memory import ConversationBufferMemory
from langchain.prompts import PromptTemplate
import time
from langchain.vectorstores import Qdrant
from langchain.vectorstores import Chroma
from langchain.vectorstores import deeplake
from langchain.chains.qa_with_sources import load_qa_with_sources_chain
from langchain.callbacks import StreamingStdOutCallbackHandler
import pandas as pd
from docx import Document
from nltk.tokenize import sent_tokenize, word_tokenize
from collections import Counter
from nltk.corpus import stopwords
import os
def print_letter_by_letter(text):
for char in text:
print(char, end='', flush=True)
time.sleep(0.02)
def main():
load_dotenv()
my_activeloop_org_id = "yazanrisheh"
my_activeloop_dataset_name = "langchain_course_customer_support"
dataset_path = f"hub://{my_activeloop_org_id}/{my_activeloop_dataset_name}"
# directory_path = input("Copy your directory path here or upload a file: ")
directory_path = "C:\\Users\\Asus\\Documents\\Vendolista"
# pdf_loader = DirectoryLoader(directory_path,
# glob="**/*.pdf",
# show_progress=True,
# use_multithreading=True,
# silent_errors=True,
# loader_cls = PyPDFLoader)
pdf_loader = PyPDFDirectoryLoader(directory_path)
documents = pdf_loader.load()
print(str(len(documents))+ " documents loaded")
llm = ChatOpenAI(temperature = 0, model_name='gpt-3.5-turbo', callbacks=[StreamingStdOutCallbackHandler()], streaming = True)
# Split into chunks
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=800,
chunk_overlap=100,
)
chunks = text_splitter.split_documents(documents)
embeddings = OpenAIEmbeddings()
persist_directory = "C:\\Users\\Asus\\OneDrive\\Documents\\Vendolista"
knowledge_base = Chroma.from_documents(chunks, embeddings, persist_directory = persist_directory)
# save to disk
knowledge_base.persist()
#To delete the DB we created at first so that we can be sure that we will load from disk as fresh db
knowledge_base = None
new_knowledge_base = Chroma(persist_directory = persist_directory, embedding_function = embeddings)
# weird_knowledge_base = deeplake(chunks, dataset_path=dataset_path, embedding=embeddings)
# knowledge_base = Qdrant(documents, embeddings)
p_template = """Given the following conversation and a follow up question, rephrase the follow up question to be a standalone question, in its original language.
If the Standalone question is empty or cannot be generated, use the follow up question as Standalone question.
Chat History:
{chat_history}
Follow Up Input: {question}
Standalone question:"""
#CONDENSE_QUESTION_PROMPT = PromptTemplate.from_template(_template)
CONDENSE_QUESTION_PROMPT = PromptTemplate(input_variables=["chat_history","question"],
template=p_template)
memory = ConversationBufferMemory(memory_key="chat_history",input_key="question",output_key='answer',return_messages=True)
chatTemplate = """
Important: You are an intelligent chatbot designed to help agents by answering questions only on Enterprise services & activities.
Answer the question only if there is information in the chat history(delimited by ) and context(delimited by ) below.
If context is not empty and answer cannot be determined from context, say "I cannot detemine the answer from context".
If you don't know the answer, just say that you don't know, don't try to make up an answer.
Do not print your answer starting with "Answer:"
{context} ----------- {chat_history} ----------- Question: {question} Answer: Answer the question only if there is information based on the chat history(delimited by ) and context(delimited by ) below. 1. If context is not empty and answer cannot be determined from context, say "I cannot detemine the answer from context". 2. If you don't know the answer, just say that you don't know, don't try to make up an answer. 3. Do not print your answer starting with "Answer:"
"""
promptHist = PromptTemplate(
input_variables=["context", "question", "chat_history"],
template=chatTemplate
)
qa = ConversationalRetrievalChain.from_llm(
llm = llm,
retriever = new_knowledge_base.as_retriever(),
condense_question_prompt=CONDENSE_QUESTION_PROMPT,
verbose=True,
memory=memory,
condense_question_llm=llm,
return_generated_question=False,
combine_docs_chain_kwargs={"prompt": promptHist},
return_source_documents=False,
)
while True:
question = input("Ask me anything about the files (type 'exit' to quit): ")
if question.lower() in ["exit"] and len(question) == 4:
end_chat = "Thank you for visiting us! Have a nice day"
print_letter_by_letter(end_chat)
break
if question:
# chat_history = []
with get_openai_callback() as cb:
response = qa({"question": question}, return_only_outputs = False)
# chat_history.append(('user', question))
# chat_history.append(('AI', response))
print("AI:", response)
print(cb)
if __name__ == '__main__':
main()
| Not answering correctly according to prompt and writes same answer 3-4 times | https://api.github.com/repos/langchain-ai/langchain/issues/11857/comments | 7 | 2023-10-16T10:23:37Z | 2023-10-16T11:47:30Z | https://github.com/langchain-ai/langchain/issues/11857 | 1,944,879,698 | 11,857 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
hi @dosu-bot , @dosu-beta :
I am not able to figure out how we can pass variale to the input_variable in prompt :
prompt_with_loader = PromptTemplate(
input_variables=["query", "username", "password"],
template=
"""
Based on the instruction you are tasked to construct a Python script that performs the following action: {query}. Utilize the specifics in the given API documentation to understand the required endpoint, method, and parameters.
Credentials for authentication:
Username: {username}
Password: {password}
Please create a Python script using the 'requests' library to execute the desired action. Ensure the script includes:
- Parsing of the necessary details from the API specification
- Basic Authentication with the provided username and password (encoded in Base 64)
- Proper use of the relevant HTTP method, headers, body, and parameters as per the API specification
- print the response of the api
- Error handling for the request
"""
)
chain_type_kwargs = {"prompt": prompt_with_loader} ,
**qa_chain = RetrievalQAWithSourcesChain.from_chain_type(
llm=llm,
chain_type="stuff",
retriever=retriever,
chain_type_kwargs=chain_type_kwargs,
return_source_documents=False,
verbose=True
)**. // when i am goining to hit this block it gives me this ERRR : document_variable_name summaries was not found in llm_chain input_variables: ['query', 'username', 'password'] (type=value_error)
pls can you guide us here
### Suggestion:
_No response_ | Issue: Not clear through docs: how we can pass variale to the input_variable in prompt | https://api.github.com/repos/langchain-ai/langchain/issues/11856/comments | 5 | 2023-10-16T10:11:06Z | 2024-02-11T16:10:57Z | https://github.com/langchain-ai/langchain/issues/11856 | 1,944,857,017 | 11,856 |
[
"langchain-ai",
"langchain"
] | Can someone help me fix this code please.
My error: Traceback (most recent call last):
File "C:\Users\Asus\Documents\Vendolista\app2.py", line 130, in <module>
main()
File "C:\Users\Asus\Documents\Vendolista\app2.py", line 99, in main
qa = ConversationalRetrievalChain.from_llm(
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\chains\conversational_retrieval\base.py", line 356, in from_llm
return cls(
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\load\serializable.py", line 75, in __init__
super().__init__(**kwargs)
File "pydantic\main.py", line 341, in pydantic.main.BaseModel.__init__
pydantic.error_wrappers.ValidationError: 1 validation error for ConversationalRetrievalChain
retriever
value is not a valid dict (type=type_error.dict)
PS C:\Users\Asus\Documents\Vendolista> python app2.py
100%|████████████████████████████████████████████████████████████████████████████████████████████████████████| 33/33 [00:02<00:00, 11.17it/s]
87 documents loaded
Traceback (most recent call last):
File "C:\Users\Asus\Documents\Vendolista\app2.py", line 130, in <module>
main()
File "C:\Users\Asus\Documents\Vendolista\app2.py", line 99, in main
qa = ConversationalRetrievalChain.from_llm(
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\chains\conversational_retrieval\base.py", line 356, in from_llm
return cls(
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\load\serializable.py", line 75, in __init__
super().__init__(**kwargs)
File "pydantic\main.py", line 341, in pydantic.main.BaseModel.__init__
pydantic.error_wrappers.ValidationError: 1 validation error for ConversationalRetrievalChain
retriever
value is not a valid dict (type=type_error.dict)
My code:
from dotenv import load_dotenv
import csv
import PyPDF2
from PyPDF2 import PdfReader
from langchain.document_loaders import DirectoryLoader, PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.callbacks import get_openai_callback
from langchain.chat_models import ChatOpenAI
from langchain.chains import ConversationalRetrievalChain
from langchain.memory import ConversationBufferMemory
from langchain.prompts import PromptTemplate
import time
from langchain.vectorstores import Qdrant
from langchain.vectorstores import Chroma
from langchain.vectorstores import deeplake
from langchain.chains.qa_with_sources import load_qa_with_sources_chain
from langchain.callbacks import StreamingStdOutCallbackHandler
import pandas as pd
from docx import Document
from nltk.tokenize import sent_tokenize, word_tokenize
from collections import Counter
from nltk.corpus import stopwords
import os
def print_letter_by_letter(text):
for char in text:
print(char, end='', flush=True)
time.sleep(0.02)
def main():
load_dotenv()
my_activeloop_org_id = "yazanrisheh"
my_activeloop_dataset_name = "langchain_course_customer_support"
dataset_path = f"hub://{my_activeloop_org_id}/{my_activeloop_dataset_name}"
# directory_path = input("Copy your directory path here or upload a file: ")
directory_path = "C:\\Users\\Asus\\Documents\\Vendolista"
pdf_loader = DirectoryLoader(directory_path,
glob="**/*.pdf",
show_progress=True,
use_multithreading=True,
silent_errors=True,
loader_cls = PyPDFLoader)
documents = pdf_loader.load()
print(str(len(documents))+ " documents loaded")
llm = ChatOpenAI(temperature = 0, model_name='gpt-3.5-turbo', callbacks=[StreamingStdOutCallbackHandler()], streaming = True)
# Split into chunks
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=800,
chunk_overlap=100,
)
chunks = text_splitter.split_documents(documents)
embeddings = OpenAIEmbeddings()
persist_directory = "C:\\Users\\Asus\\OneDrive\\Documents\\Vendolista"
knowledge_base = Chroma.from_documents(chunks, embeddings, persist_directory = persist_directory)
# save to disk
knowledge_base.persist()
#To delete the DB we created at first so that we can be sure that we will load from disk as fresh db
knowledge_base = None
new_knowledge_base = Chroma(persist_directory = persist_directory, embedding_function = embeddings)
# weird_knowledge_base = deeplake(chunks, dataset_path=dataset_path, embedding=embeddings)
# knowledge_base = Qdrant(documents, embeddings)
p_template = """Given the following conversation and a follow up question, rephrase the follow up question to be a standalone question, in its original language.
If the Standalone question is empty or cannot be generated, use the follow up question as Standalone question.
Chat History:
{chat_history}
Follow Up Input: {question}
Standalone question:"""
#CONDENSE_QUESTION_PROMPT = PromptTemplate.from_template(_template)
CONDENSE_QUESTION_PROMPT = PromptTemplate(input_variables=["chat_history","question"],
template=p_template)
memory = ConversationBufferMemory(memory_key="chat_history",input_key="question",output_key='answer',return_messages=True)
chatTemplate = """
Important: You are an intelligent chatbot designed to help agents by answering questions only on Enterprise services & activities.
Answer the question only if there is information in the chat history(delimited by ) and context(delimited by ) below.
If context is not empty and answer cannot be determined from context, say "I cannot detemine the answer from context".
If you don't know the answer, just say that you don't know, don't try to make up an answer.
Do not print your answer starting with "Answer:"
{context} ----------- {chat_history} ----------- Question: {question} Answer: Answer the question only if there is information based on the chat history(delimited by ) and context(delimited by ) below. 1. If context is not empty and answer cannot be determined from context, say "I cannot detemine the answer from context". 2. If you don't know the answer, just say that you don't know, don't try to make up an answer. 3. Do not print your answer starting with "Answer:"
"""
promptHist = PromptTemplate(
input_variables=["context", "question", "chat_history"],
template=chatTemplate
)
qa = ConversationalRetrievalChain.from_llm(
llm = llm,
retriever = new_knowledge_base,
condense_question_prompt=CONDENSE_QUESTION_PROMPT,
verbose=False,
memory=memory,
condense_question_llm=llm,
return_generated_question=True,
combine_docs_chain_kwargs={"prompt": promptHist},
return_source_documents=True,
)
while True:
question = input("Ask me anything about the files (type 'exit' to quit): ")
if question.lower() in ["exit"] and len(question) == 4:
end_chat = "Thank you for visiting us! Have a nice day"
print_letter_by_letter(end_chat)
break
if question:
# chat_history = []
# with get_openai_callback() as cb:
response = qa({"question": question}, return_only_outputs = True)
# chat_history.append(('user', question))
# chat_history.append(('AI', response))
print("AI:", response)
if __name__ == '__main__':
main()
| ConversationalRetrievalChain error | https://api.github.com/repos/langchain-ai/langchain/issues/11855/comments | 3 | 2023-10-16T09:34:13Z | 2024-02-09T16:15:18Z | https://github.com/langchain-ai/langchain/issues/11855 | 1,944,786,305 | 11,855 |
[
"langchain-ai",
"langchain"
] | ### System Info
langchain : 0.0.306
python : 3.10.12
platform : Ubuntu
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [X] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
`loader = S3DirectoryLoader(
bucket="", prefix="data/",
aws_access_key_id="",
aws_secret_access_key="")
data = loader.load()`
It should have load all the files with in the bucket from a given prefix which may contain multiple sub-folders but it does not
### Expected behavior
It should have loaded all the files as list of Document of langchain schema , but it does not | S3 Directory Loader not working as expected | https://api.github.com/repos/langchain-ai/langchain/issues/11854/comments | 4 | 2023-10-16T08:37:20Z | 2024-02-10T16:12:37Z | https://github.com/langchain-ai/langchain/issues/11854 | 1,944,679,764 | 11,854 |
[
"langchain-ai",
"langchain"
] | ### System Info
LangChain version: v0.0.314
### Who can help?
@eyurtsev @baskaryan
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [X] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```python
from langchain.document_loaders import ConfluenceLoader
loader = ConfluenceLoader(url="xxx", session=get_session())
docs = loader.load(page_ids=["xxx"], keep_newlines=True)
for doc in docs:
print(doc.page_content)
have_newlines = "\n" in doc.page_content
print(f"{have_newlines=}")
```
read the source code, I found the `keep_newlines` arguments are not passed to `process_page` function when load with `cql` and `page_ids`
file: `langchain.document_loaders.confluence.py`
function: `Confluence.load`
lines: 308-353
<img width="781" alt="image" src="https://github.com/langchain-ai/langchain/assets/19882756/1d11d7dd-f883-4782-8b5d-3365ca2dff87">
<img width="821" alt="image" src="https://github.com/langchain-ai/langchain/assets/19882756/b87f3a16-65a4-490c-851f-0c51a534d171">
### Expected behavior
The `keep_newlines` parameter is passed correctly and empty lines are correctly retained | ConfluenceLoader's keep_newlines is not passed correctly | https://api.github.com/repos/langchain-ai/langchain/issues/11853/comments | 2 | 2023-10-16T08:34:03Z | 2024-02-09T16:15:28Z | https://github.com/langchain-ai/langchain/issues/11853 | 1,944,674,010 | 11,853 |
[
"langchain-ai",
"langchain"
] | ### System Info
python = "=3.10.12"
llama-cpp-python = "=0.2.11"
langchain = "=0.0.313"
### Who can help?
@hwchase17
@agola11
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [X] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
Run the [tagging docs]() with either `Ollama` or `llama-cpp-python` shown below.
```python
from langchain.chains import create_tagging_chain
from langchain.callbacks.manager import CallbackManager
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
from langchain.llms import LlamaCpp
from langchain.callbacks.manager import CallbackManager
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
model_using = [YOUR MODEL PATH] / "llama-2-7b-chat.Q4_0.gguf"
callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
llm = LlamaCpp(
model_path=model_using,
n_gpu_layers=32,
n_batch=512,
f16_kv=True,
callback_manager=callback_manager,
verbose=True,
n_ctx=512,
seed=42,
)
# Schema
schema = {
"properties": {
"sentiment": {"type": "string"},
"aggressiveness": {"type": "integer"},
"language": {"type": "string"},
}
}
chain = create_tagging_chain(schema, llm)
inp = "Estoy increiblemente contento de haberte conocido! Creo que seremos muy buenos amigos!"
print(chain.run(inp))
```
Gives an error that OpenAI functions are not available.
```
Traceback (most recent call last):
File "/Users/jak121/Git_repos/jk/20231016_langchain_tagging_issue.py", line 34, in <module>
print(chain.run(inp))
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/chains/base.py", line 501, in run
return self(args[0], callbacks=callbacks, tags=tags, metadata=metadata)[
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/chains/base.py", line 306, in __call__
raise e
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/chains/base.py", line 300, in __call__
self._call(inputs, run_manager=run_manager)
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/chains/llm.py", line 93, in _call
response = self.generate([inputs], run_manager=run_manager)
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/chains/llm.py", line 103, in generate
return self.llm.generate_prompt(
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/llms/base.py", line 498, in generate_prompt
return self.generate(prompt_strings, stop=stop, callbacks=callbacks, **kwargs)
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/llms/base.py", line 647, in generate
output = self._generate_helper(
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/llms/base.py", line 535, in _generate_helper
raise e
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/llms/base.py", line 522, in _generate_helper
self._generate(
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/llms/base.py", line 1044, in _generate
self._call(prompt, stop=stop, run_manager=run_manager, **kwargs)
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/llms/llamacpp.py", line 291, in _call
for chunk in self._stream(
File "/Users/jak121/Git_repos/llm/env/lib/python3.10/site-packages/langchain/llms/llamacpp.py", line 343, in _stream
result = self.client(prompt=prompt, stream=True, **params)
TypeError: Llama.__call__() got an unexpected keyword argument 'functions'
ggml_metal_free: deallocating
```
### Expected behavior
As per the docs: ` {'sentiment': 'positive', 'language': 'Spanish'}` | Tagging documentation does not work with LlamaCpp or Ollama | https://api.github.com/repos/langchain-ai/langchain/issues/11847/comments | 3 | 2023-10-16T07:27:07Z | 2024-02-11T16:11:02Z | https://github.com/langchain-ai/langchain/issues/11847 | 1,944,554,193 | 11,847 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I'm encountering an error while trying to integrate confluenece attachment by the keyword include_attachments=True
Created a chunk of size 494, which is longer than the specified 100
Created a chunk of size 230, which is longer than the specified 100
Created a chunk of size 121, which is longer than the specified 100
Created a chunk of size 224, which is longer than the specified 100
Created a chunk of size 103, which is longer than the specified 100
Created a chunk of size 638, which is longer than the specified 100
Created a chunk of size 112, which is longer than the specified 100
Created a chunk of size 228, which is longer than the specified 100
Created a chunk of size 147, which is longer than the specified 100
Created a chunk of size 127, which is longer than the specified 100
Created a chunk of size 306, which is longer than the specified 100
Created a chunk of size 105, which is longer than the specified 100
Below is my code:
text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=0)
texts = text_splitter.split_documents(documents)
text_splitter = TokenTextSplitter(chunk_size=1000, chunk_overlap=10, encoding_name="cl100k_base")
texts = text_splitter.split_documents(texts)
### Suggestion:
_No response_ | Issue: Getting Error while integrating confluence attachment | https://api.github.com/repos/langchain-ai/langchain/issues/11845/comments | 2 | 2023-10-16T06:34:51Z | 2024-02-06T16:20:21Z | https://github.com/langchain-ai/langchain/issues/11845 | 1,944,475,842 | 11,845 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I'm encountering an error while trying to integrate confluenece attachment by the keyword include_attachments=True
Created a chunk of size 494, which is longer than the specified 100
Created a chunk of size 230, which is longer than the specified 100
Created a chunk of size 121, which is longer than the specified 100
Created a chunk of size 224, which is longer than the specified 100
Created a chunk of size 103, which is longer than the specified 100
Created a chunk of size 638, which is longer than the specified 100
Created a chunk of size 112, which is longer than the specified 100
Created a chunk of size 228, which is longer than the specified 100
Created a chunk of size 147, which is longer than the specified 100
Created a chunk of size 127, which is longer than the specified 100
Created a chunk of size 306, which is longer than the specified 100
Created a chunk of size 105, which is longer than the specified 100
Below is my code:
text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=0)
texts = text_splitter.split_documents(documents)
text_splitter = TokenTextSplitter(chunk_size=1000, chunk_overlap=10, encoding_name="cl100k_base")
texts = text_splitter.split_documents(texts)
### Suggestion:
_No response_ | Issue: Getting Error while integrating confluence attachment | https://api.github.com/repos/langchain-ai/langchain/issues/11844/comments | 1 | 2023-10-16T06:00:12Z | 2023-10-16T06:35:03Z | https://github.com/langchain-ai/langchain/issues/11844 | 1,944,428,691 | 11,844 |
[
"langchain-ai",
"langchain"
] | ### Issue with current documentation:
from langchain.embeddings import OpenAIEmbeddings
from langchain.vectorstores import Chroma
from langchain.chat_models import ChatOpenAI
from langchain.prompts import (
FewShotChatMessagePromptTemplate,
SemanticSimilarityExampleSelector,
ChatPromptTemplate,
)
examples = [
{"input": "2+2", "output": "4"},
{"input": "2+3", "output": "5"},
{"input": "2+4", "output": "6"},
{"input": "What did the cow say to the moon?", "output": "nothing at all"},
{
"input": "Write me a poem about the moon",
"output": "One for the moon, and one for me, who are we to talk about the moon?",
},
]
example_selector = SemanticSimilarityExampleSelector.from_examples(
examples,
OpenAIEmbeddings(),
Chroma,
k=1
)
few_shot_prompt = FewShotChatMessagePromptTemplate(
example_selector=example_selector,
example_prompt=ChatPromptTemplate.from_messages(
[("human", "{input}"), ("ai", "{output}")]
),
)
final_prompt = ChatPromptTemplate.from_messages(
[
("system", "You are a wondrous wizard of math."),
few_shot_prompt,
("human", "{input}"),
]
)
# 由于example_selector已经选择了一个样本进行匹配，这将导致少样本示例失效，因为选择的样本可能与给定的输入不相关。
print(final_prompt.format(input="What did the cow say to the moon?"))
