forked from Archives/langchain
You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
72 lines
3.9 KiB
ReStructuredText
72 lines
3.9 KiB
ReStructuredText
Chains
|
|
======
|
|
|
|
The examples here are all end-to-end chains for specific applications.
|
|
A chain is made up of links, which can be either primitives or other chains.
|
|
|
|
The following primitives exist as options to use for links:
|
|
|
|
#. `LLM: <../modules/llms.rst>`_ A language model takes text as input and outputs text.
|
|
#. `PromptTemplate: <../modules/prompt.rst>`_ A prompt template takes arbitrary string inputs and returns a final formatted string.
|
|
#. `TextSplitter: <../modules/text_splitter.rst>`_ A text splitter takes a longer document and splits it into smaller chunks.
|
|
#. `Python REPL: <../modules/python.rst>`_ A Python REPL takes a string representing a Python command to run, runs that command, and then returns anything that was printed during that run.
|
|
#. `SQL Database: <../modules/sql_database.rst>`_ A SQL database takes a string representing a SQL command as input and executes that command against the database. If any rows are returned, then those are cast to a string and returned.
|
|
#. `Search: <../modules/serpapi.rst>`_ A search object takes a string as input and executes that against a search object, returning any results.
|
|
#. `Docstore: <../modules/docstore.rst>`_ A docstore object can be used to lookup a document in a database by exact match.
|
|
#. `Vectorstore: <../modules/vectorstore.rst>`_ A vectorstore object uses embeddings stored in a vector database to take in an input string and return documents similar to that string.
|
|
|
|
With these primitives in mind, the following chains exist:
|
|
|
|
**LLMChain**
|
|
|
|
- **Links Used**: PromptTemplate, LLM
|
|
- **Notes**: This chain is the simplest chain, and is widely used by almost every other chain. This chain takes arbitrary user input, creates a prompt with it from the PromptTemplate, passes that to the LLM, and then returns the output of the LLM as the final output.
|
|
- `Example Notebook <chains/llm_chain.ipynb>`_
|
|
|
|
**LLMMath**
|
|
|
|
- **Links Used**: Python REPL, LLMChain
|
|
- **Notes**: This chain takes user input (a math question), uses an LLMChain to convert it to python code snippet to run in the Python REPL, and then returns that as the result.
|
|
- `Example Notebook <chains/llm_math.ipynb>`_
|
|
|
|
**PAL**
|
|
|
|
- **Links Used**: Python REPL, LLMChain
|
|
- **Notes**: This chain takes user input (a reasoning question), uses an LLMChain to convert it to python code snippet to run in the Python REPL, and then returns that as the result.
|
|
- `Paper <https://arxiv.org/abs/2211.10435>`_
|
|
- `Example Notebook <chains/pal.ipynb>`_
|
|
|
|
**Recursive Summarization**
|
|
|
|
- **Links Used**: TextSplitter, LLMChain
|
|
- **Notes**: This chain splits a document into chunks, runs a first LLMChain over each chunk to summarize it, and then runs a second LLMChain over those results to get a summary of the summaries.
|
|
- `Example Notebook <chains/map_reduce.ipynb>`_
|
|
|
|
**SQLDatabase Chain**
|
|
|
|
- **Links Used**: SQLDatabase, LLMChain
|
|
- **Notes**: This chain takes user input (a question), uses a first LLM chain to construct a SQL query to run against the SQL database, and then uses another LLMChain to take the results of that query and use it to answer the original question.
|
|
- `Example Notebook <chains/sqlite.ipynb>`_
|
|
|
|
|
|
**Vector Database Question-Answering**
|
|
|
|
- **Links Used**: Vectorstore, LLMChain
|
|
- **Notes**: This chain takes user input (a question), uses the Vectorstore and semantic search to find relevant documents, and then passes the documents plus to the original question to another LLM to generate a final answer.
|
|
- `Example Notebook <chains/vector_db_qa.ipynb>`_
|
|
|
|
**Question-Answering With Sources**
|
|
|
|
- **Links Used**: LLMChain
|
|
- **Notes**: This chain takes a question and multiple documents as input. It then runs a first LLMChain over all documents attempting to answer the provided question. It then runs a second LLMChain over the results of the first pass, combining the answers from documents into a single response that is returned.
|
|
- `Example Notebook <chains/combine_documents.ipynb>`_
|
|
|
|
|
|
.. toctree::
|
|
:maxdepth: 1
|
|
:glob:
|
|
:caption: Chains
|
|
:hidden:
|
|
|
|
chains/*
|