From 4f128f90759b7ec44e81432a91d28e9cd4362004 Mon Sep 17 00:00:00 2001
From: Liubov Yaronskaya <lyaronskaya@sourcerer.io>
Date: Thu, 7 Nov 2019 13:40:35 +0300
Subject: [PATCH] feat: add php project heuristics(bootstrap, illuminate)

---
 src/main/kotlin/app/extractors/Heuristics.kt | 22 ++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)

diff --git a/src/main/kotlin/app/extractors/Heuristics.kt b/src/main/kotlin/app/extractors/Heuristics.kt
index 6da1eae5..70ce3be4 100644
--- a/src/main/kotlin/app/extractors/Heuristics.kt
+++ b/src/main/kotlin/app/extractors/Heuristics.kt
@@ -122,6 +122,9 @@ val Perl6Regex = Regex(
 val PhpRegex = Regex(
     "^<\\?(?:php)?"
 )
+val PhpIlluminateRegex = Regex(
+    "(Auth|Bootstrap|Bus|Console|Events|Exceptions|Http|Providers" +
+                "|Support|Testing|Validation)")
 val PicoLispRegex = Regex(
     "^\\((de|class|rel|code|data|must)\\s",
     RegexOption.MULTILINE
@@ -235,6 +238,11 @@ val k8sExp = { buf: String ->
         && buf.contains("metadata")
 }
 
+val bootstrapWebpackExp = { buf: String ->
+    // Handling php projects with bootstrap Webpack.
+    buf.contains("/******/ (function(modules) { // webpackBootstrap")
+}
+
 /**
  * Heuristics to detect a programming language by file extension and content.
  * Inspired by GitHub Liguist heuristics (https://github.com/github/linguist).
@@ -638,8 +646,11 @@ val HeuristicsMap = mapOf<String, (String, String) -> ExtractorInterface?>(
     "jl" to { _, _ ->
         CommonExtractor(Lang.JULIA)
     },
-    "js" to { _, _ ->
-        JavascriptExtractor()
+    "js" to { buf, _ ->
+        if (bootstrapWebpackExp(buf)) {
+            null
+        }
+        else JavascriptExtractor()
     },
     "jsx" to { _, _ ->
         JavascriptExtractor()
@@ -793,10 +804,13 @@ val HeuristicsMap = mapOf<String, (String, String) -> ExtractorInterface?>(
     "pde" to { _, _ ->
         CommonExtractor(Lang.PROCESSING)
     },
-    "php" to { buf, _ ->
+    "php" to { buf, path ->
         if (buf.contains("<?hh")) {
             CommonExtractor(Lang.HACK)
-        } else {
+        } else if (PhpIlluminateRegex.containsMatchIn(path)) {
+            null
+        }
+        else {
             PhpExtractor()
         }
     },