### Idea or request for content:
In the "Dynamic Few-Shot Prompting" documentation, it is mentioned that when example_selector selects a sample for matching, it can cause the few-shot examples to lose effectiveness because the selected sample may not be relevant to the given input.
https://python.langchain.com/docs/modules/model_io/prompts/prompt_templates/few_shot_examples_chat#dynamic-few-shot-prompting | In the "Dynamic Few-Shot Prompting" documentation, it is mentioned that when example_selector selects a sample for matching, it can cause the few-shot examples to lose effectiveness because the selected sample may not be relevant to the given input. | https://api.github.com/repos/langchain-ai/langchain/issues/11843/comments | 4 | 2023-10-16T05:43:53Z | 2024-05-07T16:06:03Z | https://github.com/langchain-ai/langchain/issues/11843 | 1,944,410,321 | 11,843 |
[
"langchain-ai",
"langchain"
] | ### System Info
I attempted to create an AI agent bot by utilizing the QianfanChatEndpoint. However, I encountered an issue when I attempted to define a tool for making calls.
Subsequent to consulting with a contributor from the QianfanChatEndpoint project, it was determined that it is imperative to merge the following pull request for resolution: [link to the GitHub pull request](https://github.com/langchain-ai/langchain/pull/11107). Once this fix is implemented, my agent will work fine with Function Call integration.
@hwchase17
@agola11
### Who can help?
@hwchase17
@agola11
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [X] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```
"""Example demo for agent.py
"""
import os
from typing import Any, Dict, List
from langchain.agents import tool
from langchain.agents.format_scratchpad import format_to_openai_functions
from langchain.agents.output_parsers import OpenAIFunctionsAgentOutputParser
from langchain.chat_models import QianfanChatEndpoint
from langchain.prompts import ChatPromptTemplate, MessagesPlaceholder
# from langchain.tools.render import format_tool_to_openai_function
prompt = ChatPromptTemplate.from_messages(
[
(
"system",
"You are very powerful assistant, but bad at get"
"today's temperature of location.",
),
("user", "{input}"),
MessagesPlaceholder(variable_name="agent_scratchpad"),
]
)
@tool
def get_current_weather(location: str) -> str:
"""Returns current temperature of location."""
return "25"
tools = [get_current_weather]
if __name__ == "__main__":
os.environ["QIANFAN_AK"] = "ak"
os.environ["QIANFAN_SK"] = "sk"
llm = QianfanChatEndpoint(model="ERNIE-Bot") # 仅EB支持
llm_with_tools = llm.bind(
# functions=functions
functions=[
{
"name": "get_current_weather",
"description": "获得指定地点的天气",
"parameters": {
"type": "object",
"properties": {
"location": {
"type": "string",
"description": "省，市名，例如：河北省，石家庄",
},
"unit": {"type": "string", "enum": ["摄氏度", "华氏度"]},
},
"required": ["location"],
},
}
]
)
p = {
"input": lambda x: x["input"],
"agent_scratchpad": lambda x: format_to_openai_functions(
x["intermediate_steps"]
),
}
agent: Dict = p | prompt | llm_with_tools | OpenAIFunctionsAgentOutputParser()
# main loop
from langchain.schema.agent import AgentFinish
intermediate_steps: List[Any] = []
while True:
output = agent.invoke(
{"input": "上海市今天天气如何？", "intermediate_steps": intermediate_steps}
)
if isinstance(output, AgentFinish):
final_result = output.return_values["output"]
break
else:
print(output.tool, output.tool_input)
tool = {
"get_current_weather": get_current_weather,
}[output.tool]
observation = tool.run(output.tool_input)
# observation = "{\"temperature\": \"25\", \"description\": \"晴朗\"}"
intermediate_steps.append((output, observation))
print("result", final_result)
```
### Expected behavior
上海市今天气温25度 | QianfanChatEndpoint function call not work for agent. | https://api.github.com/repos/langchain-ai/langchain/issues/11839/comments | 2 | 2023-10-16T02:08:55Z | 2024-02-09T16:15:38Z | https://github.com/langchain-ai/langchain/issues/11839 | 1,944,207,173 | 11,839 |
[
"langchain-ai",
"langchain"
] | ### System Info
Dependencies:
langchain == 0.0.285
pytest ~= 7.4.0
pytest-asyncio ~= 0.21.1
pytest-mock ~= 3.11.1
Using Python 3.11
### Who can help?
@agola11
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [X] Chains
- [ ] Callbacks/Tracing
- [X] Async
### Reproduction
The following Python script is self-contained. It features:
- A custom `Chain` implementation that uses an `LLMChain` internally, `bar`
- We don't care about its inputs or outputs
- We simply care that it takes an `LLMChain` and uses that to compute its output.
- A test (`async` using the `pytest-asyncio` library) where:
- I mocked the `LLMChain` using an `AsyncMock` with spec `LLMChain`
- I mocked the return value of the `arun` function in the `LLMChain`
### Self-contained Python Script
<details>
<summary>Self-contained Python Script</summary>
```py
from langchain.chains.base import Chain
from langchain import LLMChain
from typing import Any
from langchain.callbacks.manager import AsyncCallbackManagerForChainRun, CallbackManagerForChainRun
from unittest.mock import AsyncMock
from pydantic import BaseModel, Extra
class CustomChain(Chain):
bar: LLMChain
class Config:
"""Configuration for this pydantic object."""
extra = Extra.forbid
arbitrary_types_allowed = True
allow_population_by_field_name = True
@property
def input_schema(self) -> type[BaseModel]:
return BaseModel # we don't care about the inputs here
@property
def input_keys(self) -> list[str]:
return []
@property
def output_keys(self) -> list[str]:
return ["output"]
async def _acall(
self,
inputs: dict[str, Any],
run_manager: AsyncCallbackManagerForChainRun | None = None,
) -> dict[str, Any]:
intermediary_result = await self.bar.arun(message="hello")
return {"output": intermediary_result}
def _call(self, inputs: dict[str, Any], run_manager: CallbackManagerForChainRun | None = None) -> dict[str, Any]:
raise NotImplementedError("Only supports async")
async def test_custom_chain():
bar = AsyncMock(LLMChain)
bar.arun.return_value = "Hi there!"
custom_chain = CustomChain(bar=bar)
await custom_chain._acall(inputs={})
bar.arun.assert_awaited_once()
```
</details>
### Observed Output
The test should naturally succeed. Problem is that _somewhere_ in the initialization of the `CustomChain`, `bar` is changed from being an `AsyncMock` to a `MagicMock`. This is the inexplicable bit to me.
Placing a debugger breakpoint right before the initialization of `CustomChain` in the test function:
```
custom_chain = CustomChain(bar=bar)
```
It shows that `bar` is `<AsyncMock spec='LLMChain' id='5218712656'>`, however from inside `CustomChain._acall` function, `self.bar` is a `<MagicMock name='mock._copy_and_set_values()' id='5221083792'>`.
<details>
<summary>Full test output</summary>
```
async def test_custom_chain():
bar = AsyncMock(LLMChain)
bar.arun.return_value = "Hi there!"
custom_chain = CustomChain(bar=bar)
> await custom_chain._acall(inputs={})
test/brain/test_foo.py:49:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = CustomChain(memory=None, callbacks=None, callback_manager=None, verbose=False, tags=None, metadata=None, bar=<MagicMock name='mock._copy_and_set_values()' id='5099464528'>)
inputs = {}, run_manager = None
async def _acall(
self,
inputs: dict[str, Any],
run_manager: AsyncCallbackManagerForChainRun | None = None,
) -> dict[str, Any]:
> intermediary_result = await self.bar.arun(message="hello")
E TypeError: object MagicMock can't be used in 'await' expression
test/brain/test_foo.py:37: TypeError
```
</details>
### Expected behavior
Naturally, I would expect the parameter to my custom chain to be exactly what I passed, unmodified at any point. Hence, I would expect the test and its assertion to succeed.
_Please note_: I tried reproducing the issue outside of LangChain, just by creating a similar class - just inheriting from a Pydantic `BaseModel` rather than LangChain's `Chain`. The test succeeded just as expected. | AsyncMock not preserved when passed to custom Chain | https://api.github.com/repos/langchain-ai/langchain/issues/11838/comments | 4 | 2023-10-15T22:27:00Z | 2024-02-10T16:12:47Z | https://github.com/langchain-ai/langchain/issues/11838 | 1,944,082,144 | 11,838 |
[
"langchain-ai",
"langchain"
] | ### System Info
0.0.314
### Who can help?
@hwchase17 @agola11 @eyurtsev
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [X] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```
from langchain.chat_models import ChatOpenAI
from langchain.agents import load_tools, initialize_agent
from langchain.agents import AgentType
from langchain.tools import AIPluginTool
tool = AIPluginTool.from_plugin_url("https://scholar-ai.net/.well-known/ai-plugin.json")
llm = ChatOpenAI(temperature=0, streaming=True, model_name="gpt-3.5-turbo-16k-0613", max_tokens=16000)
tools = [tool]
agent_chain = initialize_agent(
tools, llm, agent=AgentType.OPENAI_FUNCTIONS, verbose=True
)
agent_chain.run("What are the antiviral effects of Sillymarin?")
```
And I get
```
InvalidRequestError: This model's maximum context length is 16385 tokens. However, you requested 17952 tokens (1855 in the messages, 97 in the functions, and 16000 in the completion). Please reduce the length of the messages, functions, or completion.
```
### Expected behavior
The idea is that should provide sensible information based on the Plugin Scholar AI.
It would be terrific, if you could help us here. 🙂 | Use OpenAI ChatGPT Plugin via Python API fails with maximum number of tokens error | https://api.github.com/repos/langchain-ai/langchain/issues/11837/comments | 3 | 2023-10-15T21:59:24Z | 2024-02-11T16:11:07Z | https://github.com/langchain-ai/langchain/issues/11837 | 1,944,073,141 | 11,837 |
[
"langchain-ai",
"langchain"
] | ### System Info
LangChain Version: 0.0.314
Python 3.9.6
I have been using LangChain with OpenAI and FAISS for building RAG chatbots. Everything was working fine until last week.
I have noticed that for some reason, I have much higher response times. I have debugged the LangChain code and found that the reason is the LLM response itself. I am using OpenAI GPT-3.5-turbo, I haven't changed anything in the prompt being sent, but response time has increased dramatically. For example, a very simple prompt in September took 2 seconds to generate. Now the exact same prompt takes up to 14 seconds to get a response!
Does anybody have the same experience? Does anybody know why this can be happening?
Any help will be appreciated!
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
1. Use OpenAI as an LLM
2. Observer response times
### Expected behavior
The LLM response should be much faster | OpenAI API is extremely slow | https://api.github.com/repos/langchain-ai/langchain/issues/11836/comments | 6 | 2023-10-15T21:39:04Z | 2024-04-10T16:14:04Z | https://github.com/langchain-ai/langchain/issues/11836 | 1,944,067,045 | 11,836 |
[
"langchain-ai",
"langchain"
] | https://github.com/langchain-ai/langchain/blob/a50630277295c3884be8e839b04718d0e99b4ea4/libs/langchain/langchain/chains/moderation.py#L93C43-L93C43
Code a line no. 93 needs to be updated as below
self.client.create(input = text, model = self.model_name) | model_name is not passed as a argument to create method | https://api.github.com/repos/langchain-ai/langchain/issues/11830/comments | 1 | 2023-10-15T16:52:02Z | 2024-02-06T16:20:41Z | https://github.com/langchain-ai/langchain/issues/11830 | 1,943,966,873 | 11,830 |
[
"langchain-ai",
"langchain"
] | ### Feature request
similarly to:
```
db = Chroma(persist_directory=persist_directory, embedding_function=embeddings, client_settings=CHROMA_SETTINGS)
collection = db.get()
```
where ```collection``` is a list of documents already in the vector store, is there a way to get that list from neo4j_vecror that was initialized from_documents() ?
### Motivation
This will allow to filter out existing documents when new ones are added in folder that is being used to build an expanding document vector store.
### Your contribution
i was thinkging using .query() as a workaround but not sure how a CYPHER query can return all possible documents currently in the store. | return all documents in neo4j_vector | https://api.github.com/repos/langchain-ai/langchain/issues/11829/comments | 2 | 2023-10-15T16:41:33Z | 2024-02-06T16:20:46Z | https://github.com/langchain-ai/langchain/issues/11829 | 1,943,962,792 | 11,829 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I have installed langchain and ctransformer using -
```
pip install langchain
pip install ctransformers[cuda]
```
I am trying following piece of code -
```
from langchain.llms import CTransformers
config = {'max_new_tokens': 512, 'repetition_penalty': 1.1, 'context_length': 8000, 'temperature':0, 'gpu_layers':50}
llm = CTransformers(model = "./codellama-7b.Q4_0.gguf", model_type = "llama", gpu_layers=50, config=config)
```
Here gpu_layers parameter is specified still gpu is not being used and complete load is on cpu.
Can someone please point out if there is any step missing.
### Suggestion:
_No response_ | Issue: GPU is not used with Ctransformers even after specifying gpu_layers parameter | https://api.github.com/repos/langchain-ai/langchain/issues/11826/comments | 15 | 2023-10-15T14:35:12Z | 2024-05-16T07:56:34Z | https://github.com/langchain-ai/langchain/issues/11826 | 1,943,913,102 | 11,826 |
[
"langchain-ai",
"langchain"
] | ### System Info
Python 3.11.5
langchain==0.0.260
atlassian-python-api==3.41.2
### Who can help?
@eyurtsev
@hwchase17
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [X] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
My code:
`data = loader.load(page_ids=[page_id], include_attachments=True, limit=50)`
Error message:
`requests.exceptions.HTTPError: 404 Client Error: Not Found for url: ...`
### Expected behavior
The problem is the URL path is not vaild (you can read my post at atlassian developer community [here ](https://community.developer.atlassian.com/t/download-confluence-content-attachment-via-rest-api-get-404-error/73782))
My solution:
at document_loaders/confluence.py [line 542](https://github.com/langchain-ai/langchain/blob/a50630277295c3884be8e839b04718d0e99b4ea4/libs/langchain/langchain/document_loaders/confluence.py#L542C6-L542C6)
change
`absolute_url = self.base_url + attachment["_links"]["download"]`
to
`absolute_url = self.base_url + "/wiki" + attachment["_links"]["download"]` | Confluence document loader can not get the attachments | https://api.github.com/repos/langchain-ai/langchain/issues/11824/comments | 1 | 2023-10-15T13:44:03Z | 2023-10-16T10:03:34Z | https://github.com/langchain-ai/langchain/issues/11824 | 1,943,895,150 | 11,824 |
[
"langchain-ai",
"langchain"
] | ### System Info
window 、langchain-0.0.314、python 3.10
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [x] My own modified scripts
### Related Components
- [x] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
run
```
from langchain.chains.summarize import load_summarize_chain
from langchain.chat_models import ChatOpenAI
from langchain import LLMChain
from langchain.prompts.chat import (
ChatPromptTemplate,
HumanMessagePromptTemplate,
)
api_base_url = "http://localhost:10860/v1"
api_key= "EMPTY"
LLM_MODEL = "/***/AquilaChat2-7B"
model = ChatOpenAI(
streaming=True,
verbose=True,
# callbacks=[callback],
openai_api_key=api_key,
openai_api_base=api_base_url,
model_name=LLM_MODEL
)
chain = load_summarize_chain(model,
chain_type="map_reduce",
verbose = True,
)
output_summary = chain.run(docs[:5])
print(output_summary)
``
### Expected behavior
nomal
https://blog.csdn.net/FrenzyTechAI/article/details/131524746 | ValidationError: 1 validation error for ChatMessage | https://api.github.com/repos/langchain-ai/langchain/issues/11823/comments | 3 | 2023-10-15T12:14:33Z | 2024-02-09T16:15:53Z | https://github.com/langchain-ai/langchain/issues/11823 | 1,943,865,066 | 11,823 |
[
"langchain-ai",
"langchain"
] | ### System Info
Langchain version: 0.0.313
Python version: Python 3
OS: Mac
### Who can help?
@hwchase17
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [X] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [X] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
from langchain.chains.router.multi_prompt import MultiPromptChain
Assume have already installed langchain and other necessary libraries
Error I get:
cannot import name 'BasePromptTemplate' from 'langchain.schema'. I checked in the langchain.schema file. BasePromptTemplate is indeed not there. But then I don't know how to get that file. I see PromptTemplate file in the langchain.schema file
### Expected behavior
To import MultiPromptChain | Getting an error while importing MultiPromptChain | https://api.github.com/repos/langchain-ai/langchain/issues/11819/comments | 6 | 2023-10-15T06:32:23Z | 2024-02-28T16:08:50Z | https://github.com/langchain-ai/langchain/issues/11819 | 1,943,745,728 | 11,819 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I'm encountering an error while trying to integrate confluenece attachment using the keyword "include_attachments=True", without these keyword I am able to do question answering on Text content.
But I want to fetch attachment of the confluence space. Below is my code:
I encounter the following error:
### Suggestion:
_No response_ | Issue: Getting Error while integrating confluence attachment | https://api.github.com/repos/langchain-ai/langchain/issues/11818/comments | 2 | 2023-10-15T04:59:22Z | 2024-02-06T16:20:56Z | https://github.com/langchain-ai/langchain/issues/11818 | 1,943,712,366 | 11,818 |
[
"langchain-ai",
"langchain"
] | ### System Info
langchain=0.0.294
python=3.10
### Who can help?
@hwchase17 @mmz-001 @baskaryan
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [X] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [X] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
I have tried multiple text splitters, but none of them seem to respect the chunk_size parameter. My document is a podcast transcript that I create from a simple text file that combines all transcript sentences into a single string.
The SpacyTextSplitter then formats the document into coherent sentences and paragraphs. But I end up with over five individual small documents, even when I test a small part of the transcript of only 13,000 characters.
The RecursiveCharacterTextSplitter and CharacterTextSplitter split the documents into simpler grouped sentences, but they also ignore the chunk_size parameter.
My test file
[3kiv08TDn4vywkNAdoT2In_short_v2.txt](https://github.com/langchain-ai/langchain/files/12908607/3kiv08TDn4vywkNAdoT2In_short_v2.txt)
```
import spacy
from langchain.text_splitter import RecursiveCharacterTextSplitter, CharacterTextSplitter, SpacyTextSplitter
@app.post("/reduce_file/", response_model=ItemResponse)
async def reduce_file(item: Item_Reduce_FileURL):
logging.info("Entering function reduce_file")
try:
# Dynamic file type detection (from summarize_file)
response = requests.get(item.file_link)
if response.status_code != 200:
raise HTTPException(status_code=400, detail="Failed to download file")
content_type = response.headers.get('Content-Type')
file_extension = mimetypes.guess_extension(content_type)
if file_extension and file_extension.startswith('.'):
file_extension = file_extension[1:]
temp_file_name = f"temp_file.{file_extension}"
with open(temp_file_name, "wb") as f:
f.write(response.content)
# Load the document
file_loader = OnlinePDFLoader(file_path=temp_file_name) if file_extension == 'pdf' else UnstructuredAPIFileLoader(
file_path=temp_file_name,
infer_table_structure=True,
strategy="fast",
model_name="chipper",
mode="elements",
pdf_infer_table_structure=False,
ocr_languages=["eng"],
discard_invalid_pages=True,
encoding="utf_8",
api_key=unstructured_api_key
)
try:
doc = file_loader.load()
except Exception as load_exception:
logging.error(f"Error in file_loader.load(): {load_exception}")
return CustomError.handle_exception(load_exception, item, exception_type=str(type(load_exception)))
# Initialize text splitter and split the document (from reduce_pdf)
text_splitter = SpacyTextSplitter(
chunk_size=5000,
chunk_overlap=0,
length_function=len,
)
split_docs = text_splitter.split_documents(doc)
### Expected behavior
I expect only two documents since the chunk_size parameter is set to 10000 characters. I have tested with increasing and decreasing the chunk_size, and I continue to get a consistent amount of smaller documents of 1200 to 1500 characters. | Mutiple text_splitter ignore the chunk_size parameter | https://api.github.com/repos/langchain-ai/langchain/issues/11817/comments | 2 | 2023-10-15T04:42:23Z | 2023-10-16T07:10:32Z | https://github.com/langchain-ai/langchain/issues/11817 | 1,943,707,750 | 11,817 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I'm encountering an error while trying to integrate confluenece attachment using the keyword "include_attachments=True", without these keyword I am able to do question answering on Text content.
But I want to fetch attachment of the confluence space. Below is my code:
I encounter the following error:
### Suggestion:
_No response_ | Issue: Getting Error while integrating confluence attachment | https://api.github.com/repos/langchain-ai/langchain/issues/11816/comments | 3 | 2023-10-15T03:49:23Z | 2024-02-07T16:18:48Z | https://github.com/langchain-ai/langchain/issues/11816 | 1,943,688,748 | 11,816 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
Hi, I have lots product factsheet in PDF format and would like to perform QnA on these documents. These documents have a common layout. I parsed these documents and stored them in ChromaDB. However, when I try to query the vector store for a specific details for ex. "What is the product category of product A?", it retrieved irrelevant chunks. What approach should be taken for such homogeneous documents QnA use cases?
