<!-- Thank you for contributing to LangChain!
Replace this entire comment with:
- **Description:** a description of the change,
- **Issue:** the issue # it fixes (if applicable),
- **Dependencies:** any dependencies required for this change,
- **Tag maintainer:** for a quicker response, tag the relevant
maintainer (see below),
- **Twitter handle:** we announce bigger features on Twitter. If your PR
gets announced, and you'd like a mention, we'll gladly shout you out!
Please make sure your PR is passing linting and testing before
submitting. Run `make format`, `make lint` and `make test` to check this
locally.
See contribution guidelines for more information on how to write/run
tests, lint, etc:
https://github.com/langchain-ai/langchain/blob/master/.github/CONTRIBUTING.md
If you're adding a new integration, please include:
1. a test for the integration, preferably unit tests that do not rely on
network access,
2. an example notebook showing its use. It lives in `docs/extras`
directory.
If no one reviews your PR within a few days, please @-mention one of
@baskaryan, @eyurtsev, @hwchase17.
-->
It was passing in message instead of generation
<!-- Thank you for contributing to LangChain!
Replace this entire comment with:
- **Description:** a description of the change,
- **Issue:** the issue # it fixes (if applicable),
- **Dependencies:** any dependencies required for this change,
- **Tag maintainer:** for a quicker response, tag the relevant
maintainer (see below),
- **Twitter handle:** we announce bigger features on Twitter. If your PR
gets announced, and you'd like a mention, we'll gladly shout you out!
Please make sure your PR is passing linting and testing before
submitting. Run `make format`, `make lint` and `make test` to check this
locally.
See contribution guidelines for more information on how to write/run
tests, lint, etc:
https://github.com/langchain-ai/langchain/blob/master/.github/CONTRIBUTING.md
If you're adding a new integration, please include:
1. a test for the integration, preferably unit tests that do not rely on
network access,
2. an example notebook showing its use. It lives in `docs/extras`
directory.
If no one reviews your PR within a few days, please @-mention one of
@baskaryan, @eyurtsev, @hwchase17.
-->
* Restrict the chain to specific domains by default
* This is a breaking change, but it will fail loudly upon object
instantiation -- so there should be no silent errors for users
* Resolves CVE-2023-32786
* This is an opt-in feature, so users should be aware of risks if using
jinja2.
* Regardless we'll add sandboxing by default to jinja2 templates -- this
sandboxing is a best effort basis.
* Best strategy is still to make sure that jinja2 templates are only
loaded from trusted sources.
**Description:** Update `langchain.document_loaders.pdf.PyPDFLoader` to
store url in metadata (instead of a temporary file path) if user
provides a web path to a pdf
- **Issue:** Related to #7034; the reporter on that issue submitted a PR
updating `PyMuPDFParser` for this behavior, but it has unresolved merge
issues as of 20 Oct 2023 #7077
- In addition to `PyPDFLoader` and `PyMuPDFParser`, these other classes
in `langchain.document_loaders.pdf` exhibit similar behavior and could
benefit from an update: `PyPDFium2Loader`, `PDFMinerLoader`,
`PDFMinerPDFasHTMLLoader`, `PDFPlumberLoader` (I'm happy to contribute
to some/all of that, including assisting with `PyMuPDFParser`, if my
work is agreeable)
- The root cause is that the underlying pdf parser classes, e.g.
