TableExtractionCognitiveSearchAsync

This repository is integrating Azure Cognigive Form Recognizer with Azure Cognitive Search in asyncronous process. For it to work you need:

• Azure search and its credentials
• Azure search index/skillset/indexer in the format that is compatible with the output of the pipeline (below)
• Azure storage with data to be processed
• Azure Service Bus - Two queues named "polling_queue" and "results_queue" and their credentials
• Azure Form Recognizer, and credentials
• Create Azure Function app (python)
• VSC and Azure addon with Azure Functions and Azure Core tools
• Its useful to have Postman or other tool for testing and debugging purposes, but it's not necessary

You will need to update Setting -> Configuration of your function app adding env variables:

Variable name	value
"AZURE_WEB_JOB_STORAGE"	"",
"AZURE_WEB_JOB_STORAGE"	"",
"AZURE_SERVICE_BUS_CONNECTION_STRING"	"",
"AZURE_FORM_RECOGNIZER_ENDPOINT"	"https:...",
"AZURE_FORM_RECOGNIZER_ENDPOINT_KEY"	: "",
"AZURE_SEARCH_ENDPOINT"	"https:...",
"AZURE_SEARCH_ENDPOINT_KEY"	"",
"AZURE_SEARCH_INDEX_NAME"	"",
"MODEL"	"prebuilt-layout" OR "invoice"

git clone this repository go to left hand side panel in VSC and find Azure. Log in to Azure. Find "workspaces" menu and deploy to Azure button:

Follow the instructions to deploy your functions to the app. It should create 3 functions in your Azure Function app * Start_processing * Fetch_results * Push_results

To include the function in your azure search index you need to add the skill to skillset: Here is an example of how it can look like:

      { 
        "@odata.type":"#Microsoft.Skills.Custom.WebApiSkill",
        "name":"formrecognizer-tables",
        "description":"Analyze documents and extracts tables.",
        "uri":"{{Async_table_extractor_url}}?code={{Async_table_extractor_key}}",
        "httpMethod":"POST",
        "timeout": "PT1M",
        "context":"/document",
        "batchSize":1,
        "inputs":[ 
            { 
                "name":"metadata_storage_sas_token",
                "source":"/document/metadata_storage_sas_token"
            },
            { 
                "name":"metadata_storage_path",
                "source":"/document/metadata_storage_path"
            },
            { 
                "name":"metadata_storage_path_decoded",
                "source":"/document/metadata_storage_path_decoded"
            }
        ],
        "outputs":[ 
            { 
                "name":"tables",
                "targetName":"tables"
            }
        ]
    }