From fe0d65cf5a634ad1b3933bb745145a316e31e5c7 Mon Sep 17 00:00:00 2001
From: Offray Luna <offray.luna@mutabit.com>
Date: Tue, 9 May 2023 14:45:23 -0500
Subject: [PATCH] Data cleaning methods.

---
 src/MiniDocs/Markdeep.class.st | 37 ++++++++++++++++++++++++++++++++++
 src/MiniDocs/Markdown.class.st | 18 +++++++++++++++++
 src/MiniDocs/PubPub.class.st   | 30 ---------------------------
 3 files changed, 55 insertions(+), 30 deletions(-)
 delete mode 100644 src/MiniDocs/PubPub.class.st

diff --git a/src/MiniDocs/Markdeep.class.st b/src/MiniDocs/Markdeep.class.st
index c650a88..26e4c4b 100644
--- a/src/MiniDocs/Markdeep.class.st
+++ b/src/MiniDocs/Markdeep.class.st
@@ -393,6 +393,43 @@ Markdeep >> pubPubRawLinks [
 	^ (parser parse: self body)
 ]
 
+{ #category : #accessing }
+Markdeep >> removeAlternativeImagesArray [
+	| parser replacements |
+	self body ifNil: [^ self].
+	parser := PubPubGrammar2 new.
+	replacements := parser parse: body.
+	replacements ifEmpty: [^self].
+	replacements do: [:replacement | 
+		self body: 
+			(self body copyReplaceAll: replacement third with: '' )
+	].
+	self body: (self body copyReplaceAll: '{srcset=}' with: '').
+]
+
+{ #category : #accessing }
+Markdeep >> removeAutoGeneratedFileNotice [
+	| autoGeneratedNotice |
+	autoGeneratedNotice := '**Notice:** This file is an auto-generated download and, as such, might
+include minor display or rendering errors. For the version of record,
+please visit the HTML version or download the PDF.
+
+------------------------------------------------------------------------'.
+	self body: (self body copyReplaceAll: autoGeneratedNotice with: '')
+]
+
+{ #category : #accessing }
+Markdeep >> removeCCByLicenseDiv [
+	| licenseDiv|
+	licenseDiv := '<div>
+
+**License:** [Creative Commons Attribution 4.0 International License
+(CC-BY 4.0)](https://creativecommons.org/licenses/by/4.0/)
+
+</div>'.
+	self body: (self body copyReplaceAll: licenseDiv with: '')
+]
+
 { #category : #accessing }
 Markdeep >> replaceBackslashBreaklines [
 	self bodyReplaceAll: '\
diff --git a/src/MiniDocs/Markdown.class.st b/src/MiniDocs/Markdown.class.st
index 1117488..bcbe8c3 100644
--- a/src/MiniDocs/Markdown.class.st
+++ b/src/MiniDocs/Markdown.class.st
@@ -25,6 +25,24 @@ Markdown class >> yamlMetadataDelimiter [
 
 ]
 
+{ #category : #accessing }
+Markdown >> asMarkdeep [
+	^ Markdeep new 
+		body: self body;
+		markdownFile: self file;
+		commentYAMLMetadata
+]
+
+{ #category : #accessing }
+Markdown >> body [
+	^ body
+]
+
+{ #category : #accessing }
+Markdown >> body: aString [
+	body := aString
+]
+
 { #category : #operation }
 Markdown >> commentYAMLMetadata [
 	| newContents |
diff --git a/src/MiniDocs/PubPub.class.st b/src/MiniDocs/PubPub.class.st
deleted file mode 100644
index e50b328..0000000
--- a/src/MiniDocs/PubPub.class.st
+++ /dev/null
@@ -1,30 +0,0 @@
-Class {
-	#name : #PubPub,
-	#superclass : #Markdown,
-	#instVars : [
-		'url'
-	],
-	#category : #'MiniDocs-Core'
-}
-
-{ #category : #accessing }
-PubPub >> detectImageLinkPositions [
-	| docNodes response |
-	docNodes := self documentTree allChildren.
-	response := OrderedCollection new.
-	docNodes doWithIndex: [:each :i| 
-		((each className = 'PPCMLink') and: [ (docNodes at: i - 1) text = '!' ]) ifTrue:
-		[	response add: {i -> each } ]
-	].
-	^ response
-]
-
-{ #category : #accessing }
-PubPub >> linksWithTreePosition [
-	| response |
-	response := OrderedCollection new.
-	self documentTree allChildren doWithIndex: [:each :i | 
-		each className = 'PPCMLink' ifTrue: [ response add: { i -> each }] 
-	].
-	^ response
-]