### Suggestion:
_No response_ | Issue: Similarity Search on Chroma does not retrieve relevant chunk for homogeneous document search | https://api.github.com/repos/langchain-ai/langchain/issues/11815/comments | 2 | 2023-10-15T03:35:42Z | 2024-02-06T16:21:07Z | https://github.com/langchain-ai/langchain/issues/11815 | 1,943,684,633 | 11,815 |
[
"langchain-ai",
"langchain"
] | ### Feature request
A retriever for documents from [Outline](https://github.com/outline/outline).
The API has a search endpoint which allows this to be possible: https://www.getoutline.com/developers#tag/Documents/paths/~1documents.search/post
The implementation will be similar to the Wikipedia retriever:
https://python.langchain.com/docs/integrations/retrievers/wikipedia
### Motivation
Outline is an open source project that let's you create a knowledge base, like a wiki. Creating a retriever for Outline will let your team interact with your knowledge base using an LLM.
### Your contribution
PR will be coming soon. | Create retriever for Outline to ask questions on knowledge base | https://api.github.com/repos/langchain-ai/langchain/issues/11814/comments | 1 | 2023-10-15T01:58:24Z | 2023-11-28T20:58:07Z | https://github.com/langchain-ai/langchain/issues/11814 | 1,943,639,624 | 11,814 |
[
"langchain-ai",
"langchain"
] | ### System Info
Langchain 0.0.314
When LLM returns with action input parameter without space in-between two tables, call fails. Certain LLMs dont follow "Table1, Table2" format and returns with "Table1,Table2"
This is easily addressed by stripping out all spaces and split with ","
<code>
class InfoSQLDatabaseTool(BaseSQLDatabaseTool, BaseTool):
"""Tool for getting metadata about a SQL database."""
name: str = "sql_db_schema"
description: str = """
Input to this tool is a comma-separated list of tables, output is the schema and sample rows for those tables.
Example Input: "table1, table2, table3"
"""
def _run(
self,
table_names: str,
run_manager: Optional[CallbackManagerForToolRun] = None,
) -> str:
"""Get the schema for tables in a comma-separated list."""
return self.db.get_table_info_no_throw(table_names.split(", "))
</code>
Replace return line with the following:
<code>
import re
.
.
.
def_run(
.
.
.
return self.db.get_table_info_no_throw(re.sub(" *","",table_names).split(","))
</code>
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [X] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
This can be reproduced with Claude V2
### Expected behavior
sql_db_schema to not fail | SQL DB Tools sql_db_schema bad input | https://api.github.com/repos/langchain-ai/langchain/issues/11813/comments | 1 | 2023-10-14T23:56:03Z | 2024-02-06T16:21:11Z | https://github.com/langchain-ai/langchain/issues/11813 | 1,943,609,736 | 11,813 |
[
"langchain-ai",
"langchain"
] | ### Feature request
We propose the addition of some of the querying endpoints supported by serpapi that are not suported on langchain at the moment. Some examples, include querying google job, google trends and google finance etc.###
### Motivation
The motivation for this feature proposal is to enhance the querying capabilities of Langchain by incorporating support for various querying endpoints provided by serpapi. The proposed feature aims to improve Langchain's querying capabilities by enabling users to interact with serpapi supported endpoints seamlessly.
### Your contribution
Me and 3 others intend to submit a pull request for this issue at some point in November for a school project. Looking forward to contributing. | Add support for many of the querying endpoints with serpapi | https://api.github.com/repos/langchain-ai/langchain/issues/11811/comments | 2 | 2023-10-14T19:16:42Z | 2024-01-30T16:11:49Z | https://github.com/langchain-ai/langchain/issues/11811 | 1,943,495,649 | 11,811 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
Looking to use `AsyncChromiumLoader` to scrape URLs but I keep running into the following error:
```python
loader = AsyncChromiumLoader(["https://www.espn.com/nfl/story/_/id/38652877/lions-rb-jahmyr-gibbs-wr-amon-ra-st-brown-return-vs-bucs"])
loader.load()
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
Cell In[17], line 2
1 loader = AsyncChromiumLoader(["https://www.espn.com/nfl/story/_/id/38652877/lions-rb-jahmyr-gibbs-wr-amon-ra-st-brown-return-vs-bucs"])
----> 2 loader.load()
File /opt/conda/lib/python3.10/site-packages/langchain/document_loaders/chromium.py:90, in AsyncChromiumLoader.load(self)
81 def load(self) -> List[Document]:
82 """
83 Load and return all Documents from the provided URLs.
84
(...)
88
89 """
---> 90 return list(self.lazy_load())
File /opt/conda/lib/python3.10/site-packages/langchain/document_loaders/chromium.py:77, in AsyncChromiumLoader.lazy_load(self)
66 """
67 Lazily load text content from the provided URLs.
68
(...)
74
75 """
76 for url in self.urls:
---> 77 html_content = asyncio.run(self.ascrape_playwright(url))
78 metadata = {"source": url}
79 yield Document(page_content=html_content, metadata=metadata)
File /opt/conda/lib/python3.10/asyncio/runners.py:33, in run(main, debug)
9 """Execute the coroutine and return the result.
10
11 This function runs the passed coroutine, taking care of
(...)
30 asyncio.run(main())
31 """
32 if events._get_running_loop() is not None:
---> 33 raise RuntimeError(
34 "asyncio.run() cannot be called from a running event loop")
36 if not coroutines.iscoroutine(main):
37 raise ValueError("a coroutine was expected, got {!r}".format(main))
RuntimeError: asyncio.run() cannot be called from a running event loop
```
I am running langchain 0.0.314. I looked through [the documentation](https://python.langchain.com/docs/integrations/document_loaders/async_html) which wasn't helpful. I found [this stackoverflow thread](https://stackoverflow.com/questions/55409641/asyncio-run-cannot-be-called-from-a-running-event-loop-when-using-jupyter-no) which suggests adding an `await` but that seems useful if we are directly using the asychio class.
It's worth noting that I run into a similar error when using the `PlaywrightURLLoader` and I run into the following error:
```python
loader = PlaywrightURLLoader(["https://www.espn.com/nfl/story/_/id/38652877/lions-rb-jahmyr-gibbs-wr-amon-ra-st-brown-return-vs-bucs"])
loader.load()
---------------------------------------------------------------------------
Error Traceback (most recent call last)
Cell In[186], line 2
1 loader = PlaywrightURLLoader(["https://www.espn.com/nfl/story/_/id/38652877/lions-rb-jahmyr-gibbs-wr-amon-ra-st-brown-return-vs-bucs"])
----> 2 loader.load()
File /opt/conda/lib/python3.10/site-packages/langchain/document_loaders/url_playwright.py:154, in PlaywrightURLLoader.load(self)
150 from playwright.sync_api import sync_playwright
152 docs: List[Document] = list()
--> 154 with sync_playwright() as p:
155 browser = p.chromium.launch(headless=self.headless)
156 for url in self.urls:
File /opt/conda/lib/python3.10/site-packages/playwright/sync_api/_context_manager.py:47, in PlaywrightContextManager.__enter__(self)
45 self._own_loop = True
46 if self._loop.is_running():
---> 47 raise Error(
48 """It looks like you are using Playwright Sync API inside the asyncio loop.
49 Please use the Async API instead."""
50 )
52 # Create a new fiber for the protocol dispatcher. It will be pumping events
53 # until the end of times. We will pass control to that fiber every time we
54 # block while waiting for a response.
55 def greenlet_main() -> None:
Error: It looks like you are using Playwright Sync API inside the asyncio loop.
Please use the Async API instead.
```
### Suggestion:
_No response_ | Issue: `AsyncChromiumLoader` RuntimeError: asyncio.run() cannot be called from a running event loop | https://api.github.com/repos/langchain-ai/langchain/issues/11810/comments | 2 | 2023-10-14T16:12:24Z | 2024-02-08T16:18:46Z | https://github.com/langchain-ai/langchain/issues/11810 | 1,943,375,169 | 11,810 |
[
"langchain-ai",
"langchain"
] | ### System Info
LangChain: langchain-0.0.314
Python: Anaconda Python 3.9.18
X86
RTX3080 Laptop (16G)
CUDA 11.8
cuDNN 8.9.5
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [X] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
1.git clone https://github.com/ymcui/Chinese-LLaMA-Alpaca-2.git
2.cd Chinese-LLaMA-Alpaca-2/scripts/langchain
3.python langchain_sum.py --model_path chinese-alpaca-2-7b-hf --file_path doc.txt --chain_type refine
(langchain) zhanghui@zhanghui-OMEN-by-HP-Laptop-17-ck0xxx:~/Chinese-LLaMA-Alpaca-2/scripts/langchain$ python langchain_sum.py --model_path chinese-alpaca-2-7b-hf --file_path doc.txt --chain_type refine
/home/zhanghui/anaconda3/envs/langchain/lib/python3.9/site-packages/langchain/__init__.py:39: UserWarning: Importing HuggingFacePipeline from langchain root module is no longer supported.
warnings.warn(
loading LLM...
Loading checkpoint shards: 100%|███████████████████████████████████████████████████████████████████████████████| 2/2 [00:03<00:00, 1.86s/it]
/home/zhanghui/anaconda3/envs/langchain/lib/python3.9/site-packages/transformers/generation/configuration_utils.py:362: UserWarning: `do_sample` is set to `False`. However, `temperature` is set to `0.9` -- this flag is only used in sample-based generation modes. You should set `do_sample=True` or unset `temperature`. This was detected when initializing the generation config instance, which means the corresponding file may hold incorrect parameterization and should be fixed.
warnings.warn(
/home/zhanghui/anaconda3/envs/langchain/lib/python3.9/site-packages/transformers/generation/configuration_utils.py:367: UserWarning: `do_sample` is set to `False`. However, `top_p` is set to `0.6` -- this flag is only used in sample-based generation modes. You should set `do_sample=True` or unset `top_p`. This was detected when initializing the generation config instance, which means the corresponding file may hold incorrect parameterization and should be fixed.
warnings.warn(
Traceback (most recent call last):
File "/home/zhanghui/Chinese-LLaMA-Alpaca-2/scripts/langchain/langchain_sum.py", line 50, in <module>
model = HuggingFacePipeline.from_model_id(model_id=model_path,
File "/home/zhanghui/anaconda3/envs/langchain/lib/python3.9/site-packages/langchain/llms/huggingface_pipeline.py", line 112, in from_model_id
model.is_quantized
File "/home/zhanghui/anaconda3/envs/langchain/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1614, in __getattr__
raise AttributeError("'{}' object has no attribute '{}'".format(
AttributeError: 'LlamaForCausalLM' object has no attribute 'is_quantized'
### Expected behavior
| AttributeError: 'LlamaForCausalLM' object has no attribute 'is_quantized' | https://api.github.com/repos/langchain-ai/langchain/issues/11809/comments | 4 | 2023-10-14T13:46:33Z | 2023-10-16T23:54:22Z | https://github.com/langchain-ai/langchain/issues/11809 | 1,943,280,693 | 11,809 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I'm encountering an error while trying to use a Falcon model to chat with a bunch of text files. I've downloaded the model locally on my Google Drive, and everything seems fine up to this point. However, when I try to define the LLM model in Langchain using the following code:
```
from langchain.llms import HuggingFacePipeline
llm = HuggingFacePipeline.from_model_id(
model_id='/content/gdrive/MyDrive/Colab Notebooks/falcon model/898df1396f35e447d5fe44e0a3ccaaaa69f30d36',
task="text-generation",
model_kwargs={"temperature": 0, "max_length": 64},
)
```
I encounter the following error:
```
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-31-902f2455d555> in <cell line: 3>()
1 from langchain.llms import HuggingFacePipeline
2
----> 3 llm = HuggingFacePipeline.from_model_id(
4 model_id='/content/gdrive/MyDrive/Colab Notebooks/falcon model/898df1396f35e447d5fe44e0a3ccaaaa69f30d36',
5 task="text-generation",
AttributeError: 'FalconForCausalLM' object has no attribute 'is_quantized'
```
the whole code
```
device = torch.device('cuda')
checkpoint = "tiiuae/falcon-7b"
print(f"Checkpoint path: {checkpoint}") # Add this line for debugging
tokenizer = AutoTokenizer.from_pretrained(pretrained_model_name_or_path="/content/gdrive/MyDrive/Colab Notebooks/falcon model/898df1396f35e447d5fe44e0a3ccaaaa69f30d36")
base_model = AutoModel.from_pretrained(
pretrained_model_name_or_path='/content/gdrive/MyDrive/Colab Notebooks/falcon model/898df1396f35e447d5fe44e0a3ccaaaa69f30d36',
torch_dtype=torch.float32
)
loader = DirectoryLoader('/content/gdrive/MyDrive/Colab Notebooks/chatbot 3.0/', glob="./*.txt", loader_cls=TextLoader)
documents = loader.load()
text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=500)
texts = text_splitter.split_documents(documents)
embeddings = HuggingFaceEmbeddings(cache_folder='/content/gdrive/MyDrive/Colab Notebooks/falcon model/898df1396f35e447d5fe44e0a3ccaaaa69f30d36',)
from langchain.llms import HuggingFacePipeline
llm = HuggingFacePipeline.from_model_id(
model_id='/content/gdrive/MyDrive/Colab Notebooks/falcon model/898df1396f35e447d5fe44e0a3ccaaaa69f30d36',
task="text-generation",
model_kwargs={"temperature": 0, "max_length": 64},
)
```
### Suggestion:
_No response_ | Issue: : Issue with HuggingFacePipeline while trying to use a Falcon model for chat with text files | https://api.github.com/repos/langchain-ai/langchain/issues/11807/comments | 2 | 2023-10-14T12:59:34Z | 2024-02-07T16:18:58Z | https://github.com/langchain-ai/langchain/issues/11807 | 1,943,251,646 | 11,807 |
[
"langchain-ai",
"langchain"
] | ### Issue with current documentation:
[Link](https://python.langchain.com/docs/expression_language/cookbook/retrieval#with-memory-and-returning-source-documents)
Hi, I've been exploring the Cookbook RAG (link provided above) and noticed that `expanded_memory` is not defined anywhere in the code snippet(**With Memory and returning source documents**). I was wondering if a portion of the code is missing or if it's just a typo. Thank you in advance for looking over my issue!
### Idea or request for content:
_No response_ | DOC: LangChain Cookbook RAG typo | https://api.github.com/repos/langchain-ai/langchain/issues/11805/comments | 3 | 2023-10-14T09:54:28Z | 2023-10-15T14:39:12Z | https://github.com/langchain-ai/langchain/issues/11805 | 1,943,154,020 | 11,805 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
Sorry to raise this as an issue, but I have a quick question. I am preparing come code for `langchain` and noticed that most of the code I wish to build on lacks unit tests. I was considering first submitting a PR of purely unit tests for the existing code I wish to use in order to keep subsequent PRs more digestible. _Is submitting a PR of just unit tests acceptable?_
### Suggestion:
_No response_ | Issue: unit tests, a quick question. | https://api.github.com/repos/langchain-ai/langchain/issues/11803/comments | 2 | 2023-10-14T09:11:36Z | 2024-02-07T16:19:03Z | https://github.com/langchain-ai/langchain/issues/11803 | 1,943,107,712 | 11,803 |
[
"langchain-ai",
"langchain"
] | ### Feature request
- If user wants to get content from document which are having confidence HIGH or HIGHEST then with help of this feature Kendra will return only more relevance records.
### Motivation
- Sometimes kendra return content with LOW `ScoreConfidence` which may lead to inaccurate results.
### Your contribution
- I've checked the `_get_relevant_documents` method from `AmazonKendraRetriever` class but it is only returning top_k docs not actually filtering based on `ScoreConfidence`.
- With the help of `boto3` we can achieve that.
- I would like to do more research and open PR for it. | Add support for AWS Kendra `ScoreConfidence` filter. | https://api.github.com/repos/langchain-ai/langchain/issues/11801/comments | 2 | 2023-10-14T07:10:08Z | 2024-03-08T01:28:10Z | https://github.com/langchain-ai/langchain/issues/11801 | 1,943,036,454 | 11,801 |
[
"langchain-ai",
"langchain"
] | ### Feature request
The feature proposal is to integrate a new Langchain tool that would provide support for queries on [https://api.nasa.gov/](NASA APIs)
Here is the list of NASA APIs that will be integrated into the new tool:
- APOD
- Asteroids NeoWs
- Earth
- EONET
- EPIC
- GeneLab
- Insight
- Mars Rover Photos
- NASA Image and Video Library
- TechTransfer
- TLE API
- Vesta/Moon/Mars Trek WMTS
### Motivation
Currently, Langchain does not provide access to a large data pool of reliable information related to astronomy. The integration of a NASA API tool would allow users to retrieve updated and accurate data related to satellite imagery, mission data and other space-related information. In cases where accuracy of data is critical, the NASA API tool may also be preferred over its GPT-generated counterpart.
### Your contribution
I am part of a group of 3 other students that will be able to submit a PR for this issue sometime in November. | feat: add NASA API tool | https://api.github.com/repos/langchain-ai/langchain/issues/11787/comments | 8 | 2023-10-13T23:29:42Z | 2024-03-13T19:56:56Z | https://github.com/langchain-ai/langchain/issues/11787 | 1,942,774,990 | 11,787 |
[
"langchain-ai",
"langchain"
] | ### System Info
Python Version: 3.11.0
LangChain Version: 0.0.264
### Who can help?
_No response_
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [X] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
### define a custom chain:
```
class BlastAPIChain(Chain):
prompt: BasePromptTemplate
llm: BaseLanguageModel
@property
def input_keys(self) -> List[str]:
return ["prompt"]
@property
def output_keys(self) -> List[str]:
return ["response"]
@property
def _chain_type(self) -> str:
return "custom_chain"
def _call(
self,
inputs: Dict[str, Any],
run_manager: Optional[CallbackManagerForChainRun] = None
) -> Dict[str, str]:
'''
The primary method for this chain.
:param inputs: a dictionary containing the input keys and values.
:param run_manager: an optional callback manager for the chain run.
'''
# some custom logic goes here, but is omitted now
response = 'some custom logic would go here'
return {"response": response,}
```
### setup an instance of your custom chain
```
blast_api_chain_description = {
"name": "blast_api_chain",
"description": "lalalala, ",
"template": """ lalalala """,
}
blast_api_chain_prompt = PromptTemplate(
template=blast_api_chain_description["template"],
input_variables=[]
)
blast_api_chain = BlastAPIChain(llm=llm, verbose=True, memory=memory, prompt=blast_api_chain_prompt)
```
### add the destination chains to destination_chains
```
DESTINATION_CHAINS['blast_api_chain'] = blast_api_chain
```
### create a multi-prompt-chain from the destination chain
```
master_chain = MultiPromptChain(
router_chain=router_chain,
destination_chains=DESTINATION_CHAINS,
default_chain=default_chain,
verbose=True,
)
```
### observe the issue: the blast_router_chain destination has been reassigned
```
print (master_chain.__dict__['destination_chains']['blast_api_chain'])
print (DESTINATION_CHAINS['blast_api_chain'])
```
### Expected behavior
I expect that after defining `master_chain` with the code below, the `destination_chains` attribute should hold all of the same chains that are defined in `DESTINATION_CHAINS`.
```
master_chain = MultiPromptChain(
router_chain=router_chain,
destination_chains=DESTINATION_CHAINS,
default_chain=default_chain,
verbose=True,
)
```
However, this is not the case. Instead, the chains in DESTINATION_CHAINS have been over-written with different objects.
This can be seen here;
You can see that my custom BlastAPIChain instance has been replaced with an LLMChain instance. How am I supposed to route traffic to my custom chain?
| MultiPromptChain does not use its assigned destination chains | https://api.github.com/repos/langchain-ai/langchain/issues/11786/comments | 4 | 2023-10-13T23:23:40Z | 2024-02-09T16:15:58Z | https://github.com/langchain-ai/langchain/issues/11786 | 1,942,772,500 | 11,786 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
my Azure search index has 4 fields
id, user_id, content, content_vector.
I am trying to add a conversation to the azure search vector store for a user_id ="msunkara"
The code is given below:
memory_vector_store: AzureSearch = AzureSearch(
azure_search_endpoint=vector_store_address,
azure_search_key=vector_store_password,
index_name=memory_index_name,
embedding_function=embeddings.embed_query)
history = []
user_id = "msunkara"
session_id = "1"
query = Hi
res.content = Hi, How can I assist you today?
history.append("User: "+ query + " Assistant: " + res.content)
memory_id = session_id + user_id
metadata = [{"user_id": user_id}]
memory_vector_store.add_texts(history,keys = [memory_id],metadatas=metadata)
The above code is throwing the below error:
Result: Failure Exception: HttpResponseError: () The request is invalid. Details: The property 'metadata' does not exist on type 'search.documentFields'. Make sure to only use property names that are defined by the type.
Help me solve this problem
### Suggestion:
_No response_ | Issue: <add_texts method in AzureSearch class throwing a metadata error> | https://api.github.com/repos/langchain-ai/langchain/issues/11780/comments | 10 | 2023-10-13T19:51:52Z | 2024-02-14T16:09:28Z | https://github.com/langchain-ai/langchain/issues/11780 | 1,942,553,296 | 11,780 |
[
"langchain-ai",
"langchain"
] | ### Feature request
As it stands, `LLMRouterChain` basically only supports a single input. Here is the template:
https://github.com/langchain-ai/langchain/blob/v0.0.314/libs/langchain/langchain/chains/router/multi_prompt_prompt.py
While it is currently supported to define multiple input variables on a `PromptTemplate`, the `LLMRouterChain` basically doesn't support it because of the way the output from the LLM call is handled.
```
{{{{
"destination": string \\ name of the prompt to use or "DEFAULT"
"next_inputs": string \\ a potentially modified version of the original input
}}}}
```
With a lot of massaging, you can get to a place where an LLM causes `next_inputs` to be a comma-separated dict-like object but, down the chain (pun intended), you end up with an explosion because the way inputs are passed to chains doesn't handle this.
