From c213ab300e163c8d21f974dfc76b4d0f7df68f47 Mon Sep 17 00:00:00 2001 From: Sanjana Reddy Date: Thu, 7 Nov 2024 18:14:01 +0000 Subject: [PATCH] update hyperlinks --- .../vertex_genai/labs/retrieval_augmented_generation.ipynb | 4 ++-- .../solutions/retrieval_augmented_generation.ipynb | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/notebooks/vertex_genai/labs/retrieval_augmented_generation.ipynb b/notebooks/vertex_genai/labs/retrieval_augmented_generation.ipynb index f5cb0657..e8285557 100644 --- a/notebooks/vertex_genai/labs/retrieval_augmented_generation.ipynb +++ b/notebooks/vertex_genai/labs/retrieval_augmented_generation.ipynb @@ -365,7 +365,7 @@ "source": [ "#### Document Loading\n", "\n", - "Langchain provides classes to load data from different sources. Some useful data loaders are [Google Cloud Storage Directory Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/google_cloud_storage_directory), [Google Drive Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/google_drive), [Recursive URL Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/recursive_url_loader), [PDF Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/how_to/pdf), [JSON Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/how_to/json), [Wikipedia Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/wikipedia), and [more](https://python.langchain.com/docs/modules/data_connection/document_loaders/).\n", + "Langchain provides classes to load data from different sources. Some useful data loaders are [Google Cloud Storage Directory Loader](https://python.langchain.com/docs/integrations/document_loaders/google_cloud_storage_directory/), [Google Drive Loader](https://python.langchain.com/docs/integrations/document_loaders/google_drive), [Recursive URL Loader](https://python.langchain.com/docs/integrations/document_loaders/recursive_url/), [PDF Loader](https://python.langchain.com/docs/integrations/document_loaders/#pdfs), [JSON Loader](https://python.langchain.com/docs/integrations/document_loaders/json/), [Wikipedia Loader](https://python.langchain.com/docs/integrations/document_loaders/wikipedia/), and [more](https://python.langchain.com/docs/integrations/document_loaders/).\n", "\n", "In this notebook we will use the Wikipedia loader to create a private knowledge base of wikipedia articles about large language models, but the overall process is similiar regardless of which document loader you use." ] @@ -393,7 +393,7 @@ "#### Split text into chunks \n", "Now that we have the documents we will split them into chunks. Each chunk will become one vector in the vector store. To do this we will define a chunk size (number of characters) and a chunk overlap (amount of overlap i.e. sliding window). The perfect chunk size can be difficult to determine. Too large of a chunk size leads to too much information per chunk (individual chunks not specific enough), however too small of a chunk size leads to not enough information per chunk. In both cases, nearest neighbors lookup with a query/question embedding may struggle to retrieve the actually relevant chunks, or fail altogether if the chunks are too large to use as context with an LLM query.\n", "\n", - "In this notebook we will use a chunk size of 800 chacters and a chunk overlap of 400 characters, but feel free to experiment with other sizes! Note: you can specify a custom `length_function` with `RecursiveCharacterTextSplitter` if you want chunk size/overlap to be determined by something other than Python's len function. In addition to `RecursiveCharacterTextSplitter`, there are [other text splitters](https://python.langchain.com/docs/modules/data_connection/document_transformers/) you can consider." + "In this notebook we will use a chunk size of 800 chacters and a chunk overlap of 400 characters, but feel free to experiment with other sizes! Note: you can specify a custom `length_function` with `RecursiveCharacterTextSplitter` if you want chunk size/overlap to be determined by something other than Python's len function. In addition to `RecursiveCharacterTextSplitter`, [other text splitters](https://python.langchain.com/docs/how_to/#text-splitters) you can consider." ] }, { diff --git a/notebooks/vertex_genai/solutions/retrieval_augmented_generation.ipynb b/notebooks/vertex_genai/solutions/retrieval_augmented_generation.ipynb index aeaaddd1..3d32ff1a 100644 --- a/notebooks/vertex_genai/solutions/retrieval_augmented_generation.ipynb +++ b/notebooks/vertex_genai/solutions/retrieval_augmented_generation.ipynb @@ -352,7 +352,7 @@ "source": [ "#### Document Loading\n", "\n", - "Langchain provides classes to load data from different sources. Some useful data loaders are [Google Cloud Storage Directory Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/google_cloud_storage_directory), [Google Drive Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/google_drive), [Recursive URL Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/recursive_url_loader), [PDF Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/how_to/pdf), [JSON Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/how_to/json), [Wikipedia Loader](https://python.langchain.com/docs/modules/data_connection/document_loaders/integrations/wikipedia), and [more](https://python.langchain.com/docs/modules/data_connection/document_loaders/).\n", + "Langchain provides classes to load data from different sources. Some useful data loaders are [Google Cloud Storage Directory Loader](https://python.langchain.com/docs/integrations/document_loaders/google_cloud_storage_directory/), [Google Drive Loader](https://python.langchain.com/docs/integrations/document_loaders/google_drive), [Recursive URL Loader](https://python.langchain.com/docs/integrations/document_loaders/recursive_url/), [PDF Loader](https://python.langchain.com/docs/integrations/document_loaders/#pdfs), [JSON Loader](https://python.langchain.com/docs/integrations/document_loaders/json/), [Wikipedia Loader](https://python.langchain.com/docs/integrations/document_loaders/wikipedia/), and [more](https://python.langchain.com/docs/integrations/document_loaders/).\n", "\n", "In this notebook we will use the Wikipedia loader to create a private knowledge base of wikipedia articles about large language models, but the overall process is similiar regardless of which document loader you use." ] @@ -380,7 +380,7 @@ "#### Split text into chunks \n", "Now that we have the documents we will split them into chunks. Each chunk will become one vector in the vector store. To do this we will define a chunk size (number of characters) and a chunk overlap (amount of overlap i.e. sliding window). The perfect chunk size can be difficult to determine. Too large of a chunk size leads to too much information per chunk (individual chunks not specific enough), however too small of a chunk size leads to not enough information per chunk. In both cases, nearest neighbors lookup with a query/question embedding may struggle to retrieve the actually relevant chunks, or fail altogether if the chunks are too large to use as context with an LLM query.\n", "\n", - "In this notebook we will use a chunk size of 800 chacters and a chunk overlap of 400 characters, but feel free to experiment with other sizes! Note: you can specify a custom `length_function` with `RecursiveCharacterTextSplitter` if you want chunk size/overlap to be determined by something other than Python's len function. In addition to `RecursiveCharacterTextSplitter`, there are [other text splitters](https://python.langchain.com/docs/modules/data_connection/document_transformers/) you can consider." + "In this notebook we will use a chunk size of 800 chacters and a chunk overlap of 400 characters, but feel free to experiment with other sizes! Note: you can specify a custom `length_function` with `RecursiveCharacterTextSplitter` if you want chunk size/overlap to be determined by something other than Python's len function. In addition to `RecursiveCharacterTextSplitter`, there are [other text splitters](https://python.langchain.com/docs/how_to/#text-splitters) you can consider." ] }, {