From a5188737d3f3d0b0ed9ed45e24caa87f4df85563 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Anders=20J=C3=BCrisoo?= <jurisoo@hotmail.com>
Date: Fri, 24 Jan 2025 15:55:14 +0100
Subject: [PATCH] Add CSV reader

---
 packages/nodejs/package.json                  |  1 +
 .../src/computers/CsvFileRead/CsvFileRead.ts  | 89 +++++++++++++++++++
 .../nodejs/src/computers/CsvFileRead/index.ts |  1 +
 packages/nodejs/src/computers/index.ts        |  1 +
 yarn.lock                                     |  8 ++
 5 files changed, 100 insertions(+)
 create mode 100644 packages/nodejs/src/computers/CsvFileRead/CsvFileRead.ts
 create mode 100644 packages/nodejs/src/computers/CsvFileRead/index.ts

diff --git a/packages/nodejs/package.json b/packages/nodejs/package.json
index 4bb833d63..86a6798fc 100644
--- a/packages/nodejs/package.json
+++ b/packages/nodejs/package.json
@@ -31,6 +31,7 @@
   "dependencies": {
     "@data-story/core": "workspace:*",
     "axios": "^1.3.4",
+    "csv-parse": "^5.6.0",
     "dotenv": "^16.0.3",
     "glob": "^11.0.0",
     "openai": "^3.2.1",
diff --git a/packages/nodejs/src/computers/CsvFileRead/CsvFileRead.ts b/packages/nodejs/src/computers/CsvFileRead/CsvFileRead.ts
new file mode 100644
index 000000000..2eb9ff8fc
--- /dev/null
+++ b/packages/nodejs/src/computers/CsvFileRead/CsvFileRead.ts
@@ -0,0 +1,89 @@
+import * as glob from 'glob';
+import fs from 'fs';
+import path from 'path';
+import { parse } from 'csv-parse/sync';
+import { Computer, serializeError, str } from '@data-story/core';
+
+export const CsvFileRead: Computer = {
+  name: 'CsvFile.read',
+  label: 'CsvFile.read',
+  category: 'NodeJs',
+  inputs: [],
+  outputs: [
+    {
+      name: 'output',
+      schema: {},
+    },
+    {
+      name: 'errors',
+      schema: {},
+    },
+  ],
+  params: [
+    str({
+      name: 'file_path',
+      label: 'File path (supports glob patterns)',
+      help: 'File path, e.g., **/*.csv',
+    }),
+    str({
+      name: 'delimiter',
+      label: 'Delimiter',
+      help: 'CSV delimiter character (default: ,)',
+      value: ',',
+    }),
+  ],
+
+  async *run({ output, params }) {
+    const pathPattern = params.file_path as string;
+    const delimiter = params.delimiter as string;
+
+    // Check if the provided path is absolute
+    const isAbsolutePath = path.isAbsolute(pathPattern);
+
+    let files: string[] = [];
+    try {
+      if (isAbsolutePath) {
+        // If it's an absolute path, use it directly with glob
+        files = glob.sync(pathPattern, {
+          ignore: ['**/node_modules/**'],
+          absolute: true,
+        });
+      } else {
+        // If it's a relative path, resolve using the workspace folder
+        const cwd = process.env.WORKSPACE_FOLDER_PATH as string;
+        files = glob.sync(pathPattern, {
+          cwd, // Resolve relative paths from the workspace folder
+          ignore: ['**/node_modules/**'],
+          absolute: true,
+        });
+      }
+
+      // Process each file found by glob
+      for (const file of files) {
+        try {
+          const content = fs.readFileSync(file, 'utf-8');
+          const records = parse(content, {
+            columns: true, // Use first row as column names
+            delimiter,
+            skip_empty_lines: true,
+            trim: true,
+          });
+
+          // Add file path to each record
+          const items = records.map((record: any) => ({
+            ...record,
+            _filePath: file,
+          }));
+
+          output.push(items);
+          yield;
+        } catch (fileError: any) {
+          output.pushTo('errors', [serializeError(fileError)]);
+        }
+      }
+    } catch (error: any) {
+      output.pushTo('errors', [serializeError(error)]);
+      yield;
+    }
+  },
+};
diff --git a/packages/nodejs/src/computers/CsvFileRead/index.ts b/packages/nodejs/src/computers/CsvFileRead/index.ts
new file mode 100644
index 000000000..f4cc3faea
--- /dev/null
+++ b/packages/nodejs/src/computers/CsvFileRead/index.ts
@@ -0,0 +1 @@
+export { CsvFileRead } from './CsvFileRead'
\ No newline at end of file
diff --git a/packages/nodejs/src/computers/index.ts b/packages/nodejs/src/computers/index.ts
index 12dfbd1b8..0c78dc093 100644
--- a/packages/nodejs/src/computers/index.ts
+++ b/packages/nodejs/src/computers/index.ts
@@ -1,3 +1,4 @@
+export { CsvFileRead } from './CsvFileRead'
 export { JsonFileRead } from './JsonFileRead'
 export { JsonFileWrite } from './JsonFileWrite'
 export { ListFiles } from './ListFiles'
diff --git a/yarn.lock b/yarn.lock
index bb79beadb..ee24a7a0b 100644
--- a/yarn.lock
+++ b/yarn.lock
@@ -550,6 +550,7 @@ __metadata:
     "@types/node": "npm:18.14.2"
     "@types/ws": "npm:^8.5.4"
     axios: "npm:^1.3.4"
+    csv-parse: "npm:^5.6.0"
     dotenv: "npm:^16.0.3"
     glob: "npm:^11.0.0"
     openai: "npm:^3.2.1"
@@ -5032,6 +5033,13 @@ __metadata:
   languageName: node
   linkType: hard
 
+"csv-parse@npm:^5.6.0":
+  version: 5.6.0
+  resolution: "csv-parse@npm:5.6.0"
+  checksum: 10/4c82e11f50ae0ccbac2aed716ef2502d0468bf96552083561db789fc0258ee4bb0a30106fcfb2684f153cb4042f0413e0eac3645d5466874803b7ccdeba67ac8
+  languageName: node
+  linkType: hard
+
 "culvert@npm:^0.1.2":
   version: 0.1.2
   resolution: "culvert@npm:0.1.2"