extract xlsx

FloRul · May 3, 2024 · 9beb39f · 9beb39f
1 parent e2688a4
commit 9beb39f
Show file tree

Hide file tree

Showing 3 changed files with 33 additions and 2 deletions.
diff --git a/lambdas/excel_ingestion/lambda.tf b/lambdas/excel_ingestion/lambda.tf
@@ -20,7 +20,6 @@ module "lambda_function_container_image" {
   memory_size              = 256
   role_name                = "${local.lambda_function_name}-role"
   attach_policy_statements = true
-    # see https://github.com/terraform-aws-modules/terraform-aws-lambda/issues/346
   source_path = "${path.module}/src"
 
   policy_statements = {

diff --git a/lambdas/excel_ingestion/src/index.py b/lambdas/excel_ingestion/src/index.py
@@ -13,7 +13,6 @@ def lambda_handler(event, context):
     """
     try:
         print(event)
-
         bucket, key = event['xlsx_s3_uri'].replace("s3://", "").split("/", 1)
         print(f"File located at bucket: {bucket} and key: {key}")
 

diff --git a/state_machines/rfp_email_form_fill/state_machine.json b/state_machines/rfp_email_form_fill/state_machine.json
@@ -280,6 +280,11 @@
                         "Variable": "$",
                         "StringMatches": "*.docx",
                         "Next": "Extract text from DOCX"
+                      },
+                      {
+                        "Variable": "$",
+                        "StringMatches": "*.xlsx",
+                        "Next": "Extract CSV from XLSX"
                       }
                     ],
                     "Default": "Filter out unsupported attachment",
@@ -355,6 +360,34 @@
                     "ResultSelector": {
                       "extracted_text.$": "States.Array($.Payload.attachment_uri)"
                     }
+                  },
+                  "Extract CSV from XLSX": {
+                    "Type": "Task",
+                    "Resource": "arn:aws:states:::lambda:invoke",
+                    "OutputPath": "$.extracted_csv_uri",
+                    "Parameters": {
+                      "Payload": {
+                        "xlsx_s3_uri.$": "$"
+                      },
+                      "FunctionName": "arn:aws:lambda:us-east-1:446872271111:function:levio-esta-excel-ingestion-dev:$LATEST"
+                    },
+                    "Retry": [
+                      {
+                        "ErrorEquals": [
+                          "Lambda.ServiceException",
+                          "Lambda.AWSLambdaException",
+                          "Lambda.SdkClientException",
+                          "Lambda.TooManyRequestsException"
+                        ],
+                        "IntervalSeconds": 1,
+                        "MaxAttempts": 3,
+                        "BackoffRate": 2
+                      }
+                    ],
+                    "End": true,
+                    "ResultSelector": {
+                      "extracted_csv_uri.$": "States.Array($.Payload.extracted_csv_uri)"
+                    }
                   }
                 }
               },
-Original file line number
+Diff line change
@@ Expand Up / @@ -13,7 +13,6 @@ def lambda_handler(event, context): @@
         """
         try:
             print(event)
             bucket, key = event['xlsx_s3_uri'].replace("s3://", "").split("/", 1)
             print(f"File located at bucket: {bucket} and key: {key}")
@@ Expand Down @@