`langchain.document_loaders.parsers.pdf.PyPDFParser`, never receive
information about the url; the parsers receive a
`langchain.document_loaders.blob_loaders.blob`, which contains the pdf
contents and local file path, but not the url
- This update passes the web path directly to the parser since it's
minimally invasive and doesn't require further changes to maintain
existing behavior for local files... bigger picture, I'd consider
extending `blob` so that extra information like this can be
communicated, but that has much bigger implications on the codebase
which I think warrants maintainer input
- **Dependencies:** None
```python
# old behavior
>>> from langchain.document_loaders import PyPDFLoader
>>> loader = PyPDFLoader('https://arxiv.org/pdf/1706.03762.pdf')
>>> docs = loader.load()
>>> docs[0].metadata
{'source': '/var/folders/w2/zx77z1cs01s1thx5dhshkd58h3jtrv/T/tmpfgrorsi5/tmp.pdf', 'page': 0}
# new behavior
>>> from langchain.document_loaders import PyPDFLoader
>>> loader = PyPDFLoader('https://arxiv.org/pdf/1706.03762.pdf')
>>> docs = loader.load()
>>> docs[0].metadata
{'source': 'https://arxiv.org/pdf/1706.03762.pdf', 'page': 0}
```
- **Description:** #12273 's suggestion PR
Like other PDFLoader, loading pdf per each page and giving page
metadata.
- **Issue:** #12273
- **Twitter handle:** @blue0_0hope
---------
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
This will allow you create the schema beforehand. The check was failing
and preventing importing into existing classes.
<!-- Thank you for contributing to LangChain!
Replace this entire comment with:
- **Description:** a description of the change,
- **Issue:** the issue # it fixes (if applicable),
- **Dependencies:** any dependencies required for this change,
- **Tag maintainer:** for a quicker response, tag the relevant
maintainer (see below),
- **Twitter handle:** we announce bigger features on Twitter. If your PR
gets announced, and you'd like a mention, we'll gladly shout you out!
Please make sure your PR is passing linting and testing before
submitting. Run `make format`, `make lint` and `make test` to check this
locally.
See contribution guidelines for more information on how to write/run
tests, lint, etc:
https://github.com/langchain-ai/langchain/blob/master/.github/CONTRIBUTING.md
If you're adding a new integration, please include:
1. a test for the integration, preferably unit tests that do not rely on
network access,
2. an example notebook showing its use. It lives in `docs/extras`
directory.
If no one reviews your PR within a few days, please @-mention one of
@baskaryan, @eyurtsev, @hwchase17.
-->
- **Description:** implement [quip](https://quip.com) loader
- **Issue:** https://github.com/langchain-ai/langchain/issues/10352
- **Dependencies:** No
- pass make format, make lint, make test
---------
Co-authored-by: Hao Fan <h_fan@apple.com>
Co-authored-by: Bagatur <baskaryan@gmail.com>
latest release broken, this fixes it
---------
Co-authored-by: Roman Vasilyev <rvasilyev@mozilla.com>
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
Prior to this PR, `ruff` was used only for linting and not for
formatting, despite the names of the commands. This PR makes it be used
for both linting code and autoformatting it.
This input key was missed in the last update PR:
https://github.com/langchain-ai/langchain/pull/7391
The input/output formats are intended to be like this:
```
{"inputs": [<prompt>]}
{"outputs": [<output_text>]}
```
## Description
This PR adds support for
[lm-format-enforcer](https://github.com/noamgat/lm-format-enforcer) to
LangChain.
The library is similar to jsonformer / RELLM which are supported in
Langchain, but has several advantages such as
- Batching and Beam search support
- More complete JSON Schema support
- LLM has control over whitespace, improving quality
- Better runtime performance due to only calling the LLM's generate()
function once per generate() call.
The integration is loosely based on the jsonformer integration in terms
of project structure.
## Dependencies
No compile-time dependency was added, but if `lm-format-enforcer` is not
installed, a runtime error will occur if it is trying to be used.
## Tests
Due to the integration modifying the internal parameters of the
underlying huggingface transformer LLM, it is not possible to test
without building a real LM, which requires internet access. So, similar
to the jsonformer and RELLM integrations, the testing is via the
notebook.
## Twitter Handle
[@noamgat](https://twitter.com/noamgat)
Looking forward to hearing feedback!
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
Best to review one commit at a time, since two of the commits are 100%
autogenerated changes from running `ruff format`:
- Install and use `ruff format` instead of black for code formatting.
- Output of `ruff format .` in the `langchain` package.
- Use `ruff format` in experimental package.