Overall, this implementation is odd. It expects the LLM to massage the original input(s) and then wants to use that output directly in future things.
Wouldn't it make more sense to simply ask the LLM to reason out which destination to send to (alone) and then take the original inputs and pass them along?
This current implementation is combining two things into one:
* a routing decision
* paraphrasing user input
The whole purpose of Langchain is to chain things, so why wouldn't you have "paraphrase/rephrase user input" as a chain step, instead of smushing it into the router implementation?
### Motivation
I have a use case where various inputs are collected separately and need to be maintained separately for a custom-tuned LLM later on. This caused me to attempt to use `LLMRouterChain` with multiple inputs, whereby I hit the roadblock described above.
It doesn't seem that wild to want to be able to make a routing decision based on input A, input B, or multiple inputs, and subsequently send all of those inputs to different templates later.
### Your contribution
I don't really have sufficient coding skills to help. I could maybe make some diagrams that help illustrate how this could work. | LLMRouterChain should easily support multiple inputs | https://api.github.com/repos/langchain-ai/langchain/issues/11778/comments | 4 | 2023-10-13T19:41:14Z | 2024-02-11T16:11:17Z | https://github.com/langchain-ai/langchain/issues/11778 | 1,942,541,303 | 11,778 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I tried to follow the contributing Quick Start [here](https://github.com/langchain-ai/langchain/blob/master/.github/CONTRIBUTING.md#-quick-start), but I'm running into issues when I run the unit tests using `make test`
Here are the failing tests:
```
FAILED tests/unit_tests/agents/test_agent_iterator.py::test_agent_iterator_stopped_early - AssertionError: assert 'curses foiled again' == 'Agent stopped due to iteration limit or time limit.'
FAILED tests/unit_tests/agents/test_agent_iterator.py::test_agent_iterator_reset - assert 0.0 > 0.0
FAILED tests/unit_tests/callbacks/test_callback_manager.py::test_callback_manager_with_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
FAILED tests/unit_tests/storage/test_filesystem.py::test_yield_keys - AssertionError: assert ['key1', 'subdir\\key2'] == ['key1', 'subdir/key2']
FAILED tests/unit_tests/tools/file_management/test_utils.py::test_get_validated_relative_path_errs_on_absolute - re.error: incomplete escape \U at position 53
FAILED tests/unit_tests/tools/file_management/test_utils.py::test_get_validated_relative_path_errs_on_parent_dir - re.error: incomplete escape \U at position 69
FAILED tests/unit_tests/tools/file_management/test_utils.py::test_get_validated_relative_path_errs_for_symlink_outside_root - re.error: incomplete escape \U at position 64
FAILED tests/unit_tests/tools/file_management/test_utils.py::test_get_validated_relative_path_for_symlink_inside_root - langchain.tools.file_management.utils.FileValidationError: Path symlink_inside_root is outside of the allowed direc...
FAILED tests/unit_tests/tools/requests/test_tool.py::test_requests_get_tool - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
FAILED tests/unit_tests/tools/requests/test_tool.py::test_requests_post_tool - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
FAILED tests/unit_tests/tools/requests/test_tool.py::test_requests_patch_tool - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
FAILED tests/unit_tests/tools/requests/test_tool.py::test_requests_put_tool - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
FAILED tests/unit_tests/tools/requests/test_tool.py::test_requests_delete_tool - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/agents/test_agent_iterator.py::test_agent_async_iterator_stopped_early - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/agents/test_agent_iterator.py::test_agent_async_iterator_with_callbacks - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/agents/test_agent_iterator.py::test_agent_async_iterator_output_structure - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/callbacks/test_callback_manager.py::test_callback_manager_with_async_with_running_loop - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/callbacks/test_callback_manager.py::test_async_callback_manager - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/callbacks/test_callback_manager.py::test_async_callback_manager_sync_handler - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/chains/test_sequential.py::test_simple_sequential_functionality_with_callbacks[False] - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/chains/test_sequential.py::test_simple_sequential_functionality_with_callbacks[True] - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/indexes/test_indexing.py::test_abatch - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/output_parsers/test_json.py::test_partial_text_json_output_parser_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/output_parsers/test_json.py::test_partial_functions_json_output_parser_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/output_parsers/test_json.py::test_partial_text_json_output_parser_diff_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/output_parsers/test_json.py::test_partial_functions_json_output_parser_diff_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/retrievers/test_base.py::test_fake_retriever_v1_upgrade_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/retrievers/test_base.py::test_fake_retriever_v1_with_kwargs_upgrade_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/retrievers/test_base.py::test_fake_retriever_v2_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/retrievers/self_query/test_base.py::test__aget_relevant_documents - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_locals.py::test_put_get_async[<lambda>-foo-foo] - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_locals.py::test_put_get_async[<lambda>-input1-output1] - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_passthrough_tap_async - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_with_config - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_default_method_implementations - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_prompt - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_prompt_with_chat_model - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_prompt_with_llm - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_stream_log_retriever - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_prompt_with_llm_and_async_lambda - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_router_runnable - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_higher_order_lambda_runnable - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_map_astream - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_map_astream_iterator_input - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_deep_astream - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_deep_astream_assign - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_runnable_sequence_atransform - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_llm_with_fallbacks[llm_with_fallbacks] - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_llm_with_fallbacks[llm_with_multi_fallbacks] - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_llm_with_fallbacks[llm_chain_with_fallbacks] - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_async_retrying - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_seq_abatch_return_exceptions - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_runnable_branch_ainvoke - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_runnable_branch_ainvoke_callbacks - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_runnable_branch_abatch - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_tool_from_runnable - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_runnable_gen - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/schema/runnable/test_runnable.py::test_runnable_gen_transform - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/smith/evaluation/test_runner_utils.py::test_arun_on_dataset - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_base.py::test_create_async_tool - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_base.py::test_async_exception_handling_bool - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_base.py::test_async_exception_handling_str - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_base.py::test_async_exception_handling_callable - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_base.py::test_async_exception_handling_non_tool_exception - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_zapier.py::test_apreview - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_zapier.py::test_arun - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/test_zapier.py::test_alist - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
ERROR tests/unit_tests/tools/shell/test_shell.py::test_shell_tool_arun - pytest_socket.SocketBlockedError: A test tried to use socket.socket.
```
Here is the complete log:
[test-log.txt](https://github.com/langchain-ai/langchain/files/12895952/test-log.txt)
### Suggestion:
_No response_ | Issue: Many unit tests failing when running on Windows | https://api.github.com/repos/langchain-ai/langchain/issues/11775/comments | 3 | 2023-10-13T17:01:59Z | 2023-12-21T21:45:43Z | https://github.com/langchain-ai/langchain/issues/11775 | 1,942,295,198 | 11,775 |
[
"langchain-ai",
"langchain"
] | ### System Info
Using `langchain==0.0.300`
### Who can help?
_No response_
### Information
- [X] The official example notebooks/scripts
- [x] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [X] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
I tried running the [ElasticsearchDatabaseChain](https://python.langchain.com/docs/use_cases/qa_structured/sql#elastic-search) on a simple Elasticsearch database that contains one index as defined below, but I always get a `SerializationError` due to some faulty json.
Here are the details:
1. Get a local Elasticsearch instance running
```
docker run -p 9200:9200 -e "discovery.type=single-node" -e "xpack.security.enabled=false" -e "xpack.security.http.ssl.enabled=false" docker.elastic.co/elasticsearch/elasticsearch:8.10.2
```
2. Create the index mapping as follows
```bash
curl --request PUT \
--url 'http://localhost:9200/products-index?pretty=' \
--header 'content-type: application/json' \
--header 'user-agent: vscode-restclient' \
--data '{"settings": {"index.mapping.coerce": false},"mappings": {"properties": {"name": { "type": "text" },"price": { "type": "double" },"reviews": { "type": "text" },"createdAt": { "type": "date", "format": "epoch_second" }},"dynamic": true}}'
```
3. Bulk insert some data
```bash
curl --request POST \
--url http://localhost:9200/_bulk \
--header 'content-type: application/x-ndjson' \
--header 'user-agent: vscode-restclient' \
--data '{ "index": { "_index": "products-index" } }{ "name": "Bulk Item 666", "price": 666, "reviews": ["Good"], "createdAt": 1697213529 }{ "index": { "_index": "products-index" } }{ "name": "Bulk Item 123", "price": 123, "reviews": ["Good"], "createdAt": 1697213529 }'
```
4. Create and run the chain
```python
from elasticsearch import Elasticsearch
from langchain import OpenAI
from langchain.chains.elasticsearch_database import ElasticsearchDatabaseChain
db = Elasticsearch("http://localhost:9200")
chain = ElasticsearchDatabaseChain.from_llm(
llm=OpenAI(),
database=db,
verbose=True,
)
result = chain.run(prompt)
print(result)
```
<details>
<summary>Traceback (most recent call last):</summary>
> File "/home/sindeavour/Documents/github/sinaapps/datavacuum-old/ai-service/src/routes/v1/elastic.py", line 82, in elastic
result = chain.run(prompt)
^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/langchain/chains/base.py", line 487, in run
return self(args[0], callbacks=callbacks, tags=tags, metadata=metadata)[
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/langchain/chains/base.py", line 292, in __call__
raise e
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/langchain/chains/base.py", line 286, in __call__
self._call(inputs, run_manager=run_manager)
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/langchain/chains/elasticsearch_database/base.py", line 127, in _call
indices_info = self._get_indices_infos(indices)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/langchain/chains/elasticsearch_database/base.py", line 100, in _get_indices_infos
hits = self.database.search(
^^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/elasticsearch/_sync/client/utils.py", line 414, in wrapped
return api(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/elasticsearch/_sync/client/__init__.py", line 3924, in search
return self.perform_request( # type: ignore[return-value]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/elasticsearch/_sync/client/_base.py", line 285, in perform_request
meta, resp_body = self.transport.perform_request(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/elastic_transport/_transport.py", line 347, in perform_request
data = self.serializers.loads(raw_data, meta.mimetype)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/elastic_transport/_serializer.py", line 196, in loads
return self.get_serializer(mimetype).loads(data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/sindeavour/.pyenv/versions/3.11.5/envs/ai-service/lib/python3.11/site-packages/elastic_transport/_serializer.py", line 95, in loads
raise SerializationError(
elastic_transport.SerializationError: Unable to deserialize as JSON: b'{"took":1,"timed_out":false,"_shards":{"total":1,"successful":1,"skipped":0,"failed":0},"hits":{"total":{"value":4,"relation":"eq"},"max_score":1.0,"hits":[{"_index":"products-index","_id":"1","_score":1.0,"_source":{\n "name": "Samsung S22 Ultra",\n "price": 975.99,\n "reviews": ["Good"],\n "createdAt": 1697145247 // epoch_second\n}},{"_index":"products-index","_id":"pRG9JYsB0QTitqW3X0qT","_score":1.0,"_source":{\n "name": "iPhone 14 Pro",\n "price": 1099,\n "reviews": ["Good"],\n "createdAt": 1697145249 // epoch_second\n}},{"_index":"products-index","_id":"phG9JYsB0QTitqW3ZEq5","_score":1.0,"_source":{ "name": "Bulk Item 666", "price": 666, "reviews": ["Good"], "createdAt": 1697145250 }}]}}'
</details>
```
elastic_transport.SerializationError: Unable to deserialize as JSON: b'{"took":0,"timed_out":false,"_shards":{"total":1,"successful":1,"skipped":0,"failed":0},"hits":{"total":{"value":4,"relation":"eq"},"max_score":1.0,"hits":[{"_index":"products-index","_id":"1","_score":1.0,"_source":{\n "name": "Samsung S22 Ultra",\n "price": 975.99,\n "reviews": ["Good"],\n "createdAt": 1697145247 // epoch_second\n}},{"_index":"products-index","_id":"pRG9JYsB0QTitqW3X0qT","_score":1.0,"_source":{\n "name": "iPhone 14 Pro",\n "price": 1099,\n "reviews": ["Good"],\n "createdAt": 1697145249 // epoch_second\n}},{"_index":"products-index","_id":"phG9JYsB0QTitqW3ZEq5","_score":1.0,"_source":{ "name": "Bulk Item 666", "price": 666, "reviews": ["Good"], "createdAt": 1697145250 }}]}}'
```
If you look closely, you can see that there is an invalid comment `// epoch_second` in multiple places in the generated json, hence why the parsing fails.
### Expected behavior
The chain should run without issues when executing `chain.run(prompt)` | ElasticsearchDatabaseChain SerializationError | https://api.github.com/repos/langchain-ai/langchain/issues/11771/comments | 4 | 2023-10-13T16:35:06Z | 2023-11-04T04:21:52Z | https://github.com/langchain-ai/langchain/issues/11771 | 1,942,262,577 | 11,771 |
[
"langchain-ai",
"langchain"
] | ### Feature request
Implement a feature using Langchain's image_captions.py and audio_speech_to_text.py to produce .srt files. This system will provide both subtitles and visual scene descriptions, essentially creating closed captioning.
### Motivation
Current closed captioning primarily captures spoken content, omitting crucial visual elements of videos. This lack of comprehensive description hinders a full understanding, especially for the hearing-impaired community. Our goal is to elevate video accessibility by integrating both visual and auditory components in closed captioning, ensuring a more inclusive viewing experience.
### Your contribution
We are students at the University of Toronto, interested in contributing to Langchain for a school project. We would like to submit a PR that implements this feature sometime in November. | Video imagery to text (Closed Captioning) | https://api.github.com/repos/langchain-ai/langchain/issues/11770/comments | 8 | 2023-10-13T16:29:52Z | 2024-03-18T16:05:44Z | https://github.com/langchain-ai/langchain/issues/11770 | 1,942,255,454 | 11,770 |
[
"langchain-ai",
"langchain"
] | ### System Info
I encountered a very strange problem when using AsyncCallbackHandler. Most of the time, it works normally, but occasionally strange problems occur. Therefore, I logged the entire call process in each function, and finally learned from the logs that under normal circumstances, the logs are on_ llm_ start, then the API call log for calling openai, and finally on_llm_ end, but occasionally this is not the case, and in exceptional cases, the log order is on_ llm_ start, on_ llm_ end, followed by the API call log of OpenAI, which occasionally prevents me from obtaining the complete token concatenated string because I need to record the complete answer of AI each time. Therefore, I checked the source code and temporarily did not find how the while loop broke out. While not self. queue. empty() or not self. done.is_ Set () This needs to satisfy both that the queue is not empty and that done has been set. Where are the specific issues and I hope to receive answers
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [X] Callbacks/Tracing
- [X] Async
### Reproduction
class AsyncIteratorCallbackHandler(AsyncCallbackHandler):
"""Callback handler that returns an async iterator."""
queue: asyncio.Queue[str]
done: asyncio.Event
@property
def always_verbose(self) -> bool:
return True
def __init__(self) -> None:
self.queue = asyncio.Queue()
self.done = asyncio.Event()
async def on_llm_start(
self, serialized: Dict[str, Any], prompts: List[str], **kwargs: Any
) -> None:
# If two calls are made in a row, this resets the state
self.done.clear()
async def on_llm_new_token(self, token: str, **kwargs: Any) -> None:
if token is not None and token != "":
self.queue.put_nowait(token)
async def on_llm_end(self, response: LLMResult, **kwargs: Any) -> None:
self.done.set()
async def on_llm_error(
self, error: Union[Exception, KeyboardInterrupt], **kwargs: Any
) -> None:
self.done.set()
# TODO implement the other methods
async def aiter(self) -> AsyncIterator[str]:
while not self.queue.empty() or not self.done.is_set():
# Wait for the next token in the queue,
# but stop waiting if the done event is set
done, other = await asyncio.wait(
[
# NOTE: If you add other tasks here, update the code below,
# which assumes each set has exactly one task each
asyncio.ensure_future(self.queue.get()),
asyncio.ensure_future(self.done.wait()),
],
return_when=asyncio.FIRST_COMPLETED,
)
# Cancel the other task
if other:
other.pop().cancel()
# Extract the value of the first completed task
token_or_done = cast(Union[str, Literal[True]], done.pop().result())
# If the extracted value is the boolean True, the done event was set
if token_or_done is True:
break
# Otherwise, the extracted value is a token, which we yield
yield token_or_done
### Expected behavior
AsyncCallbackHandler can execute strictly in the order of on_llm_start, openai call(on_llm_new_token), and on_llm_end each time. | When using Streaming and AsyncCallbackHandler, sometimes it is not possible to obtain the complete token concatenation string. | https://api.github.com/repos/langchain-ai/langchain/issues/11769/comments | 3 | 2023-10-13T14:56:55Z | 2024-02-09T16:16:13Z | https://github.com/langchain-ai/langchain/issues/11769 | 1,942,118,596 | 11,769 |
[
"langchain-ai",
"langchain"
] | ### System Info
llm = ChatOpenAI(model_name="gpt-4", temperature=0.5, request_timeout=300)
qa_chain = ConversationalRetrievalChain.from_llm(
verbose=True,
llm=llm,
retriever=retriever,
return_source_documents=True,
combine_docs_chain_kwargs={"prompt": prompt},
rephrase_question=False,
)
given code I get error:
Retrying langchain.embeddings.openai.embed_with_retry.<locals>._embed_with_retry in 4.0 seconds as it raised APIConnectionError: Error communicating with OpenAI: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response')).
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [X] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
llm = ChatOpenAI(model_name="gpt-4", temperature=0.5, request_timeout=300)
qa_chain = ConversationalRetrievalChain.from_llm(
verbose=True,
llm=llm,
retriever=retriever,
return_source_documents=True,
combine_docs_chain_kwargs={"prompt": prompt},
rephrase_question=False,
)
### Expected behavior
successfully call OpenAI embedding api | Retrying langchain.embeddings.openai.embed_with_retry.<locals>._embed_with_retry in 4.0 seconds as it raised APIConnectionError: Error communicating with OpenAI: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response')). | https://api.github.com/repos/langchain-ai/langchain/issues/11767/comments | 2 | 2023-10-13T11:51:17Z | 2024-02-11T16:11:22Z | https://github.com/langchain-ai/langchain/issues/11767 | 1,941,800,234 | 11,767 |
[
"langchain-ai",
"langchain"
] | ### Feature request
I would like to propose a new vector database to be included in the Langchain ecosystem to expand Langchain's influence. Below is the official link to our vector database.
https://www.transwarp.cn/starwarp
### Motivation
I hope that everyone can more conveniently use Hippo + Langchain to build AI Agents.
### Your contribution
Integrating the Hippo vector database into the Langchain ecosystem can expand Langchain's influence. | Add a new vector store hippo to langchain | https://api.github.com/repos/langchain-ai/langchain/issues/11763/comments | 1 | 2023-10-13T10:54:02Z | 2024-02-06T16:21:51Z | https://github.com/langchain-ai/langchain/issues/11763 | 1,941,719,788 | 11,763 |
[
"langchain-ai",
"langchain"
] | ### Feature request
It would be great to make it possible to use multiple retrievers, LLM, and conversation history together.
### Motivation
Currently, when using `MultiRetrievalQAChain.fromLLMAndRetrievers`, you can provide an LLM and multiple retrievers for context. If there was a possibility to add a conversation/chat history to this flow, it could help further understand the context, especially in cases where there was a previous conversation between a user and a human agent/bot.
### Your contribution
I'm not familiar with Python but could provide help with the TypeScript implementation on top of the base logic. | Use conversation history with MultiRetrievalQAChain | https://api.github.com/repos/langchain-ai/langchain/issues/11760/comments | 2 | 2023-10-13T10:28:58Z | 2024-02-20T16:14:35Z | https://github.com/langchain-ai/langchain/issues/11760 | 1,941,683,584 | 11,760 |
[
"langchain-ai",
"langchain"
] | ### System Info
Google Colab
langchain==0.0.313
transformers==4.34.0
Python 3.10.12
[Dependencies](https://pastebin.com/raw/T1BqzVq1)
I am trying to download huggingface model and feed them to the chain. It was working prior to few weeks. I have tried to change model id to other models.
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```
!pip install langchain transformers
from langchain import HuggingFacePipeline
llm = HuggingFacePipeline.from_model_id(model_id = "bigscience/bloom-560m", task = "text-generation")
```
### Expected behavior
Successful execution
## Screenshot
| AttributeError: 'BloomForCausalLM' object has no attribute 'is_quantized' | https://api.github.com/repos/langchain-ai/langchain/issues/11759/comments | 6 | 2023-10-13T09:35:49Z | 2024-02-11T16:11:27Z | https://github.com/langchain-ai/langchain/issues/11759 | 1,941,602,588 | 11,759 |
[
"langchain-ai",
"langchain"
] | ### Issue with current documentation:
From the page
https://python.langchain.com/docs/integrations/platforms/aws
If you select See a [usage example.](https://python.langchain.com/docs/integrations/llms/amazon_api_gateway_example) the following message is generated:
**Page Not Found**
We could not find what you were looking for.
Please contact the owner of the site that linked you to the original URL and let them know their link is broken.
https://python.langchain.com/docs/integrations/llms/amazon_api_gateway_example
### Idea or request for content:
_No response_ | DOC:**Page Not Found** | https://api.github.com/repos/langchain-ai/langchain/issues/11758/comments | 1 | 2023-10-13T09:16:52Z | 2024-03-13T20:03:18Z | https://github.com/langchain-ai/langchain/issues/11758 | 1,941,572,219 | 11,758 |
[
"langchain-ai",
"langchain"
] | ### Feature request
will you update to follow the update of https://github.com/openai/openai-python/discussions/631 ?
