feat: markitdown implementation

Dockerfile

@@ -74,9 +74,16 @@ RUN apt-get update && apt-get install -y \
   texlive-latex-extra \
   texlive-latex-recommended \
   texlive-xetex \
+  python3 \
+  python3-pip \
+  pipx \
   --no-install-recommends \
+  && pipx install "markitdown[all]" \
   && rm -rf /var/lib/apt/lists/*
 
+# Add pipx bin directory to PATH
+ENV PATH="/root/.local/bin:${PATH}"
+
 # Install VTracer binary
 RUN ARCH=$(uname -m) && \
   if [ "$ARCH" = "aarch64" ]; then \

README.md

@@ -45,6 +45,7 @@ A self-hosted online file converter. Supports over a thousand different formats.
 | [FFmpeg](https://ffmpeg.org/)                                   | Video            | ~472          | ~199        |
 | [Potrace](https://potrace.sourceforge.net/)                     | Raster to vector | 4             | 11          |
 | [VTracer](https://github.com/visioncortex/vtracer)              | Raster to vector | 8             | 1           |
+| [Markitdown](https://github.com/microsoft/markitdown)           | Documents        | 6             | 1           |
 
 <!-- many ffmpeg fileformats are duplicates -->
 

src/converters/main.ts

@@ -23,6 +23,7 @@ import { convert as convertresvg, properties as propertiesresvg } from "./resvg"
 import { convert as convertImage, properties as propertiesImage } from "./vips";
 import { convert as convertVtracer, properties as propertiesVtracer } from "./vtracer";
 import { convert as convertxelatex, properties as propertiesxelatex } from "./xelatex";
+import { convert as convertMarkitdown, properties as propertiesMarkitdown } from "./markitdown";
 
 // This should probably be reconstructed so that the functions are not imported instead the functions hook into this to make the converters more modular
 
@@ -127,6 +128,10 @@ const properties: Record<
     properties: propertiesVtracer,
     converter: convertVtracer,
   },
+  markitDown: {
+    properties: propertiesMarkitdown,
+    converter: convertMarkitdown,
+  },
 };
 
 function chunks<T>(arr: T[], size: number): T[][] {

src/converters/markitdown.ts

@@ -0,0 +1,39 @@
+import { execFile as execFileOriginal } from "node:child_process";
+import { ExecFileFn } from "./types";
+
+export const properties = {
+  from: {
+    document: ["pdf", "powerpoint", "excel", "docx", "pptx", "html"],
+  },
+  to: {
+    document: ["md"],
+  },
+};
+
+export async function convert(
+  filePath: string,
+  fileType: string,
+  convertTo: string,
+  targetPath: string,
+  options?: unknown,
+  execFile: ExecFileFn = execFileOriginal,
+): Promise<string> {
+  return new Promise((resolve, reject) => {
+    execFile("markitdown", [filePath, "-o", targetPath], (err, stdout, stderr) => {
+      if (err) {
+        reject(`markitdown error: ${err}`);
+        return;
+      }
+
+      if (stdout) {
+        console.log(`stdout: ${stdout}`);
+      }
+
+      if (stderr) {
+        console.error(`stderr: ${stderr}`);
+      }
+
+      resolve("Done");
+    });
+  });
+}