From 2e353cdba464b2427c948348e760dc95f92513e1 Mon Sep 17 00:00:00 2001
From: SheetJS <dev@sheetjs.com>
Date: Sun, 9 Aug 2020 01:04:31 -0400
Subject: [PATCH] process uploads

---
 api/data/index.js   | 37 ++++++++++++-------------------------
 api/upload/index.js | 28 ++++++++++++++++++++++++++++
 package.json        |  2 +-
 public/index.html   | 34 ++++++++++++++++++++++++++++------
 src/util.js         | 33 +++++++++++++++++++++++++++++++++
 5 files changed, 102 insertions(+), 32 deletions(-)
 create mode 100644 api/upload/index.js
 create mode 100644 src/util.js

diff --git a/api/data/index.js b/api/data/index.js
index fc60e1d..40c9d47 100644
--- a/api/data/index.js
+++ b/api/data/index.js
@@ -1,36 +1,23 @@
-const XLSX = require('xlsx');
 const URL = require('url');
+const do_wb = require('../../src/util');
 const request = require('request');
 
-const do_url = (req, url, res) => {
+module.exports = function(req, res) {
+	res.setHeader('Access-Control-Allow-Origin', '*');
+	const url = URL.parse(req.url, true);
+
+	/* request url and process */
+	if(!url.query.url) return res.status(400).send("Must specify url");
 	request(url.query.url, {encoding:null}, function(err, response, body) {
-		if(err) return res.status(500).send(err.toString());
+		if(err) return res.status(502).send(err.toString());
+
+		/* response.statusCode is expected to be 200 */
 		switch(response.statusCode) {
 			case 200: break;
 			case 404: return res.status(404).send(`Cannot find ${url.query.url}`);
 			default:  return res.status(500).send(`Unrecognized status code ${response.statusCode}`);
 		}
-		const wb = XLSX.read(body, {type:'buffer'});
-		const N = url.query.N ? parseInt(url.query.N,10) : 0;
-		if(N < 0) {
-			switch(url.query.t || "csv") {
-				case "json": return res.status(200).send(JSON.stringify(wb.SheetNames.join("\n")));
-				default: return res.status(200).send(wb.SheetNames.join("\n"));
-			}
-		}
-		if(N >= wb.SheetNames.length) return res.status(500).send(`Cannot find sheet ${N}`);
-		var ws = wb.Sheets[wb.SheetNames[N]];
-		switch(url.query.t) {
-			case "json": return res.status(200).json(XLSX.utils.sheet_to_json(ws, {header:1, raw:true}));
-			case "html": return XLSX.stream.to_html(ws).pipe(res);
-			default: XLSX.stream.to_csv(ws).pipe(res);
-		}
+
+		do_wb(req, body, url, res);
 	});
 };
-
-module.exports = function(req, res) {
-	var url = URL.parse(req.url, true);
-	res.setHeader('Access-Control-Allow-Origin', '*');
-	if(!url.query.url) return res.status(500).send("Must issue command");
-	do_url(req, url, res);
-};
diff --git a/api/upload/index.js b/api/upload/index.js
new file mode 100644
index 0000000..724781b
--- /dev/null
+++ b/api/upload/index.js
@@ -0,0 +1,28 @@
+const URL = require('url'), fs = require('fs');
+const do_wb = require('../../src/util');
+const formidable = require('formidable-serverless');
+
+module.exports = function(req, res) {
+	res.setHeader('Access-Control-Allow-Origin', '*');
+	const url = URL.parse(req.url, true);
+
+	/* parse form data */
+	const form = formidable({multiples: true, maxFileSize: 2*1024*1024});
+	form.parse(req, (err, fields, files) => {
+		if(err) return res.status(400).send(err.message || err);
+		if(!url.query) url.query = fields;
+		["N", "t"].forEach(k => { if(!url.query[k] && fields[k] != null) url.query[k] = fields[k]; });
+
+		/* look for first uploaded file entry */
+		if(!files) return res.status(400).send("Missing file");
+		const fentries = Object.entries(files);
+		if(fentries.length == 0) return res.status(400).send("Missing file");
+
+		/* read file */
+		const file = fentries[0][1];
+		fs.readFile(file.path, (err, body) => {
+			if(err) return res.status(500).send(err.message || err);
+			do_wb(req, body, url, res);
+		});
+	});
+};
diff --git a/package.json b/package.json
index e9a5d92..949e802 100644
--- a/package.json
+++ b/package.json
@@ -5,7 +5,7 @@
 	"description": "Spreadsheet CSV conversion microservice",
 	"keywords": ["csv", "spreadsheet", "data", "conversion", "excel"],
 	"dependencies": {
-		"micro":"",
+		"formidable-serverless":"",
 		"request":"",
 		"xlsx":""
 	},
diff --git a/public/index.html b/public/index.html
index cccd588..13ee3af 100644
--- a/public/index.html
+++ b/public/index.html
@@ -8,7 +8,7 @@
 <pre>
 <b><a href="http://sheetjs.com">SheetJS</a> Spreadsheet Conversion Service</b>
 