### Motivation
Many issues in langchain point that when calling async api and using openai LLM(Chains like MapReduceDocumentsChain), it will raise '[Error Response payload is not complete](https://github.com/langchain-ai/langchain/issues/9187#top)' or [TimeoutError](https://github.com/langchain-ai/langchain/issues/7148)
After discussion, they all point to [an openai issue](https://github.com/openai/openai-python/issues/574), which will be fix in v1.0.0(now v1.0.0 beta is avaliable)
So if you can adapt to their updates, this critical issue can be fixed
### Your contribution
none | Adapt to openai-python v1.0.0 Beta | https://api.github.com/repos/langchain-ai/langchain/issues/11755/comments | 11 | 2023-10-13T07:18:30Z | 2024-02-15T16:08:42Z | https://github.com/langchain-ai/langchain/issues/11755 | 1,941,385,504 | 11,755 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I am trying to build an agent with multiple tools. Based on the query agent may need to execute single or multiple tools. Some of the tools may ask clarification question to the user. In that case agent should add the question to the final answer along with other tools response. But it wont work as expected. Also It executes the same tool again and again with similar kind of input if the tools response not satisfactory to it.
My agent has 4 tools. I have added portion of the code and a screenshot for better understanding.
I have also tried to modify the prefix, suffix and format_instruction in so many ways but no luck. Also I know return_direct=true will return the tools response to the user directly but this is not my case as this flag won't execute multiple tools.
**Expectation:** Agent should return the tools clarification question to the user along with other tools response instead of answering on it's own and execute same tool for similar kind of input. Also Agent should reply proper answer if the Query not related to tools. Eg: If the query is just "Hello" it produce error.
```
FORMAT_INSTRUCTIONS = """Use the following format:
Question: the input question you must answer
Thought: you should always think about what to do
Action: the action to take, should be one of [{tool_names}]
Action Input: the input to the action
Observation: the result of the action
... (this Thought/Action/Action Input/Observation can repeat N times)
Thought: I now know the final answer
Final Answer: the final answer to the original input question"""
PREFIX_PROMPT = """Have a conversation with a human, answering the following questions as best you can. You have access to the following tools:"""
SUFFIX_PROMPT = """Begin!
Previous conversation history:
{chat_history}
Question: {input}
Thought:{agent_scratchpad}"""
class ChatService(RequestAccessorService):
def __init__(self, model: SenseChatModel, request: Request):
super().__init__(request)
self.model = model
async def execute_async(self):
tools = [
QATool(self.model, self),
RecordCreateTool(self.model, self),
RecordSearchTool(self.model, self),
AppCreateTool(self.model, self)
]
llm = ChatOpenAI(openai_api_key=self.model.openai_api_key, temperature=0)
prompt = ZeroShotAgent.create_prompt(
tools,
prefix=PREFIX_PROMPT,
suffix=SUFFIX_PROMPT,
format_instructions=FORMAT_INSTRUCTIONS,
input_variables=["input", "chat_history", "agent_scratchpad"],
)
llm_chain = LLMChain(llm=llm, prompt=prompt)
agent = ZeroShotAgent(llm_chain=llm_chain, tools=tools, verbose=True)
agent_chain = AgentExecutor.from_agent_and_tools(
agent=agent, tools=tools, verbose=True
)
print(agent.llm_chain.prompt.template)
result = await agent_chain.arun(input=self.model.query)
return result
```
**Query:** Give me the all product in the inventory and create a product named dell in inventory
**Output**: The output of the query. Please see the Action, Action Input and observation to understand the whole scenario.
### Suggestion:
_No response_ | Question (Urgent): Agent with multiple tools doesn't work as expected and also caused infinite loop. | https://api.github.com/repos/langchain-ai/langchain/issues/11751/comments | 3 | 2023-10-13T06:26:38Z | 2024-03-29T16:06:25Z | https://github.com/langchain-ai/langchain/issues/11751 | 1,941,321,112 | 11,751 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
_No response_
### Suggestion:
_No response_ | langchain agent如何知道选择哪个Tool? | https://api.github.com/repos/langchain-ai/langchain/issues/11750/comments | 2 | 2023-10-13T05:25:20Z | 2024-02-08T16:19:20Z | https://github.com/langchain-ai/langchain/issues/11750 | 1,941,264,138 | 11,750 |
[
"langchain-ai",
"langchain"
] | ### System Info
```sh
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/local/lib/python3.8/dist-packages/langchain/__init__.py", line 8, in <module>
from langchain.agents import MRKLChain, ReActChain, SelfAskWithSearchChain
File "/usr/local/lib/python3.8/dist-packages/langchain/agents/__init__.py", line 2, in <module>
from langchain.agents.agent import Agent
File "/usr/local/lib/python3.8/dist-packages/langchain/agents/agent.py", line 10, in <module>
from langchain.chains.base import Chain
File "/usr/local/lib/python3.8/dist-packages/langchain/chains/__init__.py", line 2, in <module>
from langchain.chains.conversation.base import ConversationChain
File "/usr/local/lib/python3.8/dist-packages/langchain/chains/conversation/base.py", line 7, in <module>
from langchain.chains.conversation.memory import ConversationBufferMemory
File "/usr/local/lib/python3.8/dist-packages/langchain/chains/conversation/memory.py", line 7, in <module>
from langchain.chains.conversation.prompt import SUMMARY_PROMPT
File "/usr/local/lib/python3.8/dist-packages/langchain/chains/conversation/prompt.py", line 2, in <module>
from langchain.prompts.prompt import PromptTemplate
File "/usr/local/lib/python3.8/dist-packages/langchain/prompts/__init__.py", line 2, in <module>
from langchain.prompts.base import BasePromptTemplate
File "/usr/local/lib/python3.8/dist-packages/langchain/prompts/base.py", line 35, in <module>
class BasePromptTemplate(BaseModel, ABC):
File "/usr/local/lib/python3.8/dist-packages/langchain/prompts/base.py", line 41, in BasePromptTemplate
@root_validator()
File "/usr/local/lib/python3.8/dist-packages/pydantic/deprecated/class_validators.py", line 228, in root_validator
raise PydanticUserError(
pydantic.errors.PydanticUserError: If you use `@root_validator` with pre=False (the default) you MUST specify `skip_on_failure=True`. Note that `@root_validator` is deprecated and should be replaced with `@model_validator`.
```
### Who can help?
@agola11
```python
import langchain
```
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
1. Install langchain latest version on Linux
2. import langchain
### Expected behavior
Nothing. | @root_validator error | https://api.github.com/repos/langchain-ai/langchain/issues/11749/comments | 3 | 2023-10-13T04:51:51Z | 2023-10-18T18:52:26Z | https://github.com/langchain-ai/langchain/issues/11749 | 1,941,233,288 | 11,749 |
[
"langchain-ai",
"langchain"
] | ### Feature request
We would like to add Slack toolkit, which for a slack user, he can:
1. fetch information of workplaces (name, user type, time joined, etc.) he is a member of
2. fetch chat history across workplaces he is a member, and further process/analyze such data with LLM
3. fetch interaction (reply, mention, DM, etc) with other users across workplaces.
4. fetch a list of new activities across workplaces since last time login to slack platform,
5. and react to these activities.
Also, supports filters (by time, keyword etc) as appropriate for all the above functions
### Motivation
Currently lang-chain provides limited support for Slack, specifically, only supports loader of chat history zip file. On the other hand, slack platform analytics interface lacks aggregated information across workplaces; also most analytics info if provides are limited to owner/admin user type. Therefore, we want to implement the above feature:
1. Add-on current langchain toolkit: slack toolkit would be the one of the few, if not the first social media related toolkit.
2. Expose slack interface tools' functions to more user types: Only specific slack user type (workspace owner and/or admin) has access to downloading file option, along with most analytics tools embedded in slack user interface, which calls for a toolkit for other user types to access such information.
3. Better Info Aggregation: We notice that a user can be in multiple workspaces yet currently there lacks a way for user to get/compare information (real-time and logs) across workspaces they are member of. It would be nice for lang chain to provide a toolkit that provides user with such information upon request, so that the user can have all his/her user info across workspace at one place.
### Your contribution
We have started working on the feature and intend to submit PR in mid November. | Add Slack Toolkit | https://api.github.com/repos/langchain-ai/langchain/issues/11747/comments | 4 | 2023-10-13T01:32:38Z | 2024-02-11T16:11:36Z | https://github.com/langchain-ai/langchain/issues/11747 | 1,941,067,816 | 11,747 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
_No response_
### Suggestion:
_No response_ | 怎么控制langchain agent的意图分类的类别数？ | https://api.github.com/repos/langchain-ai/langchain/issues/11746/comments | 3 | 2023-10-13T00:48:54Z | 2024-02-08T16:19:30Z | https://github.com/langchain-ai/langchain/issues/11746 | 1,941,036,248 | 11,746 |
[
"langchain-ai",
"langchain"
] | ### System Info
Hi, I've implemented my own MultiRouteChain
### Who can help?
@UmerHA
@hwchase17
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [X] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
This is the implementation of mycallbackhandler, taken from this documentation:
https://github.com/langchain-ai/langchain/issues/5433
class MyCallbackHandler(BaseCallbackHandler):
def on_llm_new_token(self, token, **kwargs) -> None:
# print every token on a new line
print(f"#{token}#")
This is part of the implementation of my classs:
class NestedRouterMultiType(MultiRouteChain):
@classmethod
def from_prompts_and_retrievers(
cls,
llm: BaseLanguageModel,
prompt_infos: List[Dict[str, str]],
retrieval_infos: List[Dict[str, Any]],
router_infos: List[Dict[str, Dict[str, List[Dict[str, Any]]]]],
default_chain: Optional[Chain] = None,
#We can include default retrievers, but for now it is useful to just use GPT as the default chain
return_source_documents: bool = False,
verbose: bool = False,
**kwargs: Any,
) -> NestedRouterMultiType:
"""Convenience constructor for instantiating from destination prompts and retrievers."""
# Combine prompt and retriever infos for destinations
destinations = [f"{p['name']}: {p['description']}" for p in prompt_infos + retrieval_infos + router_infos]
destinations_str = "\n".join(destinations)
router_template = MULTI_TYPE_ROUTER_TEMPLATE_TEST.format(
destinations=destinations_str
)
router_prompt = PromptTemplate(
template=router_template,
input_variables=["input", "context"],
output_parser=RouterOutputParserConversational(),
)
router_chain = LLMRouterChainConversational.from_llm(llm, router_prompt)
destination_chains = {}
for p_info in prompt_infos:
name = p_info["name"]
prompt_template = p_info["prompt_template"]
input_variables = cls._extract_variables_from_template(prompt_template)
# Include here input variables input and history and test whether history is being passed correctly
#Deprecated line of code:
#prompt = PromptTemplate(template=prompt_template, input_variables=["question", "chat_history"])
prompt = PromptTemplate(template=prompt_template, input_variables=input_variables)
chain = UnifiedLLMChain(
prompt=prompt,
llm=llm,
verbose=verbose,
callbacks=[MyCallbackHandler()],
)
destination_chains[name] = chain
for r_info in retrieval_infos:
# refactor this to pass all arguments in PromptFactory
name = r_info["name"]
prompt = r_info.get("prompt") #try to retrieve a prompt, if empty, use None (not sure yet what this prompt is for)
retriever = r_info["retriever"]
condense_prompt_template = r_info["condense_prompt_template"]
prompt_template = r_info["prompt_template"]
# Here's where we create a Router that can route to ConversationalRetrievalChains
chain = ConversationalRetrievalChainHistory.from_llm(
llm,
retriever=retriever,
combine_docs_chain_kwargs={"prompt": prompt_template},
condense_question_prompt=condense_prompt_template,
return_source_documents = return_source_documents,
verbose=verbose,
callbacks=[MyCallbackHandler()],
)
destination_chains[name] = chain
### Expected behavior
I expect to print all tokens on each chain. The output is nonexistent. Either I'm missing something or it is a bug. The streaming parameter is set to True on BaseLanguageModel. | BaseCallbackHandler on_llm_new_token not working with custom MultiRouteChain | https://api.github.com/repos/langchain-ai/langchain/issues/11739/comments | 4 | 2023-10-12T22:41:51Z | 2024-02-10T16:13:27Z | https://github.com/langchain-ai/langchain/issues/11739 | 1,940,902,588 | 11,739 |
[
"langchain-ai",
"langchain"
] | ### System Info
Platform: Windows Server 2022
Python: 3.11.6
Langchain version: 0.0.306
### Who can help?
@agola11 @hwchase17
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [X] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [X] Agents / Agent Executors
- [X] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
### Creating a test tool
```py
from langchain.agents import Tool
tools = [
Tool(
name="test_tool",
func=print,
description="This is a test tool"
)
]
tools
```
```
[Tool(name='test_tool', description='This is a test tool', args_schema=None, return_direct=False, verbose=False, callbacks=None, callback_manager=None, tags=None, metadata=None, handle_tool_error=False, func=<built-in function print>, coroutine=None)]
```
### Setting up the pandas_dataframe_agent
```py
from langchain.agents import create_pandas_dataframe_agent
from langchain.llms import HuggingFacePipeline
import pandas as pd
llm = HuggingFacePipeline.from_model_id(
model_id="google/flan-t5-small", task="text2text-generation", device=0
)
agent = create_pandas_dataframe_agent(llm, pd.DataFrame(), verbose=True, extra_tools=tools)
agent.tools
```
```
[PythonAstREPLTool(name='python_repl_ast', description='A Python shell. Use this to execute python commands. Input should be a valid python command. When using this tool, sometimes output is abbreviated - make sure it does not look abbreviated before using it in your answer.', args_schema=<class 'langchain.tools.python.tool.PythonInputs'>, return_direct=False, verbose=False, callbacks=None, callback_manager=None, tags=None, metadata=None, handle_tool_error=False, globals={}, locals={'df': Empty DataFrame
Columns: []
Index: []}, sanitize_input=True),
Tool(name='test_tool', description='This is a test tool', args_schema=None, return_direct=False, verbose=False, callbacks=None, callback_manager=None, tags=None, metadata=None, handle_tool_error=False, func=<built-in function print>, coroutine=None)]
```
### Executing agent with debugging enabled
```py
import langchain
langchain.debug = True
agent.run('What is 2+2?')
```
```
[chain/start] [1:chain:AgentExecutor] Entering Chain run with input:
{
"input": "What is 2+2?"
}
[chain/start] [1:chain:AgentExecutor > 2:chain:LLMChain] Entering Chain run with input:
{
"input": "What is 2+2?",
"agent_scratchpad": "",
"stop": [
"\nObservation:",
"\n\tObservation:"
]
}
[llm/start] [1:chain:AgentExecutor > 2:chain:LLMChain > 3:llm:HuggingFacePipeline] Entering LLM run with input:
{
"prompts": [
"You are working with a pandas dataframe in Python. The name of the dataframe is `df`.\nYou should use the tools below to answer the question posed of you:\n\npython_repl_ast: A Python shell. Use this to execute python commands. Input should be a valid python command. When using this tool, sometimes output is abbreviated - make sure it does not look abbreviated before using it in your answer.\n\nUse the following format:\n\nQuestion: the input question you must answer\nThought: you should always think about what to do\nAction: the action to take, should be one of [python_repl_ast]\nAction Input: the input to the action\nObservation: the result of the action\n... (this Thought/Action/Action Input/Observation can repeat N times)\nThought: I now know the final answer\nFinal Answer: the final answer to the original input question\n\n\nThis is the result of `print(df.head())`:\n\n\nBegin!\nQuestion: What is 2+2?"
]
}
```
### The prompt from the above log
```
You are working with a pandas dataframe in Python. The name of the dataframe is `df`.
You should use the tools below to answer the question posed of you:
python_repl_ast: A Python shell. Use this to execute python commands. Input should be a valid python command. When using this tool, sometimes output is abbreviated - make sure it does not look abbreviated before using it in your answer.
Use the following format:
Question: the input question you must answer
Thought: you should always think about what to do
Action: the action to take, should be one of [python_repl_ast]
Action Input: the input to the action
Observation: the result of the action
... (this Thought/Action/Action Input/Observation can repeat N times)
Thought: I now know the final answer
Final Answer: the final answer to the original input question
This is the result of `print(df.head())`:
Begin!
Question: What is 2+2?
```
### Expected behavior
Where did my custom tool `test_tool` disappear? I expected it to show up after python_repl_ast? | `extra_tools` argument in `create_pandas_dataframe_agent()` doesn't seem to be working | https://api.github.com/repos/langchain-ai/langchain/issues/11737/comments | 3 | 2023-10-12T22:22:09Z | 2024-02-05T15:14:43Z | https://github.com/langchain-ai/langchain/issues/11737 | 1,940,885,814 | 11,737 |
[
"langchain-ai",
"langchain"
] | ### Feature request
There is an option to specify the prompt sent to LLM and also the CONDENSE_QUESTION prompt. But there is no way to specify the prompt that is being used by SelfQueryRetriever. This can be helpful to provide additional context to the LLM in order to construct a proper query.
### Motivation
For my use case, the default prompt sent to LLM to generate a query doesn't work. Although the question I am asking the LLM is good, it is not able to figure the correct filter it should use. I think if I was able to modify the query, it could solve the issue and start returning correct queries.
### Your contribution
Not really. | Add option to specify prompt for SelfQueryRetriever | https://api.github.com/repos/langchain-ai/langchain/issues/11735/comments | 5 | 2023-10-12T21:49:59Z | 2024-05-07T16:05:58Z | https://github.com/langchain-ai/langchain/issues/11735 | 1,940,842,588 | 11,735 |
[
"langchain-ai",
"langchain"
] | ### System Info
Langchain version: 0.0.312
Python: 3.10.11
Platform: Ubuntu
### Who can help?
@hwchase17 @agola11
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [X] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```
from langchain.memory import ConversationTokenBufferMemory
max_token_limit = 0
memory = ConversationTokenBufferMemory(
llm=llm, # Azure OpenAI GPT3.5
max_token_limit=max_token_limit,
return_messages=True,
)
context_pair = {
"inputs": {"input": "Example input message."},
"outputs": {"outputs": "Example output message."}
}
memory.save_context(**context_pair)
```
### Expected behavior
The above code produces an `IndexError: pop from empty list`
Due to some token budgeting we're doing, sometimes the space for the memory gets reduced to 0 (specifically when a user has a long message, we prioritize tokens for their message and the AI response). So we create memory with `max_token_limit=0`, and **ideally we'd like the memory to be pruned such that the buffer is empty.** However, we instead get the above error.
### Where the bug seems to be
The culprit seems to be in `memory.save_context()` , which we call directly in some cases or sometimes it happens internally like:
```
chat_chain = ConversationChain(
memory=memory, llm=llm, prompt=prompt_template
)
ai_response = chat_chain.predict(input=user_message) # internal call to save_context()
```
Within `save_context`, there's a loop:
```
while curr_buffer_length > self.max_token_limit:
pruned_memory.append(buffer.pop(0))
curr_buffer_length = self.llm.get_num_tokens_from_messages(buffer)
```
Oddly, when the buffer is an empty list, `curr_buffer_length` is 3 when it should be 0. So the loop continues and tries to pop from the empty buffer list and we get the above error.
| IndexError: pop from empty list - bug in ConversationTokenBufferMemory | https://api.github.com/repos/langchain-ai/langchain/issues/11734/comments | 6 | 2023-10-12T21:11:56Z | 2024-02-11T16:11:42Z | https://github.com/langchain-ai/langchain/issues/11734 | 1,940,800,530 | 11,734 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I am following the notebook in this github to create two agents with tools:
https://github.com/langchain-ai/langchain/blob/master/cookbook/two_agent_debate_tools.ipynb
I am also using AzureOpenAI instead of OpenAI directly because of corporate firewalls. I am getting this error:
TypeError Traceback (most recent call last)
Cell In[25], line 34
24 agent_description = AzureChatOpenAI(
25 openai_api_base=BASE_URL,
26 openai_api_version="2023-05-15",
(...)
29 openai_api_type="azure",
30 )(temperature=1.0)(agent_specifier_prompt).content
31 return agent_description
---> 34 agent_descriptions = {name: generate_agent_description(name) for name in names}
Cell In[25], line 34, in (.0)
24 agent_description = AzureChatOpenAI(
25 openai_api_base=BASE_URL,
26 openai_api_version="2023-05-15",
(...)
29 openai_api_type="azure",
30 )(temperature=1.0)(agent_specifier_prompt).content
31 return agent_description
---> 34 agent_descriptions = {name: generate_agent_description(name) for name in names}
Cell In[25], line 24, in generate_agent_description(name)
13 def generate_agent_description(name):
14 agent_specifier_prompt = [
...
29 openai_api_type="azure",
30 )(temperature=1.0)(agent_specifier_prompt).content
31 return agent_description
TypeError: BaseChatModel.__call__() missing 1 required positional argument: 'messages'
### Suggestion:
_No response_ | TypeError: BaseChatModel.__call__() missing 1 required positional argument: 'messages' | https://api.github.com/repos/langchain-ai/langchain/issues/11733/comments | 3 | 2023-10-12T20:45:25Z | 2024-07-14T15:27:49Z | https://github.com/langchain-ai/langchain/issues/11733 | 1,940,767,532 | 11,733 |
[
"langchain-ai",
"langchain"
] | ### Feature request
Hello everyone!
Please when will we have SQLDatabaseChain async?
I believe it's crucial for all Chains to be async so that we can have a serverless service serving multiple clients.
Thank you!
### Motivation
I believe it's crucial for all Chains to be async so that we can have a serverless service serving multiple clients.
### Your contribution
I dont know how to contribute. Maybe I coult test your code. | async SQLDatabaseChain | https://api.github.com/repos/langchain-ai/langchain/issues/11732/comments | 5 | 2023-10-12T20:25:16Z | 2024-03-28T16:06:32Z | https://github.com/langchain-ai/langchain/issues/11732 | 1,940,740,863 | 11,732 |
[
"langchain-ai",
"langchain"
] | ### Feature request
The proposal is to enhance our Langchain by integrating support for querying The Movie Database (TMDb) using a Tool to retrieve detailed information about movies.