- Format changes in experimental package by `ruff format`.
- Manual formatting fixes to make `ruff .` pass.
I always take 20-30 seconds to re-discover where the
`convert_to_openai_function` wrapper lives in our codebase. Chat
langchain [has no
clue](https://smith.langchain.com/public/3989d687-18c7-4108-958e-96e88803da86/r)
what to do either. There's the older `create_openai_fn_chain` , but we
haven't been recommending it in LCEL. The example we show in the
[cookbook](https://python.langchain.com/docs/expression_language/how_to/binding#attaching-openai-functions)
is really verbose.
General function calling should be as simple as possible to do, so this
seems a bit more ergonomic to me (feel free to disagree). Another option
would be to directly coerce directly in the class's init (or when
calling invoke), if provided. I'm not 100% set against that. That
approach may be too easy but not simple. This PR feels like a decent
compromise between simple and easy.
```
from enum import Enum
from typing import Optional
from pydantic import BaseModel, Field
class Category(str, Enum):
"""The category of the issue."""
bug = "bug"
nit = "nit"
improvement = "improvement"
other = "other"
class IssueClassification(BaseModel):
"""Classify an issue."""
category: Category
other_description: Optional[str] = Field(
description="If classified as 'other', the suggested other category"
)
from langchain.chat_models import ChatOpenAI
llm = ChatOpenAI().bind_functions([IssueClassification])
llm.invoke("This PR adds a convenience wrapper to the bind argument")
# AIMessage(content='', additional_kwargs={'function_call': {'name': 'IssueClassification', 'arguments': '{\n "category": "improvement"\n}'}})
```
- Prefer lambda type annotations over inferred dict schema
- For sequences that start with RunnableAssign infer seq input type as
"input type of 2nd item in sequence - output type of runnable assign"
Replace this entire comment with:
-Add MultiOn close function and update key value and add async
functionality
- solved the key value TabId not found.. (updated to use latest key
value)
@hwchase17
- **Description:** This pull request removes secrets present in raw
format,
- **Issue:** Fireworks api key was exposed when printing out the
langchain object
[#12165](https://github.com/langchain-ai/langchain/issues/12165)
- **Maintainer:** @eyurtsev
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
**Description:** Textract PDF Loader generating linearized output,
meaning it will replicate the structure of the source document as close
as possible based on the features passed into the call (e. g. LAYOUT,
FORMS, TABLES). With LAYOUT reading order for multi-column documents or
identification of lists and figures is supported and with TABLES it will
generate the table structure as well. FORMS will indicate "key: value"
with columms.
- **Issue:** the issue fixes#12068
- **Dependencies:** amazon-textract-textractor is added, which provides
the linearization
- **Tag maintainer:** @3coins
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
can get the correct token count instead of using gpt-2 model
**Description:**
Implement get_num_tokens within VertexLLM to use google's count_tokens
function.
(https://cloud.google.com/vertex-ai/docs/generative-ai/get-token-count).
So we don't need to download gpt-2 model from huggingface, also when we
do the mapreduce chain we can get correct token count.
**Tag maintainer:**
@lkuligin
**Twitter handle:**
My twitter: @abehsu1992626
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
Following this tutoral about using OpenAI Embeddings with FAISS
https://python.langchain.com/docs/integrations/vectorstores/faiss
```python
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.text_splitter import CharacterTextSplitter
from langchain.vectorstores import FAISS
from langchain.document_loaders import TextLoader
from langchain.document_loaders import TextLoader
loader = TextLoader("../../../extras/modules/state_of_the_union.txt")
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
docs = text_splitter.split_documents(documents)
embeddings = OpenAIEmbeddings()
```
This works fine
```python
db = FAISS.from_documents(docs, embeddings)
query = "What did the president say about Ketanji Brown Jackson"
docs = db.similarity_search(query)
```
But the async version is not
```python
db = await FAISS.afrom_documents(docs, embeddings) # NotImplementedError
query = "What did the president say about Ketanji Brown Jackson"
docs = await db.asimilarity_search(query) # this will use await asyncio.get_event_loop().run_in_executor under the hood and will not call OpenAIEmbeddings.aembed_query but call OpenAIEmbeddings.embed_query
```
So this PR add async/await supports for FAISS
---------
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
- Description: adding support to Activeloop's DeepMemory feature that
boosts recall up to 25%. Added Jupyter notebook showcasing the feature
and also made index params explicit.
