Neo4j has added vector index integration just recently. To allow both
ingestion and integrating it as vector RAG applications, I wrapped it as
a vector store as the implementation is completely different from
`GraphCypherQAChain`. Here, we are not generating any Cypher statements
at query time, we are simply doing the vector similarity search using
the new vector index as if we were dealing with a vector database.
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
Update google drive doc loader and retriever notebooks. Show how to use with langchain-googledrive package.
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
Fixed title for the `extras/integrations/llms/llm_caching.ipynb`.
Existing title breaks the sorted order of items in the navbar.
Updated some formatting.
* Added links to the AI Network
* Made title consistent to other tool kits
* Added `integrations/providers/` integration card page
* **No changes** in the example code!
- Fixed a broken link in the `integrations/providers/infino.mdx`
- Fixed a title in the `integration/collbacks/infino.ipynb` example
- Updated text format in this example.
## Description
The following PR enables the [grammar-based
sampling](https://github.com/ggerganov/llama.cpp/tree/master/grammars)
in llama-cpp LLM.
In short, loading file with formal grammar definition will constrain
model outputs. For instance, one can force the model to generate valid
JSON or generate only python lists.
In the follow-up PR we will add:
* docs with some description why it is cool and how it works
* maybe some code sample for some task such as in llama repo
---------
Co-authored-by: Lance Martin <lance@langchain.dev>
Co-authored-by: Bagatur <baskaryan@gmail.com>
### Description
The previous Redis implementation did not allow for the user to specify
the index configuration (i.e. changing the underlying algorithm) or add
additional metadata to use for querying (i.e. hybrid or "filtered"
search).
This PR introduces the ability to specify custom index attributes and
metadata attributes as well as use that metadata in filtered queries.
Overall, more structure was introduced to the Redis implementation that
should allow for easier maintainability moving forward.
# New Features
The following features are now available with the Redis integration into
Langchain
## Index schema generation
The schema for the index will now be automatically generated if not
specified by the user. For example, the data above has the multiple
metadata categories. The the following example
```python
from langchain.embeddings import OpenAIEmbeddings
from langchain.vectorstores.redis import Redis
embeddings = OpenAIEmbeddings()
rds, keys = Redis.from_texts_return_keys(
texts,
embeddings,
metadatas=metadata,
redis_url="redis://localhost:6379",
index_name="users"
)
```
Loading the data in through this and the other ``from_documents`` and
``from_texts`` methods will now generate index schema in Redis like the
following.
view index schema with the ``redisvl`` tool. [link](redisvl.com)
```bash
$ rvl index info -i users
```
Index Information:
| Index Name | Storage Type | Prefixes | Index Options | Indexing |
|--------------|----------------|---------------|-----------------|------------|
| users | HASH | ['doc:users'] | [] | 0 |
Index Fields:
| Name | Attribute | Type | Field Option | Option Value |
|----------------|----------------|---------|----------------|----------------|
| user | user | TEXT | WEIGHT | 1 |
| job | job | TEXT | WEIGHT | 1 |
| credit_score | credit_score | TEXT | WEIGHT | 1 |
| content | content | TEXT | WEIGHT | 1 |
| age | age | NUMERIC | | |
| content_vector | content_vector | VECTOR | | |
### Custom Metadata specification
The metadata schema generation has the following rules
1. All text fields are indexed as text fields.
2. All numeric fields are index as numeric fields.
If you would like to have a text field as a tag field, users can specify
overrides like the following for the example data
```python
# this can also be a path to a yaml file
index_schema = {
"text": [{"name": "user"}, {"name": "job"}],
"tag": [{"name": "credit_score"}],
"numeric": [{"name": "age"}],
}
rds, keys = Redis.from_texts_return_keys(
texts,
embeddings,
metadatas=metadata,
redis_url="redis://localhost:6379",
index_name="users"
)
```
This will change the index specification to
Index Information:
| Index Name | Storage Type | Prefixes | Index Options | Indexing |
|--------------|----------------|----------------|-----------------|------------|
| users2 | HASH | ['doc:users2'] | [] | 0 |
Index Fields:
| Name | Attribute | Type | Field Option | Option Value |
|----------------|----------------|---------|----------------|----------------|
| user | user | TEXT | WEIGHT | 1 |
| job | job | TEXT | WEIGHT | 1 |
| content | content | TEXT | WEIGHT | 1 |
| credit_score | credit_score | TAG | SEPARATOR | , |
| age | age | NUMERIC | | |
| content_vector | content_vector | VECTOR | | |
and throw a warning to the user (log output) that the generated schema
does not match the specified schema.
```text
index_schema does not match generated schema from metadata.
index_schema: {'text': [{'name': 'user'}, {'name': 'job'}], 'tag': [{'name': 'credit_score'}], 'numeric': [{'name': 'age'}]}
generated_schema: {'text': [{'name': 'user'}, {'name': 'job'}, {'name': 'credit_score'}], 'numeric': [{'name': 'age'}]}
```
As long as this is on purpose, this is fine.
