Skip to content

RetrievalAugmentedGenerationUsingChatGPTandMATLAB.md

Latest commit

 

History

History
183 lines (135 loc) · 6.37 KB

RetrievalAugmentedGenerationUsingChatGPTandMATLAB.md

File metadata and controls

183 lines (135 loc) · 6.37 KB

Retrieval-Augmented Generation Using ChatGPT and MATLAB

To run the code shown on this page, open the MLX file in MATLAB®: mlx-scripts/RetrievalAugmentedGenerationUsingChatGPTandMATLAB.mlx

This example shows how to use retrieval-augmented generation to generate answers to queries based on information contained in a document corpus.

The example contains three steps:

  • Download and preprocess documents.
  • Find documents relevant to a query using keyword search.
  • Generate a response using ChatGPT™ based on the both the query and the most relevant source document.

This example requires Text Analytics Toolbox™.

To run this example, you need a valid API key from a paid OpenAI™ API account.

loadenv(".env")

Download and Preprocess Documents

Specify the URLs of the reports.

url = ["https://openknowledge.worldbank.org/bitstreams/0c18c872-91f0-51a4-ba91-c36b98893b4a/download"
    "https://openknowledge.worldbank.org/bitstreams/476f037b-a17e-484f-9cc2-282a2e5a929f/download"
    "https://openknowledge.worldbank.org/bitstreams/0c18c872-91f0-51a4-ba91-c36b98893b4a/download"];

Define the local path where the reports will be saved and download the reports using the provided URLs and save them to the specified local path.

localpath = "./data/";
if ~exist(localpath, "dir")
    mkdir(localpath);
end
numFiles = numel(url);
for i = 1:numFiles
    filename = "WBD_" + i + ".pdf";
    local_file_name = fullfile(localpath, filename);
    if ~exist(local_file_name,"file")
        websave(local_file_name, url{i}, weboptions(Timeout=30));
    end
end

Define the function to read the text from the downloaded files.

readFcn = @extractFileText;
file_pattern = [".txt",".pdf",".docx",".html",".htm"];
fds = fileDatastore(localpath,'FileExtensions',file_pattern,'ReadFcn',readFcn);

str = readall(fds);
str = [str{:}];

Split the text data into paragraphs with the helper function preprocessDocuments.

documents = preprocessDocuments(str);

Initialize the chatbot with a system prompt and API key. Include your API key in the environment variable OPENAI_API_KEY or pass your key using the APIKey name-value pair. This example uses the model GPT-4.1 mini.

chat = openAIChat("You are a helpful assistant. You will get a " + ...
    "context for each question, but only use the information " + ...
    "in the context if that makes sense to answer the question. " + ...
    "Let's think step-by-step, explaining how you reached the answer.", ...
    ModelName="gpt-4.1-mini");

Retrieve Relevant Documents

Define the query, then retrieve and filter the relevant documents based on the query.

query = "What technical criteria can be used to streamline new approvals for grid-friendly DPV?";

Tokenize the query and find similarity scores between the query and documents.

embQuery = bm25Similarity(documents, tokenizedDocument(query));

Sort the documents in descending order of similarity scores.

[~, idx] = sort(embQuery, "descend");
limitWords = 1000;
selectedDocs = [];
totalWords = 0;

Iterate over sorted document indices until word limit is reached

i = 1;
while totalWords <= limitWords && i <= length(idx)
    totalWords = totalWords + doclength(documents(idx(i)));
    selectedDocs = [selectedDocs; joinWords(documents(idx(i)))];
    i = i + 1;
end

Generate Response

Define the prompt for the chatbot and generate a response.

prompt = "Context:" + join(selectedDocs, " ") + newline + ...
    "Answer the following question: " + query;
response = generate(chat, prompt);

Wrap the text for easier visualization.

wrapText(response)
ans = 
    "To streamline new approvals for grid-friendly distributed photovoltaics (DPV), 
     prudent technical criteria can be applied that balance deployment with the 
     technical requirements of the distribution system. Specifically:
     
1. **Technical Screening Criteria**: Utilities can use technical screening
     criteria based on metrics relevant to the impact of DPV on low-voltage 
     distribution grids. For example, limiting DPV capacity relative to the minimum 
     feeder daytime load helps prevent technical issues.
     
     2. **Hosting Capacity Calculations**: These calculations estimate the point at 
     which DPV penetration would induce technical impacts on system operation. 
     Screening criteria derived from these calculations allow utilities to set 
     limits that ensure grid stability.
     
     3. **Connection Size Limits**: Some regulations specify maximum DPV 
     installation sizes relative to customer peak load (e.g., 80 percent of peak 
     load in India), serving as a straightforward technical criterion for approval.
     
     4. **Advanced Inverter Functions**: Specifying that inverters have advanced 
     capabilities such as reactive power control and active power curtailment before 
     connection approval helps maintain grid stability and voltage management.
     
     5. **Grid Codes Reflecting DPV Growth**: Adopting or updating grid codes to 
     include expected growth of distributed energy resources and requirements for 
     inverter functionalities streamlines approvals by setting clear technical 
     standards.
     
     6. **Cost-effective and Forward-looking Measures**: The criteria should be easy 
     to implement, cost-efficient, and anticipate future system integration needs, 
     helping avoid costly retrofits.
     
     Thus, prudent technical criteria include setting size limits relative to load, 
     applying hosting capacity-based screens, requiring advanced inverter 
     functionalities, and updating grid codes to reflect DPV penetration—all 
     facilitating streamlined, grid-friendly DPV approvals."

Helper Functions

function allDocs = preprocessDocuments(str)
tokenized = tokenizedDocument(join(str,[newline newline]));
allDocs = splitParagraphs(tokenized);
end

function wrappedText = wrapText(text)
s = textwrap(text,80);
wrappedText = string(join(s,newline));
end

References

Energy Sector Management Assistance Program (ESMAP). 2023. From Sun to Roof to Grid: Power Systems and Distributed PV. Technical Report. Washington, DC: World Bank. License: Creative Commons Attribution CC BY 3.0 IGO

Copyright 2024 The MathWorks, Inc.