Power App Pinecone and Summarize results with Azure Open AI GPT 3
5 min readJun 17, 2023
Let’s build a Power App to use Azure Open AI ChatGPT to summarize the results from Pine Cone index
What’s needed.
- Register for Azure Open AI — https://learn.microsoft.com/en-us/azure/cognitive-services/openai/overview
- Once got approved create a azure open ai resource in Azure portal
- Select region as East US
- At the time of writing this article gpt4, gpt3.5-turbo is only available in south central US
- Create a deployment inside the resource
- Create pine cone service
- Create a Index to work with
- Enable Cosine Similarity as well
Power Flow
- Let’s create a power flow
- On the left menu in power apps click on flows
- https://make.preview.powerapps.com/
- Click on flows
- Click New Flow
- Name it as pineconegpt
- here is the entire flow
- Now we need to add a trigger as Power Apps
- Now initialize a variable caleed searchtxt
- Call the embedding service to get the vector
- Set the URL as below
https://aoaoresourcename.openai.azure.com/openai/deployments/text-deploy/embeddings?api-version=2023-03-15-preview- Set the header as below
content-type:application/json
api-key: <your_api_key>- Set Body as
{
"input": @{variables('searchtxt')}
}- now parse the vecotr from the response
- here is the schema
{
"type": "object",
"properties": {
"object": {
"type": "string"
},
"data": {
"type": "array",
"items": {
"type": "object",
"properties": {
"object": {
"type": "string"
},
"index": {
"type": "integer"
},
"embedding": {
"type": "array",
"items": {
"type": "number"
}
}
},
"required": [
"object",
"index",
"embedding"
]
}
},
"model": {
"type": "string"
},
"usage": {
"type": "object",
"properties": {
"prompt_tokens": {
"type": "integer"
},
"total_tokens": {
"type": "integer"
}
}
}
}
}- now we need to call the pinecone service to get the results
- send the vector as input
- set the URL as below
https://indexname.svc.us-east1-gcp.pinecone.io/query- set the header as below
content-type:application/json
api-key: <your_api_key>“ tabindex=”0" role=”button” style=”box-sizing: border-box; position: relative; display: flex !important; padding: 0px !important; font-size: 14px; font-weight: var( — base-text-weight-medium, 500); line-height: 20px; white-space: nowrap; vertical-align: middle; cursor: pointer; user-select: none; border: 0px; border-radius: 6px; appearance: none; color: var( — color-accent-fg); background-color: transparent; box-shadow: none; transition: color 80ms cubic-bezier(0.33, 1, 0.68, 1) 0s, background-color, box-shadow, border-color; justify-content: center !important; align-items: center !important; margin: var( — base-size-8, 8px) !important; width: var( — control-small-size, 28px); height: var( — control-small-size, 28px);”>
- body
{
"includeMetadata": true,
"includeValues": true,
"namespace": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
"topK": 5,
"vector": @{body('Parse_JSON_3')?['data']?[0]?['embedding']}
}- now parse the json response
- set the schema as
{
"type": "object",
"properties": {
"results": {
"type": "array"
},
"matches": {
"type": "array",
"items": {
"type": "object",
"properties": {
"id": {
"type": "string"
},
"score": {
"type": "number"
},
"values": {
"type": "array",
"items": {
"type": "number"
}
},
"metadata": {
"type": "object",
"properties": {
"source": {
"type": "string"
},
"text": {
"type": "string"
}
}
}
},
"required": [
"id",
"score",
"values",
"metadata"
]
}
},
"namespace": {
"type": "string"
}
}
}- Now initialize a varibale
- Apply for each
- Take each text append the text to variable
- now send to gpt 2 summarization service
- set the URL as below
https://aoairesourcename.openai.azure.com/openai/deployments/deployment/completions?api-version=2022-12-01- set the header as below
content-type:application/json
api-key: <your_api_key>- body
{
"prompt": @{concat('Summarize the content in 300 words: ' , variables('searchouttxt'))},
"max_tokens": 300,
"temperature": 0
}- now parse the json output
- schema for parsing
{
"type": "object",
"properties": {
"id": {
"type": "string"
},
"object": {
"type": "string"
},
"created": {
"type": "integer"
},
"model": {
"type": "string"
},
"choices": {
"type": "array",
"items": {
"type": "object",
"properties": {
"text": {
"type": "string"
},
"index": {
"type": "integer"
},
"logprobs": {},
"finish_reason": {
"type": "string"
}
},
"required": [
"text",
"index",
"logprobs",
"finish_reason"
]
}
}
}
}- now
- initialize a variable calle outsummary
- apply for each
- append to variable: outsummary
- now send the output back to powerapp
Power Apps
- Create a new app
- Create a input text box and a button
- Create a label to display the output
- Now create a flow and connect to the flow
- Assign the flow to button
Set(messagevar1,pineconegpt.Run(TextInput1_1.Text));- Now set the text box to display the output
- set the text property of the text box as below
Substitute(Substitute(TrimEnds(Trim(messagevar1.outputsearch)), Char(10),""), """","")- Goal is to use semantic search get the top 3 docs and then pass that with prompt engineering to chatgpt and get the response back and display it in power apps
original article — Samples2023/pinesearchgpt.md at main · balakreshnan/Samples2023 · GitHub