The schema can be defined as a yaml file or a dictionary
```yaml
text:
- name: user
- name: job
tag:
- name: credit_score
numeric:
- name: age
```
and you pass in a path like
```python
rds, keys = Redis.from_texts_return_keys(
texts,
embeddings,
metadatas=metadata,
redis_url="redis://localhost:6379",
index_name="users3",
index_schema=Path("sample1.yml").resolve()
)
```
Which will create the same schema as defined in the dictionary example
Index Information:
| Index Name | Storage Type | Prefixes | Index Options | Indexing |
|--------------|----------------|----------------|-----------------|------------|
| users3 | HASH | ['doc:users3'] | [] | 0 |
Index Fields:
| Name | Attribute | Type | Field Option | Option Value |
|----------------|----------------|---------|----------------|----------------|
| user | user | TEXT | WEIGHT | 1 |
| job | job | TEXT | WEIGHT | 1 |
| content | content | TEXT | WEIGHT | 1 |
| credit_score | credit_score | TAG | SEPARATOR | , |
| age | age | NUMERIC | | |
| content_vector | content_vector | VECTOR | | |
### Custom Vector Indexing Schema
Users with large use cases may want to change how they formulate the
vector index created by Langchain
To utilize all the features of Redis for vector database use cases like
this, you can now do the following to pass in index attribute modifiers
like changing the indexing algorithm to HNSW.
```python
vector_schema = {
"algorithm": "HNSW"
}
rds, keys = Redis.from_texts_return_keys(
texts,
embeddings,
metadatas=metadata,
redis_url="redis://localhost:6379",
index_name="users3",
vector_schema=vector_schema
)
```
A more complex example may look like
```python
vector_schema = {
"algorithm": "HNSW",
"ef_construction": 200,
"ef_runtime": 20
}
rds, keys = Redis.from_texts_return_keys(
texts,
embeddings,
metadatas=metadata,
redis_url="redis://localhost:6379",
index_name="users3",
vector_schema=vector_schema
)
```
All names correspond to the arguments you would set if using Redis-py or
RedisVL. (put in doc link later)
### Better Querying
Both vector queries and Range (limit) queries are now available and
metadata is returned by default. The outputs are shown.
```python
>>> query = "foo"
>>> results = rds.similarity_search(query, k=1)
>>> print(results)
[Document(page_content='foo', metadata={'user': 'derrick', 'job': 'doctor', 'credit_score': 'low', 'age': '14', 'id': 'doc:users:657a47d7db8b447e88598b83da879b9d', 'score': '7.15255737305e-07'})]
>>> results = rds.similarity_search_with_score(query, k=1, return_metadata=False)
>>> print(results) # no metadata, but with scores
[(Document(page_content='foo', metadata={}), 7.15255737305e-07)]
>>> results = rds.similarity_search_limit_score(query, k=6, score_threshold=0.0001)
>>> print(len(results)) # range query (only above threshold even if k is higher)
4
```
### Custom metadata filtering
A big advantage of Redis in this space is being able to do filtering on
data stored alongside the vector itself. With the example above, the
following is now possible in langchain. The equivalence operators are
overridden to describe a new expression language that mimic that of
[redisvl](redisvl.com). This allows for arbitrarily long sequences of
filters that resemble SQL commands that can be used directly with vector
queries and range queries.
There are two interfaces by which to do so and both are shown.
```python
>>> from langchain.vectorstores.redis import RedisFilter, RedisNum, RedisText
>>> age_filter = RedisFilter.num("age") > 18
>>> age_filter = RedisNum("age") > 18 # equivalent
>>> results = rds.similarity_search(query, filter=age_filter)
>>> print(len(results))
3
>>> job_filter = RedisFilter.text("job") == "engineer"
>>> job_filter = RedisText("job") == "engineer" # equivalent
>>> results = rds.similarity_search(query, filter=job_filter)
>>> print(len(results))
2
# fuzzy match text search
>>> job_filter = RedisFilter.text("job") % "eng*"
>>> results = rds.similarity_search(query, filter=job_filter)
>>> print(len(results))
2
# combined filters (AND)
>>> combined = age_filter & job_filter
>>> results = rds.similarity_search(query, filter=combined)
>>> print(len(results))
1
# combined filters (OR)
>>> combined = age_filter | job_filter
>>> results = rds.similarity_search(query, filter=combined)
>>> print(len(results))
4
```
All the above filter results can be checked against the data above.
### Other
- Issue: #3967
- Dependencies: No added dependencies
- Tag maintainer: @hwchase17 @baskaryan @rlancemartin
- Twitter handle: @sampartee
---------
Co-authored-by: Naresh Rangan <naresh.rangan0@walmart.com>
Co-authored-by: Bagatur <baskaryan@gmail.com>
This adds Xata as a memory store also to the python version of
LangChain, similar to the [one for
LangChain.js](https://github.com/hwchase17/langchainjs/pull/2217).
I have added a Jupyter Notebook with a simple and a more complex example
using an agent.
To run the integration test, you need to execute something like:
```
XATA_API_KEY='xau_...' XATA_DB_URL="https://demo-uni3q8.eu-west-1.xata.sh/db/langchain" poetry run pytest tests/integration_tests/memory/test_xata.py
```
Where `langchain` is the database you create in Xata.