- Twitter handle: will really appreciate if we could announce this on
twitter.
---------
Co-authored-by: adolkhan <adilkhan.sarsen@alumni.nu.edu.kz>
Hey, we're looking to invest more in adding cohere integrations to
langchain so would love to get more of an idea for how it's used.
Hopefully this pr is acceptable. This week I'm also going to be looking
into adding our new [retrieval augmented generation
product](https://txt.cohere.com/chat-with-rag/) to langchain.
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
<!-- Thank you for contributing to LangChain!
Replace this entire comment with:
- **Description:** a description of the change,
- **Issue:** the issue # it fixes (if applicable),
- **Dependencies:** any dependencies required for this change,
- **Tag maintainer:** for a quicker response, tag the relevant
maintainer (see below),
- **Twitter handle:** we announce bigger features on Twitter. If your PR
gets announced, and you'd like a mention, we'll gladly shout you out!
Please make sure your PR is passing linting and testing before
submitting. Run `make format`, `make lint` and `make test` to check this
locally.
See contribution guidelines for more information on how to write/run
tests, lint, etc:
https://github.com/langchain-ai/langchain/blob/master/.github/CONTRIBUTING.md
If you're adding a new integration, please include:
1. a test for the integration, preferably unit tests that do not rely on
network access,
2. an example notebook showing its use. It lives in `docs/extras`
directory.
If no one reviews your PR within a few days, please @-mention one of
@baskaryan, @eyurtsev, @hwchase17.
-->
## **Description:**
When building our own readthedocs.io scraper, we noticed a couple
interesting things:
1. Text lines with a lot of nested <span> tags would give unclean text
with a bunch of newlines. For example, for [Langchain's
documentation](https://api.python.langchain.com/en/latest/document_loaders/langchain.document_loaders.readthedocs.ReadTheDocsLoader.html#langchain.document_loaders.readthedocs.ReadTheDocsLoader),
a single line is represented in a complicated nested HTML structure, and
the naive `soup.get_text()` call currently being made will create a
newline for each nested HTML element. Therefore, the document loader
would give a messy, newline-separated blob of text. This would be true
in a lot of cases.
<img width="945" alt="Screenshot 2023-10-26 at 6 15 39 PM"
src="https://github.com/langchain-ai/langchain/assets/44193474/eca85d1f-d2bf-4487-a18a-e1e732fadf19">
<img width="1031" alt="Screenshot 2023-10-26 at 6 16 00 PM"
src="https://github.com/langchain-ai/langchain/assets/44193474/035938a0-9892-4f6a-83cd-0d7b409b00a3">
Additionally, content from iframes, code from scripts, css from styles,
etc. will be gotten if it's a subclass of the selector (which happens
more often than you'd think). For example, [this
page](https://pydeck.gl/gallery/contour_layer.html#) will scrape 1.5
million characters of content that looks like this:
<img width="1372" alt="Screenshot 2023-10-26 at 6 32 55 PM"
src="https://github.com/langchain-ai/langchain/assets/44193474/dbd89e39-9478-4a18-9e84-f0eb91954eac">
Therefore, I wrote a recursive _get_clean_text(soup) class function that
1. skips all irrelevant elements, and 2. only adds newlines when
necessary.
2. Index pages (like [this
one](https://api.python.langchain.com/en/latest/api_reference.html))
would be loaded, chunked, and eventually embedded. This is really bad
not just because the user will be embedding irrelevant information - but
because index pages are very likely to show up in retrieved content,
making retrieval less effective (in our tests). Therefore, I added a
bool parameter `exclude_index_pages` defaulted to False (which is the
current behavior — although I'd petition to default this to True) that
will skip all pages where links take up 50%+ of the page. Through manual
testing, this seems to be the best threshold.
