feat(mistral_ocr): Add outfile option and handle empty input

Introduces the ability to save OCR output to a file and adds robustness for empty file inputs. Increments version to 0.2.1.

Author: iovdin

README.md

@@ -295,7 +295,7 @@ Tune is a versatile toolkit designed for developers and users to effectively int
 ### `mistral_ocr`
 Extract text from documents and images using the [Mistral OCR API](https://mistral.ai/). Requires a `MISTRAL_KEY` set in `.env`.
 
-Supports documents: `.pdf`, `.docx`, `.pptx`, `.txt`, `.epub`, `.xml`, `.rtf`, `.odt`, `.bib`, `.fb2`, `.ipynb`, `.tex`, `.opml`, `.1`, `.man`
+Supports documents: `.pdf`, `.docx`, `.xlsx`, `.pptx`, `.txt`, `.epub`, `.xml`, `.rtf`, `.odt`, `.bib`, `.fb2`, `.ipynb`, `.tex`, `.opml`, `.1`, `.man`
 
 Supports images: `.jpg`, `.jpeg`, `.png`, `.avif`, `.tiff`, `.gif`, `.heic`, `.heif`, `.bmp`, `.webp`
 
@@ -326,6 +326,17 @@ tool_result:
 Your disk usage has exceeded 90%. Please free up space to avoid performance issues.
 ```
 
+You can also save the extracted text to a file using the `outfile` parameter:
+
+```chat
+user: @mistral_ocr
+extract text from report.pdf and save it to report.md
+
+tool_call: mistral_ocr {"filename":"report.pdf","outfile":"report.md"}
+tool_result:
+saved to report.md
+```
+
 ### `websearch`
 Search the web with web enabled llms
 Supports search with  `perplexity/sonar`, `perplexity/sonar-pro`, `gpt-4o-search-preview`, `gpt-4o-mini-search-preview` models via the `model` parameter (defaults to `perplexity/sonar`).

package.json

@@ -1,6 +1,6 @@
 {
   "name": "tune-basic-toolset",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "description": "Basic toolset for tune",
   "main": "src/index.js",
   "files": [

src/linenum.proc.js

@@ -10,7 +10,7 @@ module.exports = async function(node, args) {
   const res = Object.assign({}, node);
   res.read = async function(args) {
     let contents = await node.read(args);
-    contents = contents.split("\n");
+    contents = (contents || "").split("\n");
     const pad = `${contents.length}`.length;
     return contents.map((item, index) => {
       const lineNum = `${index + 1}`.padEnd(pad + 1, " ") + "|";

src/mistral_ocr.schema.json

@@ -6,6 +6,10 @@
       "filename": {
         "type": "string",
         "description": "Path to the file to perform OCR on"
+      },
+      "outfile": {
+        "type": "string",
+        "description": "save ocr result to this filename"
       }
     },
     "required": ["filename"]

src/mistral_ocr.tool.js

@@ -31,7 +31,7 @@ const MIME_TYPES = {
   '.webp':  { kind: 'image', mime: 'image/webp' },
 };
 
-module.exports = async function mistralOcr({ filename }, ctx) {
+module.exports = async function mistralOcr({ filename, outfile }, ctx) {
   const apiKey = await ctx.read("MISTRAL_KEY");
   if (!apiKey) {
     throw new Error("MISTRAL_KEY is not set");
@@ -73,7 +73,12 @@ module.exports = async function mistralOcr({ filename }, ctx) {
     throw new Error(`Mistral OCR error: ${response.status} ${response.statusText} - ${errorText}`);
   }
 
-  const result = await response.json();
+  let result = await response.json();
+  result = result.pages.map(page => page.markdown).join("\n\n")
+  if (outfile) {
+    await ctx.write(outfile, result)
+    return `saved to ${outfile}`
+  }
 
-  return result.pages.map(page => page.markdown).join("\n\n");
+  return result;
 };