Commit Graph

1505 Commits

Author SHA1 Message Date
Vikram Shitole
0d44746430
10634: Added the capability to inject boto3 client in SagemakerEndpointEmbeddings (#12146)
**Description: Allow to inject boto3 client for Cross account access
type of scenarios in using SagemakerEndpointEmbeddings and also updated
the documentation for same in the sample notebook**

**Issue:SagemakerEndpointEmbeddings cross account capability #10634
#10184**

Dependencies: None
Tag maintainer:
Twitter handle:lethargicoder

Co-authored-by: Vikram(VS) <vssht@amazon.com>
2023-10-23 15:08:26 -04:00
aubin_mzt
66f8cb015d
Add connection args for pgvector vector store (#11930)
- **Description:** sqlalchemy create_engine() does not take into account
connect_args which are mandatory for managed PGSQL instances on cloud
providers (ssl_context for example).
Also re-enabled create_vector_extension at post_init for using pgvector
class seamlessly
- **Tag maintainer:** @baskaryan, @eyurtsev, @hwchase17.

---------

Co-authored-by: Sami Bargaoui <bargaoui.sam@gmail.com>
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
2023-10-23 14:43:44 -04:00
NuODaniel
4d6243fa87
fix: doc string of default params in chat_models, llm qianfan (#12153)
- **Description:** a fix of the doc string in Qianfan
  - **Issue:** no
  - **Dependencies:** no
  - **Tag maintainer:** @baskaryan
  - **Twitter handle:** no
2023-10-23 14:03:18 -04:00
Predrag Gruevski
f82bdf4613
Update deprecated langchain imports with suggested new paths. (#12164)
Let's help our users find the proper import to use instead of the
deprecated top-level ones.
2023-10-23 13:52:08 -04:00
Bagatur
963ff93476
bump 321 (#12161) 2023-10-23 12:49:38 -04:00
Nuno Campos
d0505c0d47
Update default recursion_limit, update docs (#12134)
<!-- 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.
 -->
2023-10-23 16:29:17 +01:00
William FH
4f23aa677a
Fix Pickle Error (#12141)
If non-pickleable objects (like locks) get passed to the tracing
callback, they'll fail in the deepcopy. Fallback to a shallow copy in
these instances .
2023-10-23 08:22:47 -07:00
Predrag Gruevski
95a1b598fe
Update to actions/checkout@v4. (#11951)
We don't use any of the new functionality at the moment. Just making
sure we don't fall back on versions and fail to benefit from new
patches. This is an easy upgrade and it's always harder to upgrade
across multiple major versions at once.
2023-10-23 10:01:33 -04:00
William FH
7c4f340cc0
Include Parent Run ID (#12139)
If you set local callbacks
2023-10-22 17:19:11 -07:00
omahs
f3cc9bba5b
Fix typos (#12128)
Fix typos
2023-10-22 17:16:03 -07:00
Nuno Campos
1afdb40b48
Add optional config arg to RunnablePassthrough func arg (#12131)
<!-- 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.
 -->
2023-10-22 19:57:16 +01:00
Nuno Campos
325fdde8b4
Fix bug where types were lost when calling with_cconfig or bind (#12137)
<!-- 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.
 -->
2023-10-22 19:26:13 +01:00
Nuno Campos
02dce74b97
Fix type hint for older py versions (#12132)
<!-- 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.
 -->
2023-10-22 18:01:09 +01:00
Nuno Campos
d0ce374731
Allow specifying custom input/output schemas for runnables with .with_types() (#12083)
<!-- 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.
 -->
2023-10-22 17:26:48 +01:00
Harrison Chase
ee69116761
move csv agent to langchain experimental (#12113) 2023-10-21 10:26:02 -07:00
Harrison Chase
03bf6ef473
add missing init files (#12114) 2023-10-21 10:25:50 -07:00
Bagatur
ef8b180d6d
bump 320 (#12108) 2023-10-21 11:52:52 -04:00
Rotem Weiss
78d186fb44
Add Tavily Search API as a Tool (#12103)
Adding Tavily Search API as a tool. I will be the maintainer and
assaf_elovic is the twitter handler.

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
2023-10-21 11:23:21 -04:00
Bagatur
85302a9ec1
Add CI check that integration tests compile (#12090) 2023-10-21 10:52:18 -04:00
verlocks
5dbe456aae
Bug fix tongyi.py to be compatible with DashScope API (#11956)
Current ChatTongyi is not compatible with DashScope API, which will
cause error when passing api key to chat model directly.
- **Description:** Update tongyi.py to be compatible with DashScope API.
Specifically, update parameter name "dashscope_api_key" to "api_key".
  - **Issue:** None.
- **Dependencies:** Nothing new, Tongyi would require DashScope as
before.
2023-10-20 18:46:41 -04:00
Tomaz Bratanic
82f4c0589c
Add neo4j graph environment variables (#12080) 2023-10-20 14:43:01 -07:00
Mohammad Mohtashim
d5400f6502
Google Scholar Search Tool using serpapi (#11513)
- **Description:** Implementing the Google Scholar Tool as requested in
PR #11505. The tool will be using the [serpapi python
package](https://serpapi.com/integrations/python#search-google-scholar).
The main idea of the tool will be to return the results from a Google
Scholar search given a query as an input to the tool.

