From 4791cd23af08da70895204a3a7fbaf889d9af2d5 Mon Sep 17 00:00:00 2001
From: "A.J. Shulman" <Shulman.aj@gmail.com>
Date: Sat, 7 Sep 2024 12:43:05 -0400
Subject: completely restructured, added comments, and significantly reduced
 the length of the prompt (~72% shorter and cheaper)

---
 .../views/nodes/chatbot/tools/DataAnalysisTool.ts  | 59 ++++++++++++++++++++++
 1 file changed, 59 insertions(+)
 create mode 100644 src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts

(limited to 'src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts')

diff --git a/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts b/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts
new file mode 100644
index 000000000..a12ee46e5
--- /dev/null
+++ b/src/client/views/nodes/chatbot/tools/DataAnalysisTool.ts
@@ -0,0 +1,59 @@
+import { BaseTool } from './BaseTool';
+
+export class DataAnalysisTool extends BaseTool<{ csv_file_name: string | string[] }> {
+    private csv_files_function: () => { filename: string; id: string; text: string }[];
+
+    constructor(csv_files: () => { filename: string; id: string; text: string }[]) {
+        super(
+            'dataAnalysis',
+            'Analyzes, and provides insights, from one or more CSV files',
+            {
+                csv_file_name: {
+                    type: 'string',
+                    description: 'Name(s) of the CSV file(s) to analyze',
+                    required: 'true',
+                    max_inputs: '3',
+                },
+            },
+            'Provide the name(s) of up to 3 CSV files to analyze based on the user query and whichever available CSV files may be relevant.',
+            'Provides the full CSV file text for your analysis based on the user query and the available CSV file(s). '
+        );
+        this.csv_files_function = csv_files;
+    }
+
+    getFileContent(filename: string): string | undefined {
+        const files = this.csv_files_function();
+        const file = files.find(f => f.filename === filename);
+        return file?.text;
+    }
+
+    getFileID(filename: string): string | undefined {
+        const files = this.csv_files_function();
+        const file = files.find(f => f.filename === filename);
+        return file?.id;
+    }
+
+    async execute(args: { csv_file_name: string | string[] }): Promise<any> {
+        const filenames = Array.isArray(args.csv_file_name) ? args.csv_file_name : [args.csv_file_name];
+        const results = [];
+
+        for (const filename of filenames) {
+            const fileContent = this.getFileContent(filename);
+            const fileID = this.getFileID(filename);
+
+            if (fileContent && fileID) {
+                results.push({
+                    type: 'text',
+                    text: `<chunk chunk_id=${fileID} chunk_type=csv>${fileContent}</chunk>`,
+                });
+            } else {
+                results.push({
+                    type: 'text',
+                    text: `File not found: ${filename}`,
+                });
+            }
+        }
+
+        return results;
+    }
+}
-- 
cgit v1.2.3-70-g09d2