metadata pipeline via Step Functions

statemachines/pdf-index/definition.asl.json

@@ -0,0 +1,71 @@
+{
+  "Comment": "PDF metadata index pipeline — list keys, then per-key extract + DDB write in parallel.",
+  "StartAt": "ListPdfs",
+  "States": {
+    "ListPdfs": {
+      "Type": "Task",
+      "Resource": "arn:aws:states:::lambda:invoke",
+      "Parameters": {
+        "FunctionName": "${ListPdfsFunctionArn}",
+        "Payload": {}
+      },
+      "ResultSelector": {
+        "keys.$": "$.Payload.keys",
+        "count.$": "$.Payload.count"
+      },
+      "ResultPath": "$.list",
+      "Next": "ExtractAndIndex"
+    },
+    "ExtractAndIndex": {
+      "Type": "Map",
+      "ItemsPath": "$.list.keys",
+      "MaxConcurrency": 10,
+      "ItemProcessor": {
+        "ProcessorConfig": {
+          "Mode": "INLINE"
+        },
+        "StartAt": "ExtractMetadata",
+        "States": {
+          "ExtractMetadata": {
+            "Type": "Task",
+            "Resource": "arn:aws:states:::lambda:invoke",
+            "Parameters": {
+              "FunctionName": "${ExtractMetadataFunctionArn}",
+              "Payload": {
+                "key.$": "$"
+              }
+            },
+            "ResultSelector": {
+              "key.$": "$.Payload.key",
+              "pages.$": "$.Payload.pages",
+              "size_bytes.$": "$.Payload.size_bytes"
+            },
+            "Retry": [
+              {
+                "ErrorEquals": ["States.ALL"],
+                "IntervalSeconds": 1,
+                "MaxAttempts": 2,
+                "BackoffRate": 2.0
+              }
+            ],
+            "Next": "WriteToIndex"
+          },
+          "WriteToIndex": {
+            "Type": "Task",
+            "Resource": "arn:aws:states:::dynamodb:putItem",
+            "Parameters": {
+              "TableName": "${PdfIndexTableName}",
+              "Item": {
+                "key":        { "S.$": "$.key" },
+                "pages":      { "N.$": "States.Format('{}', $.pages)" },
+                "size_bytes": { "N.$": "States.Format('{}', $.size_bytes)" }
+              }
+            },
+            "End": true
+          }
+        }
+      },
+      "End": true
+    }
+  }
+}