-send a request to /api/data?url=&lt;url&gt;&amp;N=&lt;idx&gt; to convert the spreadsheet at `url` to a simpler format
+<b>Proxy:</b> send a request to /api/data to convert the spreadsheet at `url` to a simpler format
 
 <b>parameters:</b>
 - url=&lt;url&gt;      the url to request
@@ -28,12 +28,12 @@ var URLS = [
 	[],
 	["[LIST]", {url:URL2,N:-1}],
 	["[N=1] ", {url:URL2,N:1}],
-	["[N=2] ", {url:URL2,N:2}],
-	["[N=3] ", {url:URL2,N:3}],
+//	["[N=2] ", {url:URL2,N:2}],
+//	["[N=3] ", {url:URL2,N:3}],
 	["[N=4] ", {url:URL2,N:4}],
-	["[N=5] ", {url:URL2,N:5}],
-	["[N=6] ", {url:URL2,N:6}],
-	["[N=7] ", {url:URL2,N:7}],
+//	["[N=5] ", {url:URL2,N:5}],
+//	["[N=6] ", {url:URL2,N:6}],
+//	["[N=7] ", {url:URL2,N:7}],
 	["[N=8] ", {url:URL2,N:8}]
 ];
 var parms = ["N", "t"];
@@ -45,6 +45,28 @@ for(var i = 0; i < URLS.length; ++i) {
 }
 </script>
 <pre>
+
+<b>Upload:</b> upload a file to /api/upload to convert the spreadsheet at `url` to a simpler format
+
+<b>parameters:</b>
+- N=&lt;idx&gt;        the sheet index to use (-1 returns a list of sheet names in the workbook)
+- t=&lt;type&gt;       export type: "json" for json, "html" for html, default is CSV
+
+<b>example:</b>
+<form method="POST" action="api/upload" enctype="multipart/form-data"><input type="hidden" name="N" value="0">
+<label for="file">Upload a file</label>
+<input type="file" name="file" id="file">
+
+<label for="t">Export type</label>
+<select name="t" id="t">
+	<option value="csv" selected>CSV</option>
+	<option value="json">Array of arrays</option>
+	<option value="html">HTML TABLE</option>
+</select>
+
+<input type="submit" value="Click here to upload!">
+</form>
+
 <b>This service is powered by <a href="http://sheetjs.com">SheetJS</a></b>
 
 <b>Source Code:</b>
diff --git a/src/util.js b/src/util.js
new file mode 100644
index 0000000..37d0f11
--- /dev/null
+++ b/src/util.js
@@ -0,0 +1,33 @@
+const XLSX = require('xlsx');
+
+module.exports = function(req, body, url, res) {
+	/* try to read the buffer */
+	let wb;
+	try {
+		wb = XLSX.read(body, {type:'buffer'});
+	} catch(e) { return res.status(500).send(e.message || e); }
+
+	/* N parameter specifies worksheet index */
+	const N = url.query.N ? parseInt(url.query.N,10) : 0;
+
+	/* -1 -> return sheet names */
+	if(N < 0) switch(url.query.t || "csv") {
+		case "json": return res.status(200).send(JSON.stringify(wb.SheetNames.join("\n")));
+		default: return res.status(200).send(wb.SheetNames.join("\n"));
+	}
+
+	/* find worksheet */
+	if(N >= wb.SheetNames.length) return res.status(500).send(`Cannot find sheet ${N}`);
+	const ws = wb.Sheets[wb.SheetNames[N]];
+
+	/* result depends on "t" parameter:
+	 * - json: AOA (sheet_to_json with header:1, raw: true)
+	 * - html: HTML (using stream.to_html)
+	 * default: CSV (using stream.to_csv)
+	 */
+	switch(url.query.t) {
+		case "json": return res.status(200).json(XLSX.utils.sheet_to_json(ws, {header:1, raw:true}));
+		case "html": return XLSX.stream.to_html(ws).pipe(res);
+		default: return XLSX.stream.to_csv(ws).pipe(res);
+	}
+};