[https://developer.themoviedb.org/docs/getting-started](url)
[https://github.com/celiao/tmdbsimple/](url)
### Motivation
Allowing users to query The Movie Database (TMDb) would allow users to ask information regarding moives, and this tool will give the llms better/more reliable information to work with.
### Your contribution
We intend to submit a pull request for this issue at some point in November for a school project. | Add support for themoviedb querying to get movie info | https://api.github.com/repos/langchain-ai/langchain/issues/11728/comments | 3 | 2023-10-12T19:15:07Z | 2023-10-16T04:21:47Z | https://github.com/langchain-ai/langchain/issues/11728 | 1,940,635,026 | 11,728 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
The retriever object in the VectorStoreRetrieverMemory class is a VectorStoreRetriever and not AzureSearchVectorStoreRetriever.
The VectorStoreRetriever and AzureSearchVectorStoreRetriever are two different classes and doesn't support same methods.
As i want to use vector store backed memory retriever which has option for filtering with user_id metadata to maintain mutliple chat histories based on user. I require help in two things.
1. Support for AzureSearchVectorStoreRetriever object in VectorStoreRetrieverMemory
2. How to initialize the memory object that accepts azure vector store (with id, user_id, conversation, conversation_vector as fields in my case), I want to save_context with respect to user_id, the following example given in the
[](https://python.langchain.com/docs/modules/memory/types/vectorstore_retriever_memory)
memory.save_context({"input": "My favorite food is pizza"}, {"output": "that's good to know"}) doesn't say how.
similarly for the memory.load_memory_variables({"prompt": "what sport should i watch?"})["history"]
### Suggestion:
As most organizations use cases would be at scale, examples that include how to maintain memory for different users (user_id being metadata) would help.
Add AzureSearchVectorStoreRetriever object to the VectorStoreRetrieverMemory class would also help.
If i am wrong in the above cases, please correct me and help me pointing out the resources to solve this. | Issue: <VectorStoreRetrieverMemory class doesn't support AzureSearchVectorStoreRetriever retriever> | https://api.github.com/repos/langchain-ai/langchain/issues/11720/comments | 2 | 2023-10-12T16:49:35Z | 2024-02-06T16:22:36Z | https://github.com/langchain-ai/langchain/issues/11720 | 1,940,398,645 | 11,720 |
[
"langchain-ai",
"langchain"
] | ### System Info
LangChain 0.0.240
but code on latest master has same issue
### Who can help?
@agola11
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [X] Callbacks/Tracing
- [ ] Async
### Reproduction
Use a fine-tuned OpenAI model, named like "ft:gpt-2.5-turbo-0613:orgname@12abCDEF"
`with get_openai_callback() as openai_callback:`
`openai_callback.total_cost`
The cost is zero.
### Expected behavior
The cost is returned.
The issue is in https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/callbacks/openai_info.py#L67
which expects fine-tuned models' names' to start with "ft-" (dash, should be colon) | Cost calculation is always zero with fine-tuned OpenAI models | https://api.github.com/repos/langchain-ai/langchain/issues/11715/comments | 5 | 2023-10-12T14:46:31Z | 2024-02-04T19:41:20Z | https://github.com/langchain-ai/langchain/issues/11715 | 1,940,160,863 | 11,715 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
Having issue with cypherPrompt in GraphCypherQAChain .
I am trying to use GraphCypherQAChain and Bedrock to generate an answer from Neo4j database.
I am using local Neo4j Movies database and the code example from: https://js.langchain.com/docs/modules/chains/additional/cypher_chain
Here is the code:
```
const { PromptTemplate } = require("langchain/prompts");
const { GraphCypherQAChain } = require("langchain/chains/graph_qa/cypher");
const { Neo4jGraph } = require("langchain/graphs/neo4j_graph");
const { Bedrock } = require("langchain/llms/bedrock");
const url = "bolt://localhost:7687";
const username = "neo4j";
const password = "***";
const aiModelId = "ai21.j2-mid-v1";
module.exports.chainRun = async (event) => {
const graph = await Neo4jGraph.initialize({ url, username, password });
const model = new Bedrock({
model: aiModelId,
region: "***",
maxTokens: 200,
temperature: 0.5,
stopSequences: ["\n\nHuman:"],
credentials: {
accessKeyId: "***",
secretAccessKey: "***",
},
});
await graph.query(
"CREATE (a:Actor {name:'Bruce Willis'})" +
"-[:ACTED_IN]->(:Movie {title: 'Pulp Fiction'})"
);
const cypherTemplate = `Task:Generate Cypher statement to query a graph database.
Instructions:
Use only the provided relationship types and properties in the schema.
Do not use any other relationship types or properties that are not provided.
Schema:
{schema}
Note: Do not include any explanations or apologies in your responses.
Do not respond to any questions that might ask anything else than for you to construct a Cypher statement.
Do not include any text except the generated Cypher statement.
Follow these Cypher example when Generating Cypher statements:
# How many actors played in Top Gun?
MATCH (m:Movie {title:"Top Gun"})<-[:ACTED_IN]-()
RETURN count(*) AS result
The question is:
{question}`;
const cypherPrompt = new PromptTemplate({
template: cypherTemplate,
inputVariables: ["schema", "question"],
});
const chain = GraphCypherQAChain.fromLLM({
llm: model,
graph,
cypherPrompt,
});
const res = await chain.run("Who played in Pulp Fiction?");
return {
statusCode: 200,
body: JSON.stringify(res),
};
};
```
The error that I am getting:
<img width="926" alt="Screenshot 2023-10-12 at 2 08 52 PM" src="https://github.com/langchain-ai/langchain/assets/42175061/3f40f1e1-09fa-4ebf-ae29-35f6dc28e0b3">
If I run that query directly it returns result, also if I remove that part from cypher prompt there are no issues.
Can you help me to find what is the issue?
### Suggestion:
_No response_ | Issue: Invalid prompt schema: Missing value for input [GraphCypherQAChain][cypherPrompt] | https://api.github.com/repos/langchain-ai/langchain/issues/11706/comments | 2 | 2023-10-12T12:12:12Z | 2023-10-16T11:55:48Z | https://github.com/langchain-ai/langchain/issues/11706 | 1,939,853,669 | 11,706 |
[
"langchain-ai",
"langchain"
] | ### Feature request
I propose a feature that allows the **API key** to be set as an optional **parameter** when changing the `openai_api_base` in LangChain's implementation of OpenAI. This feature would make it possible to set the API key to an empty string or any arbitrary string, similar to how it's done in **LiteLLM**.
Relevant resources:
- LangChain OpenAI: [GitHub Repo](https://github.com/langchain-ai/langchain)
- LiteLLM Proxy: [Documentation](https://docs.litellm.ai/docs/proxy_server#create-a-proxy-for-multiple-llms)
### Motivation
The motivation for this proposal is to facilitate testing and development scenarios where the `openai_api_base` is pointed to a local or mock server that doesn't require an API key. Currently, when the API key is set to empty or an arbitrary string, LangChain OpenAI throws an `AuthenticationError`. This behavior is frustrating when trying to test or develop without a valid OpenAI API key.
### Your contribution
I am willing to help implement this feature and submit a PR, but I would appreciate guidance from the maintainers or community to ensure the changes are made correctly and in line with the project's standards and practices. | Allow API_KEY to be optional when API_BASE is selected for OpenAI. | https://api.github.com/repos/langchain-ai/langchain/issues/11700/comments | 4 | 2023-10-12T10:47:50Z | 2024-02-10T16:13:37Z | https://github.com/langchain-ai/langchain/issues/11700 | 1,939,721,293 | 11,700 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
如何获取打印的日志呢？不知道怎么关闭这些打印的日志
### Suggestion:
_No response_ | Agent stopped due to iteration limit or time limit. | https://api.github.com/repos/langchain-ai/langchain/issues/11699/comments | 5 | 2023-10-12T09:38:36Z | 2024-02-10T16:13:42Z | https://github.com/langchain-ai/langchain/issues/11699 | 1,939,606,027 | 11,699 |
[
"langchain-ai",
"langchain"
] | ### System Info
Current master, all python versions
### Who can help?
@hwchase17
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [X] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
The LocalAI embeddings class required the openai api key to be set even though this might not be required by the locally hosted server: https://github.com/langchain-ai/langchain/blob/44da27c07b2bd0ccac355c8236a3ab1dd26870eb/libs/langchain/langchain/embeddings/localai.py#L200
Initializing the class like this fails:
```
from langchain.embeddings.localai import LocalAIEmbeddings
LocalAIEmbeddings(openai_api_key=None)
# Did not find openai_api_key, please add an environment variable `OPENAI_API_KEY` which contains it, or pass
`openai_api_key` as a named parameter. (type=value_error)
```
### Expected behavior
it should work fine, the key is not required for the class to function. | LocalAI embeddings require OpenAI key | https://api.github.com/repos/langchain-ai/langchain/issues/11698/comments | 1 | 2023-10-12T08:58:04Z | 2024-02-06T16:22:51Z | https://github.com/langchain-ai/langchain/issues/11698 | 1,939,531,951 | 11,698 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
Dear Langchain Developers,
Thank you very much for developing Langchain.
I am writing to seek clarification on a few aspects of Langchain that I find intriguing. Specifically, I am interested in understanding more about the role and functionality of 'metadata' and 'page_content' within the system.
Could you kindly explain what the purpose of these two components is within Langchain? I am curious to know how they contribute to the overall functionality of the system. Additionally, I am interested in understanding how these elements are stored in vectors.
Furthermore, when a user question is used to match the most suitable text chunk, how does the system process this? After the user question is converted into a vector, does the system search the 'metadata' or the 'page_content'?
Your insights would be greatly appreciated. I am sure that understanding these aspects will enhance my appreciation of Langchain even more. Thank you in advance for your time and assistance.
Best regards,
### Suggestion:
_No response_ | Understanding the Role and Functionality of Metadata and Page_Content in Langchain | https://api.github.com/repos/langchain-ai/langchain/issues/11697/comments | 2 | 2023-10-12T08:31:39Z | 2024-02-09T16:16:58Z | https://github.com/langchain-ai/langchain/issues/11697 | 1,939,487,505 | 11,697 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
How to get the same values of the Float numbers generated as embeddings -
1/ Am comparing values generated from OpenAI - from langchain.embeddings.openai import OpenAIEmbeddings
embedding = OpenAIEmbeddings()
2/ comparing with the values generated from -- gpt4all
from langchain.embeddings import GPT4AllEmbeddings
gpt4all_embd = GPT4AllEmbeddings()
```python
def get_embeddings_gpt4all():
"""
"""
from langchain.embeddings import GPT4AllEmbeddings
gpt4all_embd = GPT4AllEmbeddings()
#text = "This is a test document."
text = "i like dogs"
query_result = gpt4all_embd.embed_query(text)
print("--query_result-",type(query_result))
print("--query_result-",len(query_result))
print("--query_result-",query_result[0:10])
print("--query_result-",query_result[11])
print("--query_result-",query_result[15])
doc_result = gpt4all_embd.embed_documents([text])
print("--doc_result-",type(doc_result)) # List of LIST --
print("--doc_result-",len(doc_result)) # 1
print("--doc_result-",type(doc_result[0])) # List of Floats --
print("--doc_result-",len(doc_result[0])) # 384
```
- terminal prints as below --
```
Found model file at C:\\\\Users\\\\rohit.dhankar\\\\.cache\\\\gpt4all\\ggml-all-MiniLM-L6-v2-f16.bin
--query_result- <class 'list'>
--query_result- 384
--query_result- [-0.04856949672102928, -0.04622294753789902, 0.060556311160326004, 0.043546244502067566, -0.05463598296046257, -0.004457559902220964, 0.05100328475236893, -0.02644169144332409, 0.08410052955150604, 0.05480821430683136]
--query_result- -0.07515878230333328
--query_result- 0.05370383337140083
--doc_result- <class 'list'>
--doc_result- 1
--doc_result- <class 'list'>
--doc_result- 384
```
### Suggestion:
_No response_ | Issue: GPT4AllEmbeddings - dont compare with - OpenAIEmbeddings | https://api.github.com/repos/langchain-ai/langchain/issues/11696/comments | 2 | 2023-10-12T07:12:54Z | 2024-02-09T16:17:03Z | https://github.com/langchain-ai/langchain/issues/11696 | 1,939,352,699 | 11,696 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
_No response_
### Suggestion:
_No response_ | agent是否支持llama2系列及其衍生模型 | https://api.github.com/repos/langchain-ai/langchain/issues/11695/comments | 2 | 2023-10-12T07:07:11Z | 2023-10-12T07:44:50Z | https://github.com/langchain-ai/langchain/issues/11695 | 1,939,343,638 | 11,695 |
[
"langchain-ai",
"langchain"
] | ### System Info
APIError: HTTP code 200 from API ("{\"rate_limit_usage\": {\)
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/streamlit/runtime/scriptrunner/script_runner.py", line 541, in _run_script
exec(code, module.__dict__)
File "[...]/gitlab/run_finQA_webUI.py", line 158, in <module>
generate_answer(clicked)
File "[...]/gitlab/run_finQA_webUI.py", line 123, in generate_answer
response = rq_agent.finQAinConversation(user_query, callbacks=[stream_handler])
File "[...]/gitlab/report_query/report_query.py", line 187, in finQAinConversation
response = self._convR_qa(query, callbacks=callbacks)
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/base.py", line 306, in __call__
raise e
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/base.py", line 300, in __call__
self._call(inputs, run_manager=run_manager)
File "[...]/gitlab/report_query/custom_intention_chain.py", line 131, in _call
answer = self.combine_docs_chain.run(
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/base.py", line 506, in run
return self(kwargs, callbacks=callbacks, tags=tags, metadata=metadata)[
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/base.py", line 306, in __call__
raise e
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/base.py", line 300, in __call__
self._call(inputs, run_manager=run_manager)
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/combine_documents/base.py", line 119, in _call
output, extra_return_dict = self.combine_docs(
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/combine_documents/stuff.py", line 171, in combine_docs
return self.llm_chain.predict(callbacks=callbacks, **inputs), {}
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/llm.py", line 257, in predict
return self(kwargs, callbacks=callbacks)[self.output_key]
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/base.py", line 306, in __call__
raise e
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/base.py", line 300, in __call__
self._call(inputs, run_manager=run_manager)
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/llm.py", line 93, in _call
response = self.generate([inputs], run_manager=run_manager)
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chains/llm.py", line 103, in generate
return self.llm.generate_prompt(
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chat_models/base.py", line 469, in generate_prompt
return self.generate(prompt_messages, stop=stop, callbacks=callbacks, **kwargs)
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chat_models/base.py", line 359, in generate
raise e
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chat_models/base.py", line 349, in generate
self._generate_with_cache(
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chat_models/base.py", line 501, in _generate_with_cache
return self._generate(
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chat_models/openai.py", line 357, in _generate
return _generate_from_stream(stream_iter)
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chat_models/base.py", line 60, in _generate_from_stream
for chunk in stream:
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/langchain/chat_models/openai.py", line 326, in _stream
for chunk in self.completion_with_retry(
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/openai/api_resources/abstract/engine_api_resource.py", line 166, in <genexpr>
return (
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/openai/api_requestor.py", line 693, in <genexpr>
self._interpret_response_line(
File "[...]/anaconda3/envs/sjj_env/lib/python3.10/site-packages/openai/api_requestor.py", line 755, in _interpret_response_line
raise error.APIError(
Seems like this error just appears starting from oct 12. The answer can be directly generated with an HTTP code 200 in Openai response, but an error appears at the end of generation. @hwchase17
### Who can help?
_No response_
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
answer = self.combine_docs_chain.run(
input_documents=docs, callbacks=_run_manager.get_child(), **new_inputs
)
### Expected behavior
No error happens. | APIError: HTTP code 200 from API ("{\"rate_limit_usage\": {\) | https://api.github.com/repos/langchain-ai/langchain/issues/11694/comments | 7 | 2023-10-12T04:24:29Z | 2023-10-12T07:43:45Z | https://github.com/langchain-ai/langchain/issues/11694 | 1,939,149,824 | 11,694 |
[
"langchain-ai",
"langchain"
] | ### Feature request
Would it make sense to use an llm to split a large document into smaller documents that have complete context?
### Motivation
My best guess is that most of the existing splitters may drop context ?
### Your contribution
Could try | text split by llm? | https://api.github.com/repos/langchain-ai/langchain/issues/11691/comments | 5 | 2023-10-12T01:29:23Z | 2024-02-26T16:07:38Z | https://github.com/langchain-ai/langchain/issues/11691 | 1,939,011,196 | 11,691 |
[
"langchain-ai",
"langchain"
] | ### Feature request
At the moment the agent spends additional rounds of agent <> LLM to format the JSON data I request to be in a sentence format. Is there a way to avoid this?
### Motivation
If the user is asking for JSON the agent should provide JSON and not a properly formatted sentence(that costs more token>money)
### Your contribution
Test, Deploy to production workload | A way to use csv_agent to simply return JSON | https://api.github.com/repos/langchain-ai/langchain/issues/11683/comments | 1 | 2023-10-11T21:31:49Z | 2024-01-30T00:26:49Z | https://github.com/langchain-ai/langchain/issues/11683 | 1,938,775,462 | 11,683 |
[
"langchain-ai",
"langchain"
] | ### System Info
Version `0.0.312`
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [X] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
# Description
The parameter `include_df_in_prompt` on the method `create_pandas_dataframe_agent` is a boolean value which should be eigther `True` or `False`.
If you specify your own suffix in the create_pandas_dataframe_agent method it results in an error saying that the variable include_df_in_prompt should be False.
## For example
```python
agent = create_pandas_dataframe_agent(ChatOpenAI(temperature=0),df, AgentType.OPENAI_FUNCTIONS, suffix=sufix)
```
Results in
```
ValueError: If suffix is specified, include_df_in_prompt should not be.
```
This is also the correct behavior. However, if you set `include_df_in_prompt` to `False`, the error still occurs.
## For example
```python
agent = create_pandas_dataframe_agent(ChatOpenAI(temperature=0),df, AgentType.OPENAI_FUNCTIONS, suffix=sufix, include_df_in_prompt=False)
```
But everything just works fine if I use `None` instade of `False`
## For example
```python
agent = create_pandas_dataframe_agent(ChatOpenAI(temperature=0),df, AgentType.OPENAI_FUNCTIONS, suffix=sufix, include_df_in_prompt=None)
``
### Expected behavior
It should also work when I use `False` instade of `None` | include_df_in_prompt should be compared as boolean value and not None | https://api.github.com/repos/langchain-ai/langchain/issues/11677/comments | 2 | 2023-10-11T20:47:09Z | 2024-06-28T16:30:12Z | https://github.com/langchain-ai/langchain/issues/11677 | 1,938,701,583 | 11,677 |
[
"langchain-ai",
"langchain"
] | ### Feature request
Similar to the discussion in this thread https://github.com/langchain-ai/langchain/issues/7239, I would like to have a way to show inline citations in QA tasks.
Specially, I haven't had a whole lot of success integrating `create_citation_fuzzy_match_chain` with `ConversationalRetrievalChain`
### Motivation
The point of this change is to enable QA based chains (e.g. ConversationalRetrievalChain) to easily do inline citations, and surface the citations as part of metadata fields
### Your contribution
TBD | Integrating `create_citation_fuzzy_match_chain` with `ConversationalRetrievalChain` | https://api.github.com/repos/langchain-ai/langchain/issues/11673/comments | 18 | 2023-10-11T19:24:42Z | 2023-12-15T10:43:03Z | https://github.com/langchain-ai/langchain/issues/11673 | 1,938,566,368 | 11,673 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
Similar to the discussion in this thread https://github.com/langchain-ai/langchain/issues/7239, I would like to have a way to show inline citations in QA tasks.
Specially, I haven't had a whole lot of success integrating `create_citation_fuzzy_match_chain` with `ConversationalRetrievalChain`
### Suggestion:
_No response_ | DELETE | https://api.github.com/repos/langchain-ai/langchain/issues/11671/comments | 0 | 2023-10-11T19:16:40Z | 2023-10-24T01:07:38Z | https://github.com/langchain-ai/langchain/issues/11671 | 1,938,552,713 | 11,671 |
[
"langchain-ai",
"langchain"
] | ### Feature request
Currently, the Bedrock `chat_model` does not support async streaming, which is odd, as the previous [PR](https://github.com/langchain-ai/langchain/pull/10393/files#diff-9874347f7fa335df661ff4089b0922b3214e08a92e9879610424522f806358f7R62) included it; seems that the async method was not added properly.
### Motivation
Without async streaming you cannot pretty much build chatbot infra.
### Your contribution
Not sure, possibly just copying-pasting what was [here](https://github.com/langchain-ai/langchain/pull/10393/files#diff-9874347f7fa335df661ff4089b0922b3214e08a92e9879610424522f806358f7R62) already? | Add async streaming for Bedrock | https://api.github.com/repos/langchain-ai/langchain/issues/11668/comments | 6 | 2023-10-11T16:06:06Z | 2024-06-18T08:41:48Z | https://github.com/langchain-ai/langchain/issues/11668 | 1,938,174,193 | 11,668 |
[
"langchain-ai",
"langchain"
] | ### System Info
on Ubuntu with Python 3.11.5, Langchain 0.0.307
### Who can help?
@hwchase17
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [X] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
What my code basically does is returns an answer to a given question using conversational retrieval chain with a qdrant vectorstore that stores publications. I want the chain to use only the documents published in specified years.