- **Tag maintainer:** @baskaryan, @eyurtsev, @hwchase17
2023-10-20 17:35:55 -04:00
Holt Skinner
f5be2d525a
fix: Add _serving_config property to GoogleVertexAISearchRetriever (#12084)
- Fixes error:

```
ValueError: "GoogleVertexAISearchRetriever" object has no field "_serving_config"
```

Introduced in #11736

@baskaryan, @eyurtsev, @hwchase17 if you could review and merge quickly,
that would be appreciated :)
2023-10-20 15:16:42 -04:00
Nuno Campos
5fee61a207
Support runnable factories in .configurable_alts() (#12065)
<!-- 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.
 -->
2023-10-20 15:22:09 +01:00
Zhitao Xu
a4c3a44712
Fix documentation typo in Clickhouse Class (#12047)
<!-- 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:** The return info in the documentation for
similarity_search_by_vector and similarity_search_with_relevance_scores
is wrong
2023-10-19 17:00:22 -04:00
William FH
25418b9b4d
Always add run ID (#12046)
in eval callback handler.

Useful if you're using a custom run evaluator and don't want to thread
things through.
2023-10-19 12:38:07 -07:00
Eugene Yurtsev
44d7763580
Add zapier deprecation warning (#12045)
Add zapier deprecation
2023-10-19 15:27:56 -04:00
John Mai
4188f046ec
Add Tencent Hunyuan chat model (#12022)
### Description:
The Tencent Hunyuan model, developed by Tencent, is a large language
model by robust Chinese text generation capabilities, adeptness in
logical reasoning within complex contexts, and reliable task execution
proficiency.For more information, see
[https://cloud.tencent.com/document/product/1729](https://cloud.tencent.com/document/product/1729)
2023-10-19 15:10:12 -04:00
Eugene Yurtsev
68599d98c2
More security notes (#12040)
Add more security notes
2023-10-19 14:49:09 -04:00
Bagatur
0006075b08
bump 319 (#12041) 2023-10-19 11:45:27 -07:00
John Mai
8eb40b5fe2
baichuan_secret_key use pydantic.types.SecretStr & Add Baichuan tests (#12031)
### Description
- `baichuan_secret_key` use pydantic.types.SecretStr
- Add Baichuan tests
2023-10-19 14:37:41 -04:00
Nuno Campos
85bac75729
nc/runnable-dynamic-schemas-from-config (#12038)
<!-- 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.
 -->
2023-10-19 19:34:35 +01:00
Nuno Campos
85eaa4ccee
Revert "nc/runnable-dynamic-schemas-from-config" (#12037)
This reverts commit a46eef64a7.

<!-- 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.
 -->
2023-10-19 19:27:02 +01:00
Nuno Campos
a46eef64a7 nc/runnable-dynamic-schemas-from-config 2023-10-19 19:17:48 +01:00
Nuno Campos
d392e030be
Add default value (#12032)
<!-- 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.
 -->
2023-10-19 18:30:05 +01:00
Kenneth Choe
62efe1ffb9
support add_embeddings for elasticsearch (#11002)
- **Description:** Provide a way to use different text for embedding.
- For example, if you are ingesting stack-overflow Q&As for RAG, you
would want to embed the questions and return the answer(s) for the hits.
With this change, the consumer of langchain can implement that easily.
- I noticed the similar function is added on faiss.py with #1912 which
was for performance reason, but I see the same function can be used to
achieve what I thought. So instead of changing Document class to have
embedding_content, I mimicked the implementation of faiss.py.
- The test should provide some guidance on how to use it. It would be
more intuitive if I just pass texts and embedding_texts as separate
arguments, but I chose to use `zip`-ed object for the consistency with
faiss.py implementation.
      - I plan to make similar pull request for OpenSearch.
  - **Issue:** N/A
  - **Dependencies:** None other than the existing ones.

Co-authored-by: Bagatur <baskaryan@gmail.com>
2023-10-19 09:43:51 -07:00
Bagatur
76d3afaef0
bump 318 (#12030) 2023-10-19 09:33:39 -07:00
Dmitry Tyumentsev
5dd2161c4b
add _acall method to YandexGPT (#12029)
- **Description:** Add async support for YandexGPT LLM model

Co-authored-by: Dmitry Tyumentsev <dmitry.tyumentsev@raftds.com>
2023-10-19 09:15:26 -07:00
Peter Krenesky
8425f33363
Pydantic v2 support for OpenAPI Specs (#11936)
- **Description:** Adding Pydantic v2 support for OpenAPI Specs 

- **Issue:**
- OpenAPI spec support was disabled because `openapi-schema-pydantic`
doesn't support Pydantic v2:
     #9205
     
     - Caused errors in `get_openapi_chain`
   
    - This may be the cause of #9520.