## Other Information:
- **Issue:** n/a
- **Dependencies:** n/a
- **Tag maintainer:** n/a
- **Twitter handle:** @andrewthezhou
---------
Co-authored-by: Andrew Zhou <andrew@heykona.com>
Co-authored-by: Bagatur <baskaryan@gmail.com>
- **Description:**
* Add unit tests for document_transformers/beautiful_soup_transformer.py
* Basic functionality is tested (extract tags, remove tags, drop lines)
* add a FIXME comment about the order of tags that is not preserved
(and a passing test, but with the expected tags now out-of-order)
- **Issue:** None
- **Dependencies:** None
- **Tag maintainer:** @rlancemartin
- **Twitter handle:** `peter_v`
Please make sure your PR is passing linting and testing before
submitting.
=> OK: I ran `make format`, `make test` (passing after install of
beautifulsoup4) and `make lint`.
- **Description:** Added masking of the API Key for AI21 LLM when
printed and improved the docstring for AI21 LLM.
- Updated the AI21 LLM to utilize SecretStr from pydantic to securely
manage API key.
- Made improvements in the docstring of AI21 LLM. It now mentions that
the API key can also be passed as a named parameter to the constructor.
- Added unit tests.
- **Issue:** #12165
- **Tag maintainer:** @eyurtsev
---------
Co-authored-by: Anirudh Gautam <anirudh@Anirudhs-Mac-mini.local>
Currently this gives a bug:
```
from langchain.schema.runnable import RunnableLambda
bound = RunnableLambda(lambda x: x).with_config({"callbacks": []})
# ConfigError: field "callbacks" not yet prepared so type is still a ForwardRef, you might need to call RunnableConfig.update_forward_refs().
```
Rather than deal with cyclic imports and extra load time, etc., I think
it makes sense to just have a separate Callbacks definition here that is
a relaxed typehint.
1. Allow run evaluators to return {"results": [list of evaluation
results]} in the evaluator callback.
2. Allows run evaluators to pick the target run ID to provide feedback
to
(1) means you could do something like a function call that populates a
full rubric in one go (not sure how reliable that is in general though)
rather than splitting off into separate LLM calls - cheaper and less
code to write
(2) means you can provide feedback to runs on subsequent calls.
Immediate use case is if you wanted to add an evaluator to a chat bot
and assign to assign to previous conversation turns
have a corresponding one in the SDK
In the GoogleSerperResults class, the name field is defined as
'google_serrper_results_json'. This looks like a typo, and perhaps
should be 'google_serper_results_json'.
<!-- Thank you for contributing to LangChain!
Replace this entire comment with:
- **Description:** a description of the change,
- **Issue:** the issue # it fixes (if applicable),
- **Dependencies:** any dependencies required for this change,
- **Tag maintainer:** for a quicker response, tag the relevant
maintainer (see below),
- **Twitter handle:** we announce bigger features on Twitter. If your PR
gets announced, and you'd like a mention, we'll gladly shout you out!
Please make sure your PR is passing linting and testing before
submitting. Run `make format`, `make lint` and `make test` to check this
locally.
See contribution guidelines for more information on how to write/run
tests, lint, etc:
https://github.com/langchain-ai/langchain/blob/master/.github/CONTRIBUTING.md
If you're adding a new integration, please include:
1. a test for the integration, preferably unit tests that do not rely on
network access,
2. an example notebook showing its use. It lives in `docs/extras`
directory.
If no one reviews your PR within a few days, please @-mention one of
@baskaryan, @eyurtsev, @hwchase17.
-->
Add Redis langserve template! Eventually will add semantic caching to
this too. But I was struggling to get that to work for some reason with
the LCEL implementation here.
