feat(routes): add doc-to-txt route

README.md

@@ -13,6 +13,7 @@
 
 Docsmith is a RESTful API, built using Node.js and the [Fastify](https://fastify.io/) web framework, that can convert files from:
 
+-   DOC to TXT
 -   DOCX to HTML
 -   DOCX to TXT
 -   PDF to HTML

package.json

@@ -101,6 +101,7 @@
 		"@fastify/static": "^6.10.1",
 		"@fastify/swagger": "^8.3.1",
 		"@fastify/under-pressure": "^8.2.0",
+		"cfb": "^1.2.2",
 		"clean-css": "^5.3.2",
 		"cssesc": "^3.0.0",
 		"cssom": "^0.5.0",
@@ -128,6 +129,7 @@
 		"redoc": "^2.0.0",
 		"secure-json-parse": "^2.7.0",
 		"tesseract.js": "^4.0.5",
-		"upath": "^2.0.1"
+		"upath": "^2.0.1",
+		"word-extractor": "^1.0.4"
 	}
 }

src/config/index.js

@@ -263,6 +263,11 @@ async function getConfig() {
 						: undefined,
 				},
 				tags: [
+					{
+						name: "DOC",
+						description:
+							"Endpoints used for the conversion of DOC documents",
+					},
 					{
 						name: "DOCX",
 						description:

src/plugins/doc-to-txt/index.js

@@ -0,0 +1,44 @@
+const fp = require("fastify-plugin");
+const WordExtractor = require("word-extractor");
+
+/**
+ * @author Frazer Smith
+ * @description Pre-handler plugin that uses Word-Extractor to convert Buffer containing
+ * DOC file in `req.body` to TXT.
+ * `req` object is decorated with `conversionResults.body` holding the converted document.
+ * @param {object} server - Fastify instance.
+ */
+async function plugin(server) {
+	const wordExtractor = new WordExtractor();
+
+	server.addHook("onRequest", async (req) => {
+		req.conversionResults = { body: undefined };
+	});
+
+	server.addHook("preHandler", async (req, res) => {
+		try {
+			const results = await wordExtractor.extract(req.body);
+
+			const value = `${results.getHeaders({
+				includeFooters: false,
+			})}\n${results.getTextboxes({
+				includeHeadersAndFooters: false,
+			})}\n${results.getBody()}\n${results.getEndnotes()}\n${results.getFootnotes()}\n${results.getFooters()}`.trim();
+
+			req.conversionResults.body = value;
+			res.type("text/plain; charset=utf-8");
+		} catch {
+			/**
+			 * Word-Extractor will throw if the .doc file provided
+			 * by client is malformed, thus client error code
+			 */
+			throw server.httpErrors.badRequest();
+		}
+	});
+}
+
+module.exports = fp(plugin, {
+	fastify: "4.x",
+	name: "doc-to-txt",
+	dependencies: ["@fastify/sensible"],
+});

src/plugins/doc-to-txt/plugin.test.js

@@ -0,0 +1,91 @@
+const fs = require("fs/promises");
+const Fastify = require("fastify");
+const isHtml = require("is-html");
+const sensible = require("@fastify/sensible");
+const plugin = require(".");
+
+describe("DOC-to-TXT conversion plugin", () => {
+	let server;
+
+	beforeAll(async () => {
+		server = Fastify();
+
+		server.addContentTypeParser(
+			"application/msword",
+			{ parseAs: "buffer" },
+			async (_req, payload) => payload
+		);
+
+		await server.register(sensible).register(plugin);
+
+		server.post("/", (req, res) => {
+			res.header("content-type", "application/json").send(
+				req.conversionResults
+			);
+		});
+
+		await server.ready();
+	});
+
+	afterAll(async () => {
+		await server.close();
+	});
+
+	it("Converts DOC file to TXT", async () => {
+		const response = await server.inject({
+			method: "POST",
+			url: "/",
+			body: await fs.readFile(
+				"./test_resources/test_files/valid_doc.doc"
+			),
+			headers: {
+				"content-type": "application/msword",
+			},
+		});
+
+		const { body } = JSON.parse(response.payload);
+
+		// String found in first heading of the test document
+		expect(body).toMatch(
+			"Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nunc ac faucibus odio."
+		);
+		// String found at end of the test document
+		expect(body).toMatch(
+			/Nullam venenatis commodo imperdiet. Morbi velit neque, semper quis lorem quis, efficitur dignissim ipsum. Ut ac lorem sed turpis imperdiet eleifend sit amet id sapien$/m
+		);
+		expect(isHtml(body)).toBe(false);
+		expect(response.statusCode).toBe(200);
+	});
+
+	// TODO: use `it.concurrent.each()` once it is no longer experimental
+	it.each([
+		{ testName: "is missing" },
+		{
+			testName: "is not a valid DOC file",
+			readFile: true,
+		},
+	])(
+		"Returns HTTP status code 400 if DOC file $testName",
+		async ({ readFile }) => {
+			const response = await server.inject({
+				method: "POST",
+				url: "/",
+				headers: {
+					"content-type": "application/msword",
+				},
+				body: readFile
+					? await fs.readFile(
+							"./test_resources/test_files/invalid_doc.doc"
+					  )
+					: undefined,
+			});
+
+			expect(JSON.parse(response.payload)).toEqual({
+				error: "Bad Request",
+				message: "Bad Request",
+				statusCode: 400,
+			});
+			expect(response.statusCode).toBe(400);
+		}
+	);
+});