-
Notifications
You must be signed in to change notification settings - Fork 820
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
docs: fix broken links #2042
docs: fix broken links #2042
Conversation
/azp run |
Hey @JessicaXYWang 👋! We use semantic commit messages to streamline the release process. Examples of commit messages with semantic prefixes:
To test your commit locally, please follow our guild on building from source. |
Azure Pipelines successfully started running 1 pipeline(s). |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Summary by GPT-4
In this notebook, we'll demonstrate how to develop a context-aware question answering framework for any form of a document using OpenAI models, SynapseML and Azure AI Services. In this notebook, we assume that PDF documents are the source of data, however, the same framework can be easily extended to other document formats too.
We’ll cover the following key steps:
- Preprocessing PDF Documents: Learn how to load the PDF documents into a Spark DataFrame, read the documents using the Azure AI Document Intelligence in Azure AI Services, and use SynapseML to split the documents into chunks.
- Embedding Generation and Storage: Learn how to generate embeddings for the chunks using SynapseML and Azure OpenAI Services, store the embeddings in a vector store using Azure Cognitive Search, and search the vector store to answer the user’s question.
- Question Answering Pipeline: Learn how to retrieve relevant document based on the user’s question and provide the answer using Langchain.
We utilize SynapseML, an ecosystem of tools designed to enhance the distributed computing framework Apache Spark. SynapseML introduces advanced networking capabilities to the Spark ecosystem and offers user-friendly SparkML transformers for various Azure AI Services.
Additionally, we employ AnalyzeDocument from Azure AI Services to extract the complete document content and present it in designated columns called "output_content" and "paragraph."
To produce embeddings for each chunk, we utilize both SynapseML and Azure OpenAI Service. By integrating Azure OpenAI service with SynapseML, we can leverage Apache Spark's distributed computing framework power to process numerous prompts using OpenAI service. This integration enables SynapseML embedding client to generate embeddings in a distributed manner, enabling efficient processing of large volumes of data. If you're interested in applying large language models at a distributed scale using Azure OpenAI and Azure Synapse Analytics, you can refer to this approach. For more detailed information on generating embeddings with Azure OpenAI, you can look here.
Azure Cognitive Search offers a user-friendly interface for creating a vector database, as well as storing and retrieving data using vector search. If you're interested in learning more about vector search, you can look here.
After processing the document, we can proceed to pose a question. We will use SynapseML to convert users' questions into an embedding and then utilize cosine similarity to retrieve top K document chunks that closely match users' questions. It's worth mentioning that alternative similarity metrics can also be employed.
Suggestions
No suggestions are needed as the changes in this PR are clear and accurate.
TODO (not this PR): [build.sbt] Line 501 in cde6834
[pull_request_template]
[website channel manifest]
|
Codecov Report
@@ Coverage Diff @@
## master #2042 +/- ##
===========================================
- Coverage 87.07% 70.46% -16.62%
===========================================
Files 306 306
Lines 16063 16063
Branches 852 852
===========================================
- Hits 13987 11318 -2669
- Misses 2076 4745 +2669 |
/azp run |
Azure Pipelines successfully started running 1 pipeline(s). |
* fix broken links * fix broken links
Related Issues/PRs
#xxx
What changes are proposed in this pull request?
Briefly describe the changes included in this Pull Request.
How is this patch tested?
Does this PR change any dependencies?
Does this PR add a new feature? If so, have you added samples on website?
website/docs/documentation
folder.Make sure you choose the correct class
estimators/transformers
and namespace.DocTable
points to correct API link.yarn run start
to make sure the website renders correctly.<!--pytest-codeblocks:cont-->
before each python code blocks to enable auto-tests for python samples.WebsiteSamplesTests
job pass in the pipeline.