1. Create vectorstore
qdrant_client = QdrantClient(QDRANT_HOST, port=int(QDRANT_PORT))
retriever = Qdrant(client=qdrant_client, collection_name=QDRANT_COLLECTION, embeddings=get_encoder(ENCODER))
2. Create chain:
rag_chain = ConversationalRetrievalChain.from_llm(llm=llm,
retriever = retriever.as_retriever(search_kwargs={"filter": {"year": ["2009","2010"]}),
return_source_documents=return_source_documents,
verbose=verbose)
3. When I run the chain it retrieves no documents even though it should.
### Expected behavior
The chain should return documents related to a given question. When I checked the implementation of qdrant.py under langchain vectorstores I saw that _qdrant_filter_from_dict function makes a "must" match; however in my case I need a "should" match since no document can have more than one publication date. | ConversationalRetrievalChain Vectorstore cannot apply filter to a key with given multiple values | https://api.github.com/repos/langchain-ai/langchain/issues/11666/comments | 5 | 2023-10-11T15:29:24Z | 2024-02-12T16:11:33Z | https://github.com/langchain-ai/langchain/issues/11666 | 1,938,100,043 | 11,666 |
[
"langchain-ai",
"langchain"
] | ### System Info
langchain==0.0.306
chromadb==0.4.13
Python 3.11.5
### Who can help?
@hwchase17 @atroyn
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [X] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```python
from langchain.vectorstores import Chroma
from langchain.docstore.document import Document
from langchain.embeddings import GPT4AllEmbeddings
docs = [
Document(page_content="This is a demo document"),
Document(page_content="This is another demo document"),
]
other_docs = [
Document(page_content="Unrelated document"),
Document(page_content="Another unrelated document"),
]
# Create two (independent) vector stores
vectorstore = Chroma.from_documents(documents=docs, embedding=GPT4AllEmbeddings())
other_vectorstore = Chroma.from_documents(documents=other_docs, embedding=GPT4AllEmbeddings())
# Expect two stores to be independent
assert len(vectorstore.similarity_search("document", k=999)) == len(docs) # OK
assert len(other_vectorstore.similarity_search("document", k=999)) == len(other_docs) # FAILS
# Documents from the first vector store are actually retrieved in the `other_vectorstore`
print(other_vectorstore.similarity_search("document"))
# Actually the two vector stores contain the union of `docs` and `other_docs`
assert len(vectorstore.get()["documents"]) == 2 # FAILS
assert len(other_vectorstore.get()["documents"]) == 2 # FAILS
```
### Expected behavior
When I initialize two separate Chroma in-memory vectorstores with `Chroma.from_documents` using two distinct groups of documents, the vector store that was initialized last will contain the union of the two groups of documents.
In summary, running this code:
```python
vectorstore = Chroma.from_documents(documents=documents, …)
other_vectorstore = Chroma.from_documents(documents=other_documents, …)
```
the `other_vectorstore` will contain both `documents` and `other_documents`.
---
It seems that the two vectorstores are sharing the same chroma client & collection.
This could be a side effect of commit https://github.com/langchain-ai/langchain/commit/822cdb161b6631d8255e341f3321d8f0b9e28e8a which made the `chromadb.Client` shared among all instances of the `Chroma` class.
One can avoid the bug by explicitly passing two different `collection_name` values when initializing the vector store, but I thought this would have been the default behaviour (or at least I expected a warning telling that the new vectorstore was reusing an existing collection).
Moreover, it is not clear to me why the first vector store will return the documents it was initialized with if I use `vectorstore.similarity_search(…)`, but shows the union of all documents if I use `vectorstore.get()`.
If you confirm what is the intended behaviour I am available to to work a PR to fix this. | `Chroma.from_documents` adds docs to same collection creating duplicates | https://api.github.com/repos/langchain-ai/langchain/issues/11665/comments | 5 | 2023-10-11T14:05:47Z | 2023-10-11T20:53:10Z | https://github.com/langchain-ai/langchain/issues/11665 | 1,937,873,604 | 11,665 |
[
"langchain-ai",
"langchain"
] | ### System Info
0.0.3
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
https://api.python.langchain.com/en/latest/vectorstores/langchain.vectorstores.faiss.FAISS.html
### Expected behavior
get concrete docs's index or content | db = FAISS.from_documents(docs, embeddings) ;how to get docs's index or content from db? | https://api.github.com/repos/langchain-ai/langchain/issues/11661/comments | 6 | 2023-10-11T13:20:45Z | 2023-10-11T14:46:52Z | https://github.com/langchain-ai/langchain/issues/11661 | 1,937,769,538 | 11,661 |
[
"langchain-ai",
"langchain"
] | 
I am using the ConversationalRetrievalChain along with DirectoryLoader from LangChain to load PDF files only and answer questions based on the different files in the directory along with returning the source of the answer from the document it got it from. As shown in the 2 pictures, in the first picture I asked it to answer in 1 single sentence but it kept answering and answering then repeated the answer. In the second picture, it rewrote the question again and then answered along with repeating the answers. In the 3rd picture, it answered a question thats not in any of the files in the Directory and it then shows the error of exceeding the rate limit.
I'd like help on how can I make it "listen" to me and how do I fix the rate limit error? I am using OpenAI embeddings for my Q&A chatbot. Does anyone recommend other embeddings that are not OpenAI which I could use for this use case? I plan on later moving my embeddings to the cloud and deploy it.
Thanks in advance
| Chat model answering incorrectly and weirdly | https://api.github.com/repos/langchain-ai/langchain/issues/11657/comments | 3 | 2023-10-11T11:39:43Z | 2024-02-08T16:20:16Z | https://github.com/langchain-ai/langchain/issues/11657 | 1,937,562,333 | 11,657 |
[
"langchain-ai",
"langchain"
] | I am trying to use a prompt with my chain but I am getting an error which I'm not sure how to solve. Also, I was wondering is it possible to use a constitutional principle to apply it to the ConversationalRetrievalChain? If so, can someone guide me on how to do so? Sorry for beginner questions but I am new to all of this including coding. Thanks in advance!
Code:
from dotenv import load_dotenv
import csv
import PyPDF2
from PyPDF2 import PdfReader
from langchain.document_loaders import DirectoryLoader, PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.callbacks import get_openai_callback
from langchain.chat_models import ChatOpenAI
from langchain.chains import ConversationalRetrievalChain
from langchain.memory import ConversationBufferMemory
from langchain.prompts import PromptTemplate
import time
from langchain.chains.qa_with_sources import load_qa_with_sources_chain
from langchain.vectorstores import Chroma
from langchain.callbacks import StreamingStdOutCallbackHandler
import pandas as pd
from docx import Document
from nltk.tokenize import sent_tokenize, word_tokenize
from collections import Counter
from nltk.corpus import stopwords
import os
def print_letter_by_letter(text):
for char in text:
print(char, end='', flush=True)
time.sleep(0.02)
def main():
load_dotenv()
# directory_path = input("Copy your directory path here or upload a file: ")
directory_path = "C:\\Users\\Asus\\OneDrive\\Documents\\Vendolista"
pdf_loader = DirectoryLoader(directory_path,
glob="**/*.pdf",
show_progress=True,
use_multithreading=True,
silent_errors=True,
loader_cls = PyPDFLoader)
documents = pdf_loader.load()
print(str(len(documents))+ " documents loaded")
llm = ChatOpenAI(temperature = 0, model_name='gpt-3.5-turbo', callbacks=[StreamingStdOutCallbackHandler()], streaming = True)
# Split into chunks
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=1000,
chunk_overlap=200,
)
chunks = text_splitter.split_documents(documents)
embeddings = OpenAIEmbeddings()
persist_directory = "C:\\Users\\Asus\\OneDrive\\Documents\\Vendolista"
knowledge_base = Chroma.from_documents(chunks, embeddings, persist_directory = persist_directory)
#save to disk
knowledge_base.persist()
#To delete the DB we created at first so that we can be sure that we will load from disk as fresh db
knowledge_base = None
new_knowledge_base = Chroma(persist_directory = persist_directory, embedding_function = embeddings)
custom_template = ("""
Help the users with answering questions based on the document.
"""
)
# prompt = PromptTemplate(input_variables=[], template=custom_template)
prompt = PromptTemplate.from_template(custom_template)
memory = ConversationBufferMemory(memory_key = "chat_history", return_messages = True)
model = ConversationalRetrievalChain.from_llm(llm = llm,
chain_type = "refine",
retriever = new_knowledge_base.as_retriever(),
memory = memory,
combine_docs_chain_kwargs = {"prompt":prompt}
)
while True:
question = input("Ask me anything about the files (type 'exit' to quit): ")
if question.lower() in ["exit"] and len(question) == 4:
end_chat = "Thank you for visiting us! Have a nice day"
print_letter_by_letter(end_chat)
break
if question:
chat_history = []
# with get_openai_callback() as cb:
response = model({"question": question}, return_only_outputs = True)
chat_history.append(('user', question))
chat_history.append(('AI', response))
print("AI:", response)
if __name__ == '__main__':
main()
Error:
86 documents loaded
Traceback (most recent call last):
File "C:\Users\Asus\Documents\Vendolista\app.py", line 100, in <module>
main()
File "C:\Users\Asus\Documents\Vendolista\app.py", line 73, in main
model = ConversationalRetrievalChain.from_llm(llm = llm,
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\chains\conversational_retrieval\base.py", line 341, in from_llm
doc_chain = load_qa_chain(
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\chains\question_answering\__init__.py", line 249, in load_qa_chain
return loader_mapping[chain_type](
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\chains\question_answering\__init__.py", line 206, in _load_refine_chain
return RefineDocumentsChain(
File "C:\Users\Asus\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\load\serializable.py", line 97, in __init__
super().__init__(**kwargs)
File "pydantic\main.py", line 341, in pydantic.main.BaseModel.__init__
pydantic.error_wrappers.ValidationError: 1 validation error for RefineDocumentsChain
prompt
extra fields not permitted (type=value_error.extra) | Prompt invalid with ConversationalRetrievalChain | https://api.github.com/repos/langchain-ai/langchain/issues/11656/comments | 5 | 2023-10-11T11:29:12Z | 2024-03-13T19:57:58Z | https://github.com/langchain-ai/langchain/issues/11656 | 1,937,545,931 | 11,656 |
[
"langchain-ai",
"langchain"
] | ### Feature request
I want to make similar implementation of SQLAlchemyCache that can index prompts by md5 hash.
### Motivation
Postgres have limitation for 2704 bytes on index. Processing on simple index can be faster.
### Your contribution
I have prepared cache implementation – I will add it in contributing PR | Using hash for indexing prompts SQLAlchemyCache | https://api.github.com/repos/langchain-ai/langchain/issues/11655/comments | 3 | 2023-10-11T10:56:25Z | 2023-10-23T05:39:08Z | https://github.com/langchain-ai/langchain/issues/11655 | 1,937,471,577 | 11,655 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
In creating an app, ideally, you would like to let's say
db = SQLDatabase.from_uri(DATABASE_URI)
in one endpoint, and use the db in other endpoints,
The second endpoint could be what users get access to, so in getting responses, it should be quick, hence only accessing the already loaded db
but I cannot serialize db from the first endpoint, hence i have to include
db = SQLDatabase.from_uri(DATABASE_URI)
in the user endpoint, meaning anytime they hit the endpoint, db is run which delays significantly, and doesn't make for a good user experience, how do I solve this issue
### Suggestion:
_No response_ | Issue: Inability to serialize langchain modules such as conversational retrieval chain, SQL database | https://api.github.com/repos/langchain-ai/langchain/issues/11654/comments | 3 | 2023-10-11T10:35:10Z | 2024-02-08T16:20:20Z | https://github.com/langchain-ai/langchain/issues/11654 | 1,937,424,892 | 11,654 |
[
"langchain-ai",
"langchain"
] | ### System Info
IDE==PyCharm
Python==3.11.4
langchain==0.0.306
pinecone-client==2.2.2
### Who can help?
_No response_
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [X] Vector Stores / Retrievers
- [X] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [X] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
The following function in my code with **ConversationalRetrievalChain** runs perfectly. It responds to the question "Hello" appropriately. It answers the questions related to the data in the vector_store:
```python
def get_conversation_chain(vector_store):
llm = ChatOpenAI()
memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
conversation_chain = ConversationalRetrievalChain.from_llm(
llm=llm,
retriever=vector_store.as_retriever(),
memory=memory
)
return conversation_chain
```
When the chain is changed to **RetrievalQAWithSourcesChain** with new arguments then the bot responds to the user prompt "hello" with statements about Michael Jackson. It also does not answer questions to data that is contained within the vector store:
```python
def get_conversation_chain(vector_store):
llm = ChatOpenAI()
memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True, output_key="answer")
conversation_chain = RetrievalQAWithSourcesChain.from_chain_type(
llm=llm,
chain_type="stuff",
retriever=vector_store.as_retriever(),
# return_source_documents=True,
memory=memory
)
return conversation_chain
```
Everything else in my code remains the same when switching between the 2 functions. The issue only arises when I use the RetrievalQAWithSourcesChain function above. Also when the return_source_documents=True argument is active, it returns all documents within vector_store instead of only the documents being referenced in the answer.
### Expected behavior
I would expect that the answers from the RetrievalQAWithSourcesChain are the same as with ConversationalRetrievalChain:
- responds appropriately to the user prompt "hello"
- provides an answer referencing data that is contained within the vector_store
- plus returning relevant document sources related to answer. | Hallucinations, ignoring data in vector store and returning all documents as sources | https://api.github.com/repos/langchain-ai/langchain/issues/11653/comments | 2 | 2023-10-11T10:27:20Z | 2024-02-08T16:20:25Z | https://github.com/langchain-ai/langchain/issues/11653 | 1,937,406,691 | 11,653 |
[
"langchain-ai",
"langchain"
] | ### Issue with current documentation:
Hello folks,
I am not quite sure if my issue is an issue regarding the docs or rather the implementation.
### What is the differnece between HuggingFaceEmbeddings and HuggingFaceHubEmbeddings?
Currently in [v0.0.312](https://github.com/langchain-ai/langchain/releases/tag/v0.0.312), there are two pretty similar modules in the `langchain` package, namely `libs/langchain/langchain/embeddings/huggingface.py` and `libs/langchain/langchain/embeddings/huggingface_hub.py` (compare https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/embeddings/huggingface.py and https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/embeddings/huggingface_hub.py).
Moreover, these modules provide pretty similar classes: `HuggingFaceEmbeddings` and `HuggingFaceHubEmbeddings`.
When should I use which one? This question arises as a consequence of the described example below.
### Relevant Example
Considering the task of using a LLM with additional documents, you could consider the following code:
```
from langchain.llms.huggingface_hub import HuggingFaceHub
from langchain.prompts import PromptTemplate
from langchain.vectorstores.faiss import FAISS
from langchain.chains import LLMChain
from langchain.embeddings import HuggingFaceHubEmbeddings, HuggingFaceEmbeddings
from langchain.text_splitter import RecursiveCharacterTextSplitter
model = HuggingFaceHub(repo_id=llm, model_kwargs={"temperature": 0.2, "max_length": 64, "raw_response": True}, huggingfacehub_api_token=HUGGINGFACEHUB_API_KEY)
prompt = PromptTemplate(input_variables=["question", "docs"],
template="""
Answer the following question: {question}
By searching the following documents: {docs}
If you feel like you don't have enough information to answer the question, say "I don't know".
Your answers should be verbose and detailed.
""")
chain = LLMChain(llm=model, prompt=prompt)
docs = "bla bli blub ..."
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=5,
chunk_overlap=0
)
documents = text_splitter.split_text(document)
# A)
embeddings = HuggingFaceEmbeddings()
# B) alternatively
# embeddings = HuggingFaceHubEmbeddings(huggingfacehub_api_token=self.huggingfacehub_api_key)
self.documents_db = FAISS.from_texts(documents, embeddings)
docs = self.documents_db.similarity_search(prompt, k=4)
docs_page_content = " ".join([d.page_content for d in docs])
response = self.llm_chain.run(question=prompt, docs=docs_page_content)
response = response.replace("\n", "")
```
### Observed Behaviour
#### A)
Everything seems to work reliable and smooth without any errors or unexpected outputs.
#### B)
I did not conduct a study, but in my empirical testings I could observe the following.
Sometimes it works like A) and everything seems to work smooth and reliable,
However, sometimes (approximately 50% of the time i am running this script), I am getting one of the following two different errors:
##### KeyError
```
You're using a different task than the one specified in the repository. Be sure to know what you're doing :)
Traceback (most recent call last):
File "/qa_docs.py", line 59, in <module>
qa.add_document_to_context(document)
File "/qa.py", line 79, in add_document_to_context
self.documents_db = FAISS.from_texts(documents, embeddings)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/lib/python3.11/site-packages/langchain/vectorstores/faiss.py", line 603, in from_texts
return cls.__from(
^^^^^^^^^^^
File "/lib/python3.11/site-packages/langchain/vectorstores/faiss.py", line 562, in __from
index = faiss.IndexFlatL2(len(embeddings[0]))
~~~~~~~~~~^^^
KeyError: 0
```
##### ChunkedEncodingError
```
You're using a different task than the one specified in the repository. Be sure to know what you're doing :)
Traceback (most recent call last):
File "/lib/python3.11/site-packages/urllib3/response.py", line 710, in _error_catcher
yield
File "/lib/python3.11/site-packages/urllib3/response.py", line 1073, in read_chunked
self._update_chunk_length()
File "/lib/python3.11/site-packages/urllib3/response.py", line 1008, in _update_chunk_length
raise InvalidChunkLength(self, line) from None
urllib3.exceptions.InvalidChunkLength: InvalidChunkLength(got length b'', 0 bytes read)
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/lib/python3.11/site-packages/requests/models.py", line 816, in generate
yield from self.raw.stream(chunk_size, decode_content=True)
File "/lib/python3.11/site-packages/urllib3/response.py", line 933, in stream
yield from self.read_chunked(amt, decode_content=decode_content)
File "/lib/python3.11/site-packages/urllib3/response.py", line 1061, in read_chunked
with self._error_catcher():
File "/lib/python3.11/contextlib.py", line 155, in __exit__
self.gen.throw(typ, value, traceback)
File "/lib/python3.11/site-packages/urllib3/response.py", line 727, in _error_catcher
raise ProtocolError(f"Connection broken: {e!r}", e) from e
urllib3.exceptions.ProtocolError: ("Connection broken: InvalidChunkLength(got length b'', 0 bytes read)", InvalidChunkLength(got length b'', 0 bytes read))
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "qa_docs.py", line 59, in <module>
qa.add_document_to_context(document)
File "qa.py", line 79, in add_document_to_context
self.documents_db = FAISS.from_texts(documents, embeddings)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/lib/python3.11/site-packages/langchain/vectorstores/faiss.py", line 602, in from_texts
embeddings = embedding.embed_documents(texts)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/lib/python3.11/site-packages/langchain/embeddings/huggingface_hub.py", line 90, in embed_documents
responses = self.client(inputs=texts, params=_model_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/lib/python3.11/site-packages/huggingface_hub/inference_api.py", line 190, in __call__
response = get_session().post(self.api_url, headers=self.headers, json=payload, data=data)
File "/lib/python3.11/site-packages/requests/sessions.py", line 637, in post
return self.request("POST", url, data=data, json=json, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/lib/python3.11/site-packages/requests/sessions.py", line 589, in request
resp = self.send(prep, **send_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/lib/python3.11/site-packages/requests/sessions.py", line 747, in send
r.content
File "/lib/python3.11/site-packages/requests/models.py", line 899, in content
self._content = b"".join(self.iter_content(CONTENT_CHUNK_SIZE)) or b""
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/lib/python3.11/site-packages/requests/models.py", line 818, in generate
raise ChunkedEncodingError(e)
requests.exceptions.ChunkedEncodingError: ("Connection broken: InvalidChunkLength(got length b'', 0 bytes read)", InvalidChunkLength(got length b'', 0 bytes read))
```
### Remarks about the documents
Important note: the value of param `documents` is calculated by another script and is not fixed. It is a textual inference of another model. I already thought about that being the cause of the errors I encounter in some script executions but I have no idea how this could be. I can assure that there is always at least 100 characters that are being the value of `documents`.
However, I think that this problem already may be answered if I fully understand the differences between the two classes `HuggingFaceHubEmbeddings` and `HuggingFaceEmbeddings`.
### Related langchain documentation
- https://python.langchain.com/docs/integrations/vectorstores/faiss
- https://python.langchain.com/docs/integrations/providers/huggingface
- https://python.langchain.com/docs/integrations/llms/huggingface_hub.html
- https://python.langchain.com/docs/integrations/text_embedding/huggingfacehub.html
If there is more information needed, feel free to ask. I am looking forward for every help I can get!
### Idea or request for content:
When to use HuggingFaceEmbeddings vs. when to use HuggingFaceHubEmbeddings. | DOC: HuggingFaceEmbeddings vs. HuggingFaceHubEmbeddings | https://api.github.com/repos/langchain-ai/langchain/issues/11651/comments | 4 | 2023-10-11T09:37:15Z | 2024-02-11T16:11:51Z | https://github.com/langchain-ai/langchain/issues/11651 | 1,937,320,086 | 11,651 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I've tried building a Bot and now I see the following issue.
Architecture:
1. Cloud Run service 1 ; Streamlit app which accepts the user_input (question) on the topic and sends it to Flask API part of service 2.
2. Cloud Run service 2: Flask code which calls the LLM code and returns the response to streamlit.
I'm using qa as below.
qa = ConversationalRetrievalChain.from_llm(
llm,
retriever,
condense_question_prompt=CONDENSE_QUESTION_PROMPT,
verbose=False,
memory=memory,
condense_question_llm =llm,
return_generated_question= True,
combine_docs_chain_kwargs={"prompt": promptHist},
return_source_documents=True,
)
Issue: If different Users are asking questions through various streamlit sessions, it is using the same chat_history buffer and mixing responses.
Question: Is there a way to separate the "qa" sessions specific to User and avoid common memory/chat buffer being used ? Is there a utility in Langchain for this?