Still working out interface/notebooks + need discord data dump to test
out things other than copy+paste
Update:
- Going to remove the 'user_id' arg in the loaders themselves and just
standardize on putting the "sender" arg in the extra kwargs. Then can
provide a utility function to map these to ai and human messages
- Going to move the discord one into just a notebook since I don't have
a good dump to test on and copy+paste maybe isn't the greatest thing to
support in v0
- Need to do more testing on slack since it seems the dump only includes
channels and NOT 1 on 1 convos
-
---------
Co-authored-by: Harrison Chase <hw.chase.17@gmail.com>
The Graph Chains are different in the way that it uses two LLMChains
instead of one like the retrievalQA chains. Therefore, sometimes you
want to use different LLM to generate the database query and to generate
the final answer.
This feature would make it more convenient to use different LLMs in the
same chain.
I have also renamed the Graph DB QA Chain to Neo4j DB QA Chain in the
documentation only as it is used only for Neo4j. The naming was
ambigious as it was the first graphQA chain added and wasn't sure how do
you want to spin it.
Uses the shorter import path
`from langchain.document_loaders import` instead of the full path
`from langchain.document_loaders.assemblyai`
Applies those changes to the docs and the unit test.
See #9667 that adds this new loader.
Note: There are no changes in the file names!
- The group name on the main navbar changed: `Agent toolkits` -> `Agents
& Toolkits`. Examples here are the mix of the Agent and Toolkit examples
because Agents and Toolkits in examples are always used together.
- Titles changed: removed "Agent" and "Toolkit" suffixes. The reason is
the same.
- Formatting: mostly cleaning the header structure, so it could be
better on the right-side navbar.
Main navbar is looking much cleaner now.
This PR adds a new document loader `AssemblyAIAudioTranscriptLoader`
that allows to transcribe audio files with the [AssemblyAI
API](https://www.assemblyai.com) and loads the transcribed text into
documents.
- Add new document_loader with class `AssemblyAIAudioTranscriptLoader`
- Add optional dependency `assemblyai`
- Add unit tests (using a Mock client)
- Add docs notebook
This is the equivalent to the JS integration already available in
LangChain.js. See the [LangChain JS docs AssemblyAI
page](https://js.langchain.com/docs/modules/data_connection/document_loaders/integrations/web_loaders/assemblyai_audio_transcription).
At its simplest, you can use the loader to get a transcript back from an
audio file like this:
```python
from langchain.document_loaders.assemblyai import AssemblyAIAudioTranscriptLoader
loader = AssemblyAIAudioTranscriptLoader(file_path="./testfile.mp3")
docs = loader.load()
```
To use it, it needs the `assemblyai` python package installed, and the
environment variable `ASSEMBLYAI_API_KEY` set with your API key.
Alternatively, the API key can also be passed as an argument.
Twitter handles to shout out if so kindly 🙇
[@AssemblyAI](https://twitter.com/AssemblyAI) and
[@patloeber](https://twitter.com/patloeber)
---------
Co-authored-by: Bagatur <22008038+baskaryan@users.noreply.github.com>
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
Improve internal consistency in LangChain documentation
- Change occurrences of eg and eg. to e.g.
- Fix headers containing unnecessary capital letters.
- Change instances of "few shot" to "few-shot".
- Add periods to end of sentences where missing.
- Minor spelling and grammar fixes.
This PR introduces a persistence layer to help with indexing workflows
into
vectostores.
The indexing code helps users to:
1. Avoid writing duplicated content into the vectostore
2. Avoid over-writing content if it's unchanged
Importantly, this keeps on working even if the content being written is
derived
via a set of transformations from some source content (e.g., indexing
children
documents that were derived from parent documents by chunking.)
The two main components are:
1. Persistence layer that keeps track of which keys were updated and
when.
Keeping track of the timestamp of updates, allows to clean up old
content
safely, and with minimal complexity.
2. HashedDocument which is used to hash the contents (including
metadata) of
the documents. We rely on the hashes for identifying duplicates.
The indexing code works with **ANY** document loader. To add
transformations
to the documents, users for now can add a custom document loader
that composes an existing loader together with document transformers.
---------
Co-authored-by: Bagatur <baskaryan@gmail.com>
The Docugami loader was not returning the source metadata key. This was
triggering this exception when used with retrievers, per
https://github.com/langchain-ai/langchain/blob/master/libs/langchain/langchain/schema/prompt_template.py#L193C1-L195C41
The fix is simple and just updates the metadata key name for the
document each chunk is sourced from, from "name" to "source" as
expected.
I tested by running the python notebook that has an end to end scenario
in it.
Tagging DataLoader maintainers @rlancemartin @eyurtsev
This pull request corrects the URL links in the Async API documentation
to align with the updated project layout. The links had not been updated
despite the changes in layout.
Not obvious what the error is when you cannot index. This pr adds the
ability to log the first errors reason, to help the user diagnose the
issue.
Also added some more documentation for when you want to use the
vectorstore with an embedding model deployed in elasticsearch.
Credit: @elastic and @phoey1
