From d00f19145627312125c593f35193f04733b4df4e Mon Sep 17 00:00:00 2001
From: Phil Burton <phil@d3r.com>
Date: Tue, 11 Jun 2019 14:17:29 +0100
Subject: First commit

---
 src/Etym.php | 125 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 125 insertions(+)
 create mode 100644 src/Etym.php

(limited to 'src/Etym.php')

diff --git a/src/Etym.php b/src/Etym.php
new file mode 100644
index 0000000..b796a3a
--- /dev/null
+++ b/src/Etym.php
@@ -0,0 +1,125 @@
+<?php
+
+namespace App;
+
+use PHPHtmlParser\Dom;
+
+/**
+ * Get the etymology of a word
+ *
+ * @author Phil Burton <phil@pgburton.com>
+ */
+class Etym
+{
+    protected $baseURL;
+    protected $domSearch;
+    protected $pasteCmd;
+
+    /**
+     * Initalise with config options
+     *
+     * @param string $baseURL
+     * @param string $domSearch
+     * @param string $pasteCmd
+     * @author Phil Burton <phil@pgburton.com>
+     */
+    public function __construct(string $baseURL, string $domSearch, string $pasteCmd)
+    {
+        $this->baseURL = $baseURL;
+        $this->domSearch = $domSearch;
+        $this->pasteCmd = $pasteCmd;
+    }
+
+    /**
+     * Search for and return etym defeiniton
+     *
+     * @return string
+     * @author Phil Burton <phil@d3r.com>
+     */
+    public function getDefinition(): string
+    {
+        $input = $this->readStdin();
+
+        $targetURL = $this->baseURL . $input;
+
+        $dom = new Dom;
+        $dom->load($targetURL);
+        $html = $dom->find($this->domSearch);
+
+        if (!$html instanceof \PHPHtmlParser\Dom\Collection || count($html) <= 0) {
+            echo "No entry found for '$input'!\n";
+            exit(2);
+        }
+
+        $definition = $this->cleanUpHtml($html[0]->innerHtml());
+        $definition = $this->handleTruncation($input, $definition, $targetURL);
+
+        // echo $definition . "\n";
+        return $definition;
+    }
+
+    /**
+     * Get and return the user input
+     *
+     * @return string
+     * @author Phil Burton <phil@pgburton.com>
+     */
+    protected function readStdin(): string
+    {
+        $input = fgets(STDIN);
+
+        if ($input === false) {
+            echo "No input supplied!\n";
+            exit(1);
+        }
+
+        $input = rtrim($input, "\n");
+
+        return $input;
+    }
+
+    /**
+     * Tidy and return the hmtl
+     *
+     * @param string $input
+     * @return string
+     * @author Phil Burton <phil@pgburton.com>
+     */
+    protected function cleanUpHtml(string $input)
+    {
+        // Strip HTML Tags
+        $clear = strip_tags($input);
+        // Clean up things like &amp;
+        $clear = html_entity_decode($clear);
+        // Strip out any url-encoded stuff
+        $clear = urldecode($clear);
+        // Replace Multiple spaces with single space
+        $clear = preg_replace('/ +/', ' ', $clear);
+        // Trim the string of leading/trailing space
+        $clear = trim($clear);
+        // Capitalise the first char.
+        $clear = ucfirst($clear);
+
+        return $clear;
+    }
+
+    protected function handleTruncation($input, $definition, $url)
+    {
+        $truncated = $definition;
+
+        $MAX_CHARACTERS = 350;
+        if (strlen($truncated) >= $MAX_CHARACTERS) {
+            // Create that povjee link.
+            // Capitalise the first char of the input.
+            $input   = ucfirst($input);
+            $defAndUrl = "\"$input\"" . "\n\n" . $definition . "\n\n" . "[Original at: $url]";
+            $safeDef = escapeshellarg($defAndUrl);
+
+            $pasteBinCmd = "echo $safeDef | " . $this->pasteCmd;
+            $pasteBinLink = exec($pasteBinCmd);
+            $truncated = substr($truncated, 0, $MAX_CHARACTERS) . "... [More info at $pasteBinLink]";
+        }
+
+        return $truncated;
+    }
+}
-- 
cgit v1.2.3