### Suggestion:
_No response_ | ConversationalRetrievalChain + Memory ( New issue for me) | https://api.github.com/repos/langchain-ai/langchain/issues/11650/comments | 9 | 2023-10-11T08:41:59Z | 2024-02-13T16:10:52Z | https://github.com/langchain-ai/langchain/issues/11650 | 1,937,199,873 | 11,650 |
[
"langchain-ai",
"langchain"
] | ### System Info
Langchain version : 0.0.312
python version: 3.11.4
conda env platform: osx-arm64 / x86
system : macos M1 13.2.1 (22D68)
### Who can help?
@hwchase17 @agola11
### Information
- [x] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```python
template = """Question: {question}
Answer: Let's think step by step."""
prompt = PromptTemplate(template=template, input_variables=["question"])
local_path = ("/Users/langchain-usecase/04-gpt4all/llama-2-7b-chat.ggmlv3.q4_0.bin")
# Callbacks support token-wise streaming
callbacks = [StreamingStdOutCallbackHandler()]
# Verbose is required to pass to the callback manager
llm = GPT4All(model=local_path, callbacks=callbacks, verbose=True)
llm_chain = LLMChain(prompt=prompt, llm=llm)
question = "What NFL team won the Super Bowl in the year Justin Bieber was born?"
llm_chain.run(question)
```
bash output
``` bash
Traceback (most recent call last):
File "/Users/mac/EasyGithub/test/gpt4all.py", line 45, in <module>
llm = GPT4All(model=local_path, callbacks=callbacks, verbose=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/mac/miniforge3/envs/langchain/lib/python3.11/site-packages/langchain/load/serializable.py", line 97, in __init__
super().__init__(**kwargs)
File "/Users/mac/miniforge3/envs/langchain/lib/python3.11/site-packages/pydantic/v1/main.py", line 341, in __init__
raise validation_error
pydantic.v1.error_wrappers.ValidationError: 1 validation error for GPT4All
__root__ -> __root__
Serializable.__init__() takes 1 positional argument but 2 were given (type=type_error)
```
### Expected behavior
The same code runs on my other M2 MacMini and generates a reply. I don't know what caused the error.
M2 MacMini env info:
python==3.11.4
langchain==0.0.312
conda env platform: osx-arm64
macos==13.4.1 | GPT4All Model Serializable.__init__ Error | https://api.github.com/repos/langchain-ai/langchain/issues/11649/comments | 1 | 2023-10-11T07:41:44Z | 2023-10-11T07:53:52Z | https://github.com/langchain-ai/langchain/issues/11649 | 1,937,072,756 | 11,649 |
[
"langchain-ai",
"langchain"
] | ### System Info
langchain=0.0.312
### code:
from langchain.llms import Minimax
minimax = Minimax(minimax_api_key=minimax_api_key, minimax_group_id=minimax_group_id)
print(minimax("What is the difference between panda and bear?"))
### or:
from langchain.chat_models import MiniMaxChat
from langchain.schema import HumanMessage
chat = MiniMaxChat()
res = chat(
[
HumanMessage(
content="Translate this sentence from English to French. I love programming."
)
]
)
print(res)
### error log:
File "E:\AI\Projects\AIProject\AIApiTest\langchain_minimax.py", line 43, in <module>
llm = Minimax()
File "D:\anaconda3\envs\ai\lib\site-packages\langchain\llms\minimax.py", line 113, in __init__
self._client = _MinimaxEndpointClient(
File "pydantic\main.py", line 357, in pydantic.main.BaseModel.__setattr__
ValueError: "Minimax" object has no field "_client"
### Who can help?
_No response_
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [X] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
### code:
from langchain.llms import Minimax
minimax = Minimax(minimax_api_key=minimax_api_key, minimax_group_id=minimax_group_id)
print(minimax("What is the difference between panda and bear?"))
### or:
from langchain.chat_models import MiniMaxChat
from langchain.schema import HumanMessage
chat = MiniMaxChat()
res = chat(
[
HumanMessage(
content="Translate this sentence from English to French. I love programming."
)
]
)
print(res)
### Expected behavior
Fix this bug | ValueError: "Minimax" object has no field "_client" | https://api.github.com/repos/langchain-ai/langchain/issues/11648/comments | 2 | 2023-10-11T06:52:49Z | 2024-02-08T16:20:41Z | https://github.com/langchain-ai/langchain/issues/11648 | 1,936,992,404 | 11,648 |
[
"langchain-ai",
"langchain"
] | ### System Info
langchain==0.0.311
python==3.9
I am trying to find out the similarity search score. but I got the score is abnormal. I hope the score is between 0 and 1.
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [X] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [X] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```python
from torch import cuda
from langchain.embeddings.huggingface import HuggingFaceEmbeddings
from langchain.vectorstores import (FAISS, Chroma)
from langchain.document_loaders import CSVLoader
embed_model_id = 'E:/modelscope/model/hggingface/text2vec-large-chinese'
device = f'cuda:{cuda.current_device()}' if cuda.is_available() else 'cpu'
embedding = HuggingFaceEmbeddings(
model_name=embed_model_id,
model_kwargs={'device': device},
encode_kwargs={'device': device, 'batch_size': 8}
)
def save_faiss_index(file_path: str = './label_remark_openai.csv', encoding: str = 'utf-8',
index_name: str = 'label_remark_openai_index',
save_path: str = "E:/pythonProject/information_demo/faiss_data"):
document = CSVLoader(file_path=file_path, encoding=encoding).load()
FAISS.normalize_L2 = True
db = FAISS.from_documents(documents=document, embedding=embedding, normalize_L2=True)
db.save_local(index_name=index_name, folder_path=save_path)
def get_faiss_index(index_name: str = 'label_remark_openai_index',
folder_path: str = "E:/pythonProject/information_demo/faiss_data"):
local_db = FAISS.load_local(index_name=index_name, embeddings=embedding, folder_path=folder_path, normalize_L2=True)
return local_db
if __name__ == '__main__':
index = get_faiss_index()
docs_and_scores = index.similarity_search_with_score("query text", k=10)
for doc, score in docs_and_scores:
print(f"Content: {doc.page_content}, Metadata: {doc.metadata}, Score: {score}")
```
### Expected behavior
I hope the score is between 0 and 1. | FAISS similarity search with score | https://api.github.com/repos/langchain-ai/langchain/issues/11645/comments | 2 | 2023-10-11T03:21:59Z | 2024-02-13T16:10:58Z | https://github.com/langchain-ai/langchain/issues/11645 | 1,936,721,958 | 11,645 |
[
"langchain-ai",
"langchain"
] | ### System Info
Python version - Python 3
Langchain version - I am not completely sure, but I think 0.0.196
### Who can help?
@eyurtsev - can you help with the above issue as I see you look after DataLoader Abstractions?
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [X] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
Code I wrote:
!pip install arxiv
!pip install pymupdf
from langchain.document_loaders import ArxivLoader
docs = ArxivLoader(query="2303.10130", load_max_docs=5, load_all_available_meta='FALSE').load()
### Expected behavior
Problem:
The problem is that I expected 'docs' variable to have at least 5 pages (as the article has 20+ pages) given i have passed the parameter load_max_docs = 5, but it only has 1. I tried with different numbers as well, but to no avail. | ArxivLoader always returns 1 document | https://api.github.com/repos/langchain-ai/langchain/issues/11644/comments | 5 | 2023-10-11T02:24:01Z | 2023-10-14T04:35:52Z | https://github.com/langchain-ai/langchain/issues/11644 | 1,936,650,375 | 11,644 |
[
"langchain-ai",
"langchain"
] | ### System Info
Flask==3.0.0
Flask-Cors==4.0.0
langchain==0.0.312
openai==0.28.1
pandas==2.1.1
tabulate==0.9.0
numpy==1.26.0
### Who can help?
@eyurtsev
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [X] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [X] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
```python
agent = create_csv_agent(
ChatOpenAI(temperature=0, model="gpt-4", openai_api_key=api_key, verbose=True),
["https://loyp2qzippf1q4hd.public.blob.vercel-storage.com/Mock%20Data%20_%20Dumping%20Costs%20-%20Dumping%20Cost-SIU4feVnqtCwoYg2g5Rgy9K5ejfKx0.csv"],
verbose=True,
agent_type=AgentType.OPENAI_FUNCTIONS,
return_intermediate_steps=True,
agent_executor_kwargs={"handle_parsing_errors": True},
)
```
### Expected behavior
The CSV file should be loaded accordingly | create_csv_agent cannot read files hosted on Vercel blob | https://api.github.com/repos/langchain-ai/langchain/issues/11637/comments | 4 | 2023-10-11T00:00:18Z | 2023-10-11T03:26:55Z | https://github.com/langchain-ai/langchain/issues/11637 | 1,936,494,704 | 11,637 |
[
"langchain-ai",
"langchain"
] | ### Issue you'd like to raise.
I submitted a PR. Vercel 'deployment' failed. There is no other info. I dont seem to have access for this PR on vercel. How do i debug this?
### Suggestion:
_No response_ | How to debug vercel failures? | https://api.github.com/repos/langchain-ai/langchain/issues/11635/comments | 7 | 2023-10-10T23:10:20Z | 2024-05-14T20:03:32Z | https://github.com/langchain-ai/langchain/issues/11635 | 1,936,432,030 | 11,635 |
[
"langchain-ai",
"langchain"
] | ### Feature request
Currently, the `save` method for `ChatPromptTemplate` is not implemented, preventing developers from saving templates as files.
### Motivation
This is a core feature that should be supported by all templates.
### Your contribution
If I have time I can make a PR. | feat: implement ChatPromptTemplate.save | https://api.github.com/repos/langchain-ai/langchain/issues/11631/comments | 3 | 2023-10-10T22:20:50Z | 2024-02-09T16:17:29Z | https://github.com/langchain-ai/langchain/issues/11631 | 1,936,362,500 | 11,631 |
[
"langchain-ai",
"langchain"
] | ### System Info
langchain version: 0.0.268
python version: 3.11
MacOS: Ventura 13.6
### Who can help?
@agola11
### Information
- [ ] The official example notebooks/scripts
- [X] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [X] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
Using a Chroma vector store's from_documents method through an AWS API Gateway. The response size appears to be greater than 6MB. Is there a way to send a request whos response becomes smaller than 6MB?
Retrying langchain.embeddings.openai.embed_with_retry.<locals>._embed_with_retry in 10.0 seconds as it raised APIError: HTTP code 501 from API (Response size larger than Lambda 6MB Limit).
Traceback (most recent call last):
File "/Users/john_ramos/mckinsey-git/policy-y/.venv/lib/python3.11/site-packages/openai/api_requestor.py", line 753, in _interpret_response_line
data = json.loads(rbody)
^^^^^^^^^^^^^^^^^
File "/usr/local/Cellar/python@3.11/3.11.4_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/json/__init__.py", line 346, in loads
return _default_decoder.decode(s)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/Cellar/python@3.11/3.11.4_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/json/decoder.py", line 337, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/Cellar/python@3.11/3.11.4_1/Frameworks/Python.framework/Versions/3.11/lib/python3.11/json/decoder.py", line 355, in raw_decode
raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)
### Expected behavior
Embeddings are stored in the data store after the response comes back through API gateway. | Issue as a result of Chroma.from_documents, OpenAI and AWS API Gateway | https://api.github.com/repos/langchain-ai/langchain/issues/11629/comments | 2 | 2023-10-10T22:04:33Z | 2024-02-06T16:23:46Z | https://github.com/langchain-ai/langchain/issues/11629 | 1,936,346,884 | 11,629 |
[
"langchain-ai",
"langchain"
] | ### Feature request
### Feature request
Azure Cosmos DB for MongoDB vCore enables users to efficiently store, index, and query high dimensional vector data stored directly in Azure Cosmos DB for MongoDB vCore. It contains similarity measures such as COS (cosine distance), L2 (Euclidean distance) or IP (inner product) which measures the distance between the data vectors and your query vector. The data vectors that are closest to your query vector are the ones that are found to be most similar semantically and retrieved during query time. The accompanying PR would add support for Langchain Python users to store vectors from document embeddings generated from APIs such as Azure OpenAI Embeddings or Hugging Face on Azure.
[Azure Cosmos DB for MongoDB vCore](https://learn.microsoft.com/en-us/azure/cosmos-db/mongodb/vcore/vector-search)
### Motivation
This capability described in the feature request is currently not available for Langchain Python.
### Your contribution
I will be submitting a PR for this feature request. | Add AzureCosmosDBVectorSearch VectorStore | https://api.github.com/repos/langchain-ai/langchain/issues/11627/comments | 1 | 2023-10-10T20:55:53Z | 2023-10-11T20:56:47Z | https://github.com/langchain-ai/langchain/issues/11627 | 1,936,230,465 | 11,627 |
[
"langchain-ai",
"langchain"
] | ### System Info
I'm using Langchain's Human Tool as part of my application. However, I am having difficulty integrating it into streamlit.
I'm mainly trying to change the input_func argument in the function:
`Human_Tool =load_tools(["human"], lm=llm, input_func=get_input)`
I've tried many different approaches but have not had success yet.
**I want to ask what might be a `get_input` function for langchains' Human Tool that integrates successfully with streamlit.**
### Who can help?
@hwchase17
@agola11
### Information
- [X] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [X] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
Here are some of the things I've tried:
```
import streamlit as st
def get_input() -> str:
st.write("Insert your text below.")
# Updated this line to include a label and hide it using label_visibility
user_input = st.text_area(label="Your Text", value="", height=100, key=None, help=None, on_change=None, args=None, kwargs=None, label_visibility="hidden")
submit_button = st.button("Submit")
if submit_button:
return user_input.strip() # This will return the text entered by the user
if not submit_button:
with st.empty():
for remaining in range(3600, 0, -1):
st.write(f"You have {remaining//60} minutes and {remaining%60} seconds to enter the text.")
time.sleep(1)
st.empty()
return "" # Returning empty if no input after an hour or not submitted
```
The code above opens a streamlit box that and accepts the user's input but it does not pass the users input to the agent.
import streamlit as st
import uuid
def get_input() -> str:
unique_key = str(uuid.uuid4())
user_input = st.text_area(
"Insert your text below:",
key=f"uniqueTextAreaKey_{unique_key}",
value="",
height=200,
)
if st.button("Submit", key=f"uniqueButtonKey_{unique_key}"):
if user_input: # check if user input is not empty
st.session_state['user_input'] = user_input
st.session_state['submitted'] = True
st.success("Input submitted successfully!")
return user_input.strip()
elif 'submitted' in st.session_state and st.session_state['submitted']:
st.success("Input already submitted.")
return st.session_state['user_input'].strip()
else:
st.warning("Please enter text and press submit.")
return ""
This code also opens up a streamlit text box but gets stuck in an infinite loop.
### Expected behavior
I want the human tool to integrate with streamlit. | How to integrate Langchain's Human Tool into Streamlit | https://api.github.com/repos/langchain-ai/langchain/issues/11626/comments | 9 | 2023-10-10T20:26:57Z | 2024-06-19T06:09:17Z | https://github.com/langchain-ai/langchain/issues/11626 | 1,936,153,248 | 11,626 |
[
"langchain-ai",
"langchain"
] | ### System Info
The method signatures of similaritySearchWithScore and similaritySearch are identical in Javascript. I assume this is a mistake.
```
export declare abstract class VectorStore extends Serializable {
...
similaritySearch(query: string, k?: number, filter?: this["FilterType"] | undefined, _callbacks?: Callbacks | undefined): Promise<Document[]>;
similaritySearchWithScore(query: string, k?: number, filter?: this["FilterType"] | undefined, _callbacks?: Callbacks | undefined): Promise<[Document, number][]>;
```
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [X] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
Open the base.d.ts file and search for the abstract class VectorStore. It contains these methods.
### Expected behavior
I would assume the signature for similaritySearchWithScore should contain some kind of Score parameter that we could pass to it to only return documents with a certain score. | similaritySearchWithScore and similaritySearch method signatures are identical in javascript | https://api.github.com/repos/langchain-ai/langchain/issues/11625/comments | 4 | 2023-10-10T20:24:24Z | 2024-02-09T16:17:34Z | https://github.com/langchain-ai/langchain/issues/11625 | 1,936,149,678 | 11,625 |
[
"langchain-ai",
"langchain"
] | ### System Info
https://github.com/langchain-ai/langchain/releases/tag/v0.0.304
Reverting to 3.0.3 or under fixes it. The error looks something like (truncated confidential parts of the trace)
```
Traceback (most recent call last):
...
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/ddtrace/_monkey.py", line 205, in patch_all
patch(raise_errors=False, **modules)
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/ddtrace/_monkey.py", line 245, in patch
when_imported(module)(
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/ddtrace/vendor/wrapt/importer.py", line 284, in register
register_post_import_hook(hook, name)
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/ddtrace/vendor/wrapt/decorators.py", line 470, in _synchronized
return wrapped(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/ddtrace/vendor/wrapt/importer.py", line 82, in register_post_import_hook
hook(module)
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/ddtrace/_monkey.py", line 168, in on_import
imported_module.patch()
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/ddtrace/contrib/langchain/patch.py", line 763, in patch
if hasattr(langchain.embeddings, text_embedding_model):
^^^^^^^^^^^^^^^^^^^^
File "/Users/sid.gupta/.pyenv/versions/3.11.3/lib/python3.11/site-packages/langchain/__init__.py", line 332, in __getattr__
raise AttributeError(f"Could not find: {name}")
AttributeError: Could not find: embeddings
```
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
Be on `ddtrace 1.17.6` and a langchain version >= 3.0.4. Running an webapp on hypercorn and fastAPI.
### Expected behavior
Don't throw this trace and block the application. | Release 3.0.4 breaks compatibility with ddtrace. | https://api.github.com/repos/langchain-ai/langchain/issues/11624/comments | 3 | 2023-10-10T20:12:25Z | 2024-02-06T16:23:56Z | https://github.com/langchain-ai/langchain/issues/11624 | 1,936,133,212 | 11,624 |
[
"langchain-ai",
"langchain"
] | ### System Info
from langchain.docstore.document import Document
from typing import Dict
from langchain import PromptTemplate, SagemakerEndpoint
from langchain.llms.sagemaker_endpoint import LLMContentHandler
from langchain.chains.question_answering import load_qa_chain
import json
example_doc_1 = """
Peter and Elizabeth took a taxi to attend the night party in the city. While in the party, Elizabeth collapsed and was rushed to the hospital.
Since she was diagnosed with a brain injury, the doctor told Peter to stay besides her until she gets well.
Therefore, Peter stayed with her at the hospital for 3 days without leaving.
"""
docs = [
Document(
page_content=example_doc_1,
)
]
query = """How long was Elizabeth hospitalized?
"""
prompt_template = """Use the following pieces of context to answer the question at the end.
{context}
Question: {question}
Answer:"""
PROMPT = PromptTemplate(
template=prompt_template, input_variables=["context", "question"]
)
class ContentHandler(LLMContentHandler):
content_type = "application/json"
accepts = "application/json"
def transform_input(self, prompt: str, model_kwargs: Dict) -> bytes:
input_dict = {"inputs": prompt, "parameters": model_kwargs}
return json.dumps(input_dict).encode('utf-8')
def transform_output(self, output: bytes) -> str:
response_json = json.loads(output.read().decode("utf-8"))
return response_json[0]["generated_text"]
content_handler = ContentHandler()
chain = load_qa_chain(
llm=SagemakerEndpoint(
endpoint_name="endpoint",
region_name="ap-southeast-2",
credentials_profile_name="dev",
model_kwargs={"temperature": 1e-10},
endpoint_kwargs={"CustomAttributes": 'accept_eula=true'},
content_handler=ContentHandler(),
),
prompt=PROMPT,
)
chain({"input_documents": docs, "question": query}, return_only_outputs=True)
Stacktrace :
ValueError: Error raised by inference endpoint: An error occurred (ModelError) when calling the InvokeEndpoint operation: Received client error (424) from primary with message "{
"code":424,
"message":"prediction failure",
"error":"string indices must be integers"
### Who can help?
_No response_
### Information
- [ ] The official example notebooks/scripts
- [ ] My own modified scripts
### Related Components
- [ ] LLMs/Chat Models
- [ ] Embedding Models
- [ ] Prompts / Prompt Templates / Prompt Selectors
- [ ] Output Parsers
- [ ] Document Loaders
- [ ] Vector Stores / Retrievers
- [ ] Memory
- [ ] Agents / Agent Executors
- [ ] Tools / Toolkits
- [ ] Chains
- [ ] Callbacks/Tracing
- [ ] Async
### Reproduction
run the same python code to reproduce.
### Expected behavior
I am expecting a response generated by the chain. | AWS Sagemaker "code":424, | https://api.github.com/repos/langchain-ai/langchain/issues/11618/comments | 3 | 2023-10-10T17:47:53Z | 2024-02-10T16:14:07Z | https://github.com/langchain-ai/langchain/issues/11618 | 1,935,901,944 | 11,618 |
[
"langchain-ai",
"langchain"
] | ### Feature request
Add self-querying capability for MongoDBAtlasVectorSearch so that any natural language query could be converted to a MongoDB query and then be used to search the MongoDB vector store
### Motivation
I was testing self-quering using MongoDBAtlasVectorSearch when I got an error saying that self-querying is not supported. I think it will be useful to have that functionality since MongoDB vector store is well used.
### Your contribution
We will work on this feature and submit a pull request in the coming 1-2 months. | Self-querying retriever for MongoDBAtlasVectorSearch | https://api.github.com/repos/langchain-ai/langchain/issues/11616/comments | 2 | 2023-10-10T16:37:06Z | 2024-01-30T18:27:37Z | https://github.com/langchain-ai/langchain/issues/11616 | 1,935,763,708 | 11,616 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.