- **Tag maintainer:** @eyurtsev
- **Twitter handle:** kreneskyp


The root cause was that `openapi-schema-pydantic` hasn't been updated in
some time but
[openapi-pydantic](https://github.com/mike-oakley/openapi-pydantic)
forked and updated the project.
2023-10-19 11:06:11 -04:00
Joe McElroy
c9f1768cb9
Elasticsearch Query Retriever: Use match + fuzziness for LIKE (#12023)
Updated the elasticsearch self query retriever to use the match clause
for LIKE operator instead of the non-analyzed fuzzy search clause.

Other small updates include:
- fixing the stack inference integration test where the index's default
pipeline didn't use the inference pipeline created
- adding a user-agent to the old implementation to track usage
- improved the documentation for ElasticsearchStore filters
2023-10-19 09:47:21 -04:00
Nuno Campos
7db6aabf65
Update chat model output type (#11833)
---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
2023-10-19 00:55:15 -07:00
Simon Dai
ed62984cb2
update Weaviate to support multi tenancy (#11842)
- **Description:** update Weaviate to support multi tenancy
  - **Issue:** 9956
  - **Dependencies:** 
  - **Tag maintainer:** hwchase17
  - **Twitter handle:** dsx1986_
2023-10-19 00:49:30 -07:00
hiigao
f818ec49b8
Encapsulate alicloud pai-eas access method for chatmodels and llms (#11852)
### Description: 
To provide an eas llm service access methods in this pull request by
impletementing `PaiEasEndpoint` and `PaiEasChatEndpoint` classes in
`langchain.llms` and `langchain.chat_models` modules. Base on this pr,
langchain users can build up a chain to call remote eas llm service and
get the llm inference results.

### About EAS Service
EAS is a Alicloud product on Alibaba Cloud Machine Learning Platform for
AI which is short for AliCloud PAI. EAS provides model inference
deployment services for the users. We build up a llm inference services
on EAS with a general llm docker images. Therefore, end users can
quickly setup their llm remote instances to load majority of the
hugginface llm models, and serve as a backend for most of the llm apps.

### Dependencies
This pr does't involve any new dependencies.

---------

Co-authored-by: 子洪 <gaoyihong.gyh@alibaba-inc.com>
2023-10-19 00:20:18 -07:00
John Mai
a6b483dcbc
Supported RetryOutputParser & RetryWithErrorOutputParser max_retries (#11903)
Description: Supported RetryOutputParser & RetryWithErrorOutputParser
max_retries
- max_retries: Maximum number of retries to parser.

Issue: None
Dependencies: None
Tag maintainer: @baskaryan 
Twitter handle:
2023-10-18 23:57:16 -07:00
Hugues Chocart
008c7df80d
[LLMonitorCallbackHandler] Refactor + add llmonitor-py dependency (#11948)
We now require uses to have the pip package `llmonitor` installed. It
allows us to have cleaner code and avoid duplicates between our library
and our code in Langchain.
2023-10-18 23:54:10 -07:00
Sian Cao
77fc2f7644
fix: impl missing embeddings method (#10823)
FAISS does not implement embeddings method and use embed_query to
embedding texts which is wrong for some embedding models.

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
2023-10-18 23:51:28 -07:00
Holt Skinner
2661dc94f3
feat: Google Vertex AI Search Retriever - Add support for Website Data Stores (#11736)
- Only works for Data stores with Advanced Website Indexing
-
https://cloud.google.com/generative-ai-app-builder/docs/about-advanced-features
- Minor restructuring - Follow up to #10513
- Remove outdated docs (readded in
https://github.com/langchain-ai/langchain/pull/11620)
  - Move legacy class into new py file to clean up the directory
- Shouldn't cause backwards compatibility issues as the import works the
same way for users
2023-10-18 23:41:48 -07:00
Shorthills AI
4b6fdd7bf0
Update modal.py (#11588)
feat: Raise KeyError when 'prompt' key is missing in JSON response

This commit updates the error handling in the code to raise a KeyError
when the 'prompt' key is not found in the JSON response. This change
makes the code more explicit about the nature of the error, helping to
improve clarity and debugging.

@baskaryan, @eyurtsev.
2023-10-18 23:40:37 -07:00
William FH
dfb4baa3f9
Fix Fireworks Callbacks (#12003)
I may be missing something but it seems like we inappropriately overrode
the 'stream()' method, losing callbacks in the process. I don't think
(?) it gave us anything in this case to customize it here?

See new trace:

https://smith.langchain.com/public/fbb82825-3a16-446b-8207-35622358db3b/r

and confirmed it streams.

Also fixes the stopwords issues from #12000
2023-10-18 23:33:09 -07:00
Wang Wei
e26559f512
Add ERNIE-Bot-4 model support for ErnieBotChat. (#11969)
- **Description:** According to the document
https://cloud.baidu.com/doc/WENXINWORKSHOP/s/clntwmv7t, add ERNIE-Bot-4
model support for ErnieBotChat.
- **Dependencies:** Before using the ERNIE-Bot-4, you should have the
model's access authority.
2023-10-18 14:55:29 -07:00