- **Description:** Introduces the Redis LangServe template. A simple RAG
based app built on top of Redis that allows you to chat with company's
public financial data (Edgar 10k filings)
- **Issue:** None
- **Dependencies:** The template contains the poetry project
requirements to run this template
- **Tag maintainer:** @baskaryan @Spartee
- **Twitter handle:** @tchutch94
**Note**: this requires the commit here that deletes the
`_aget_relevant_documents()` method from the Redis retriever class that
wasn't implemented. That was breaking the langserve app.
---------
Co-authored-by: Sam Partee <sam.partee@redis.com>
-**Description** Adds returning the reranking score when using semantic
search
-**Issue:* #12317
---------
Co-authored-by: Adam Law <adamlaw@microsoft.com>
Co-authored-by: Bagatur <baskaryan@gmail.com>
**Description:** Improve handling of empty queries in timescale-vector.
For timescale-vector it is more efficient to get a None embedding when
the embedding has no semantic meaning. It allows timescale-vector to
perform more optimizations. Thus, when the query is empty, use a None
embedding.
Also pass down constructor arguments to the timescale vector client.
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
This code path is hit in the following case:
- Start in langchain code and manually provide a tracer
- Handoff to the traceable
- Hand back to langchain code.
Which happens for evaluating `@traceable` functions unfortunately
- **Description: To handle the hybrid search with RRF(Reciprocal Rank
Fusion) in the Elasticsearch, rrf argument was added for adjusting
'rank_constant' and 'window_size' to combine multiple result sets with
different relevance indicators into a single result set. (ref:
https://www.elastic.co/kr/blog/whats-new-elastic-enterprise-search-8-9-0),
- **Issue:** the issue # it fixes (if applicable),
- **Dependencies:** No dependencies changed,
- **Tag maintainer:** @baskaryan,
Nice to meet you,
I'm a newbie for contributions and it's my first PR.
I only changed the langchain/vectorstores/elasticsearch.py file.
I did make format&lint
I got this message,
```shell
make lint_diff
./scripts/check_pydantic.sh .
./scripts/check_imports.sh
poetry run ruff .
[ "langchain/vectorstores/elasticsearch.py" = "" ] || poetry run black langchain/vectorstores/elasticsearch.py --check
All done! ✨🍰✨
1 file would be left unchanged.
[ "langchain/vectorstores/elasticsearch.py" = "" ] || poetry run mypy langchain/vectorstores/elasticsearch.py
langchain/__init__.py: error: Source file found twice under different module names: "mvp.nlp.langchain.libs.langchain.langchain" and "langchain"
Found 1 error in 1 file (errors prevented further checking)
make: *** [lint_diff] Error 2
```
Thank you
---------
Co-authored-by: 황중원 <jwhwang@amorepacific.com>
My postgres out of connections after continuous PGVector usage, and the
reason because it constantly creates new connections, so adding a
reusable pre established connection seems like solves an issue
---------
Co-authored-by: Roman Vasilyev <rvasilyev@mozilla.com>
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
See discussion here:
https://github.com/langchain-ai/langchain/discussions/11680
The code is available for usage from langchain_experimental. The reason
for the deprecation is that the agents are relying on a Python REPL. The
code can only be run safely with appropriate sandboxing.
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
The changes introduced in #12267 and #12190 broke the cost computation
of the `completion` tokens for fine-tuned models because of the early
return. This PR aims at fixing this.
@baskaryan.
**Description:**
Revise `libs/langchain/langchain/document_loaders/async_html.py` to
store the HTML Title and Page Language in the `metadata` of
`AsyncHtmlLoader`.
Compare predicted json to reference. First canonicalize (sort keys, rm
whitespace separators), then return normalized string edit distance.
Not a silver bullet but maybe an easy way to capture structure
differences in a less flakey way
---------
Co-authored-by: Bagatur <22008038+baskaryan@users.noreply.github.com>
Will run all CI because of _test change, but future PRs against CLI will
only trigger the new CLI one
Has a bunch of file changes related to formatting/linting.
No mypy yet - coming soon
