@cumulus/es-client 18.3.4 → 19.0.0

package/executions.js

@@ -0,0 +1,99 @@
+//@ts-check
+
+const Logger = require('@cumulus/logger');
+
+const { getEsClient, defaultIndexAlias } = require('./search');
+
+const log = new Logger({ sender: '@cumulus/es-client/executions' });
+
+/**
+ * Generates a list of execution record IDs from an Elasticsearch Index based on a collection ID
+ *
+ * @param {Object} esClient - The Elasticsearch client object.
+ * @param {string} index - The name of the Elasticsearch index.
+ * @param {string} collectionId - The ID of the collection to match.
+ * @param {number} batchSize - The number of records to fetch in each batch.
+ * @returns {Promise<Array<string>>} A promise that resolves to an array of record IDs.
+ * @throws {Error} Throws an error if fetching the record IDs fails.
+ */
+const _fetchEsRecordIds = async (esClient, index, collectionId, batchSize) => {
+  try {
+    const response = await esClient.client.search({
+      index,
+      type: 'execution',
+      scroll: process.env.ES_SCROLL_TIME || '5m',
+      body: {
+        query: {
+          match: {
+            collectionId,
+          },
+        },
+        _source: false,
+      },
+      size: batchSize,
+    });
+    return response.body.hits.hits.map((hit) => hit._id);
+  } catch (error) {
+    log.error(`Failed to get recordIds ${JSON.stringify(error)}`);
+    throw new Error(`Failed to fetch record IDs from Elasticsearch index ${index}: ${error.message}`);
+  }
+};
+
+/**
+ * Deletes executions in batches by collection ID.
+ *
+ * @param {Object} params - The parameters for the function.
+ * @param {string} [params.index] - The index to delete from.
+ * @param {string} params.collectionId - The ID of the collection.
+ * @param {number} params.batchSize - The size of the batches to delete.
+ * @returns {Promise<void>} A promise that resolves when the deletions are complete.
+ */
+const batchDeleteExecutionsByCollection = async ({
+  index = defaultIndexAlias,
+  collectionId,
+  batchSize,
+}) => {
+  try {
+    const esClient = await getEsClient();
+    if (!esClient.client) {
+      throw new Error('ES client not initialized!');
+    }
+    let recordIds;
+    let failures = 0;
+    while (recordIds === undefined || recordIds.length > 0) {
+      // eslint-disable-next-line no-await-in-loop
+      recordIds = await _fetchEsRecordIds(
+        esClient,
+        index,
+        collectionId,
+        batchSize
+      );
+      if (recordIds.length > 0) {
+        const body = recordIds.map((id) => ({
+          delete: { _index: index, _type: 'execution', _id: id },
+        }));
+        // eslint-disable-next-line no-await-in-loop
+        const response = await esClient.client.bulk({ body, refresh: 'true' });
+        if (response.body.errors) {
+          failures += 1;
+          log.error(`Bulk deletion encountered errors: ${JSON.stringify(
+            response.body.errors
+          )}`);
+        }
+      }
+      log.info(
+        `Successfully deleted ${recordIds.length} execution records from ${index} for collection ${collectionId}`
+      );
+      if (failures > 0) {
+        log.info(`${failures} errors encountered during deletion - please check logs for details`);
+      }
+    }
+  } catch (error) {
+    log.error(`Failed to delete execution records from ${index}: ${error.message}`);
+    throw error;
+  }
+};
+
+module.exports = {
+  batchDeleteExecutionsByCollection,
+};

package/indexer.js

@@ -90,7 +90,7 @@ async function genericRecordUpdate(esClient, id, doc, index, type, parent) {
     try {
       indexResponse = await actualEsClient.client.index(params);
     } catch (error) {
-      if (error.name === 'ResponseError' && error.meta.body.message.includes('The security token included in the request is expired')) {
+      if (error.name === 'ResponseError' && error?.meta?.body?.message.includes('The security token included in the request is expired')) {
         logger.warn(`genericRecordUpdate encountered a ResponseError ${JSON.stringify(error)}, updating credentials and retrying`);
         await actualEsClient.refreshClient();
         indexResponse = await actualEsClient.client.index(params);
@@ -99,7 +99,7 @@ async function genericRecordUpdate(esClient, id, doc, index, type, parent) {
       }
     }
   } catch (error) {
-    logger.error(`Error thrown on index ${JSON.stringify(error)}`);
+    logger.error(`Error thrown on index ${JSON.stringify(error.message)}`);
     throw error;
   }
   return indexResponse.body;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cumulus/es-client",
-  "version": "18.3.4",
+  "version": "19.0.0",
   "description": "Utilities for working with Elasticsearch",
   "keywords": [
     "CUMULUS",
@@ -33,11 +33,11 @@
   "author": "Cumulus Authors",
   "license": "Apache-2.0",
   "dependencies": {
-    "@aws-sdk/credential-providers": "^3.535.0",
-    "@cumulus/common": "18.3.4",
-    "@cumulus/errors": "18.3.4",
-    "@cumulus/logger": "18.3.4",
-    "@cumulus/message": "18.3.4",
+    "@aws-sdk/credential-providers": "^3.621.0",
+    "@cumulus/common": "19.0.0",
+    "@cumulus/errors": "19.0.0",
+    "@cumulus/logger": "19.0.0",
+    "@cumulus/message": "19.0.0",
     "@elastic/elasticsearch": "^5.6.20",
     "aws4": "^1.12.0",
     "lodash": "~4.17.21",
@@ -45,9 +45,9 @@
     "p-limit": "^1.2.0"
   },
   "devDependencies": {
-    "@cumulus/aws-client": "18.3.4",
-    "@cumulus/test-data": "18.3.4",
+    "@cumulus/aws-client": "19.0.0",
+    "@cumulus/test-data": "19.0.0",
     "p-each-series": "^2.1.0"
   },
-  "gitHead": "d2ddab76522b140fd7b92d865a8e0ec5d80485c7"
+  "gitHead": "3a619db6e6fa20846eb75802e89d5a065fd62fcc"
 }

package/search.js

@@ -204,7 +204,7 @@ class EsClient {
   /**
    * Initializes a new instance of the `EsClient` class.
    *
-   * @param {string} host - The host URL for the Elasticsearch instance.
+   * @param {string} [host] - The host URL for the Elasticsearch instance.
    * @param {boolean} [metrics=false] - A flag indicating whether metrics are enabled.
    */
   constructor(host, metrics = false) {
@@ -441,8 +441,8 @@ class Search extends BaseSearch {}
 /**
  * Initializes and returns an instance of an `EsClient` Class
  *
- * @param {string} host - The host URL for the Elasticsearch instance.
- * @param {boolean} metrics - A flag indicating whether metrics are enabled.
+ * @param {string} [host] - The host URL for the Elasticsearch instance.
+ * @param {boolean} [metrics] - A flag indicating whether metrics are enabled.
  * @returns {Promise<EsClient>} A promise that resolves to an instance of `EsClient`.
  */
 const getEsClient = async (host, metrics) => {