From 6aa9cfb20b2c8773fb51c6ffb9dcb517be01c4d0 Mon Sep 17 00:00:00 2001
From: Starbeamrainbowlabs <sbrl@starbeamrainbowlabs.com>
Date: Mon, 16 Aug 2021 23:17:46 +0100
Subject: [PATCH] Add optional FILTER_FEED environment variable

---
 README.md             | 13 +++++++++++++
 package-lock.json     | 11 +++++++++++
 package.json          |  1 +
 src/index.11tydata.js | 10 ++++++++++
 src/lib/check_text.js | 24 ++++++++++++++++++++++++
 5 files changed, 59 insertions(+)
 create mode 100644 src/lib/check_text.js

diff --git a/README.md b/README.md
index 421535b..4d94fae 100644
--- a/README.md
+++ b/README.md
@@ -103,6 +103,19 @@ If you're debugging the feed compatibility code, then setting the `DEBUG_FEEDITE
 
 If you're working on the CSS, note that Eleventy will *not* auto-rebuild & reload the browser. For that, you need to re-save a file it *does* watch, such as `index.html`, `.eleventy.js`, etc (even if you haven't made any changes).
 
+### Environment Variables
+The following environment variables can be used to influence the behaviour of the aggregator.
+
+Environment Variable	| Purpose
+------------------------|------------------------------
+`FILTER_FEED`			| Set to any value to enable the filtering of posts based on whether they contain any bad words. Uses [futility](https://www.npmjs.com/package/futility), but with a custom algorithm on top to avoid the "[Scunthorpe Problem](https://en.wikipedia.org/wiki/Scunthorpe_problem)".
+
+Example of setting an environment variable when running a build:
+
+```bash
+FILTER_FEED=true npm run build
+```
+
 
 ### Design Decisions
 A number of decisions were made in the design process of this website. These are documented with the reasoning behind them here.
diff --git a/package-lock.json b/package-lock.json
index fc04bbf..b737bb2 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -12,6 +12,7 @@
 				"dateformat": "^4.5.1",
 				"feedme": "^2.0.2",
 				"file-type": "^16.5.3",
+				"futility": "^1.1.2",
 				"html-entities": "^2.3.2",
 				"p-reflect": "^2.1.0",
 				"phin": "^3.6.0",
@@ -1217,6 +1218,11 @@
 			"resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.1.tgz",
 			"integrity": "sha512-yIovAzMX49sF8Yl58fSCWJ5svSLuaibPxXQJFLmBObTuCr0Mf1KiPopGM9NiFjiYBCbfaa2Fh6breQ6ANVTI0A=="
 		},
+		"node_modules/futility": {
+			"version": "1.1.2",
+			"resolved": "https://registry.npmjs.org/futility/-/futility-1.1.2.tgz",
+			"integrity": "sha1-w55ZYFAH8jnL7Vpu/RZRuFiHy/s="
+		},
 		"node_modules/get-caller-file": {
 			"version": "2.0.5",
 			"resolved": "https://registry.npmjs.org/get-caller-file/-/get-caller-file-2.0.5.tgz",
@@ -4598,6 +4604,11 @@
 			"resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.1.tgz",
 			"integrity": "sha512-yIovAzMX49sF8Yl58fSCWJ5svSLuaibPxXQJFLmBObTuCr0Mf1KiPopGM9NiFjiYBCbfaa2Fh6breQ6ANVTI0A=="
 		},
+		"futility": {
+			"version": "1.1.2",
+			"resolved": "https://registry.npmjs.org/futility/-/futility-1.1.2.tgz",
+			"integrity": "sha1-w55ZYFAH8jnL7Vpu/RZRuFiHy/s="
+		},
 		"get-caller-file": {
 			"version": "2.0.5",
 			"resolved": "https://registry.npmjs.org/get-caller-file/-/get-caller-file-2.0.5.tgz",
diff --git a/package.json b/package.json
index c6c8aa1..3c542d4 100644
--- a/package.json
+++ b/package.json
@@ -25,6 +25,7 @@
 		"dateformat": "^4.5.1",
 		"feedme": "^2.0.2",
 		"file-type": "^16.5.3",
+		"futility": "^1.1.2",
 		"html-entities": "^2.3.2",
 		"p-reflect": "^2.1.0",
 		"phin": "^3.6.0",
diff --git a/src/index.11tydata.js b/src/index.11tydata.js
index 1107f10..0f455da 100644
--- a/src/index.11tydata.js
+++ b/src/index.11tydata.js
@@ -6,6 +6,7 @@ const dateformat = require("dateformat");
 const striptags = require("striptags");
 
 const fetch_feed = require("./lib/fetch_feed.js");
+const check_text = require("./lib/check_text.js");
 
 // The length of auto-generated descriptions if one isn't provided.
 const DESCRIPTION_LENGTH = 200;
@@ -72,6 +73,15 @@ module.exports = async function() {
 		})));
 		global.feed_items.sort((a, b) => b.pubdate_obj - a.pubdate_obj);
 		// console.log(feed_items.map(el => el.title));
+		
+		if(process.env.FILTER_FEED) {
+			const feed_items_count = global.feed_items.length;
+			global.feed_items = global.feed_items.filter((item) => !check_text(item.title)
+				&& !check_text(item.description)
+				&& !check_text(item.content));
+			
+			console.log(`>>> Feed filtering enabled, removed ${feed_items_count - global.feed_items.length} posts`);
+		}
 	}
 	
 	return {
diff --git a/src/lib/check_text.js b/src/lib/check_text.js
new file mode 100644
index 0000000..b7201e9
--- /dev/null
+++ b/src/lib/check_text.js
@@ -0,0 +1,24 @@
+const Futility = require("futility").default;
+const striptags = require("striptags");
+
+const checker = new Futility();
+
+/**
+ * Checks the words in the given string for bad (swear) words.
+ * Checks on a word level - therefore avoiding the "scunthorpe problem".
+ * @param	{string}	str		The string to check.
+ * @return	{boolean}	Truee if the given string contains any bad words, or false if it does not.
+ */
+function check_text(str) {
+	const words = striptags(str).split(/\s+/);
+	
+	for(let word of words) {
+		if(word.trim().search(/^\*+$/) > -1) continue;
+		let replaced = checker.censor(word, "*");
+		if(replaced.trim().search(/^\*+$/) > -1)
+			return true;
+	}
+	return false;
+}
+
+module.exports = check_text;