From 6fee512f809750baddbf36f857ecbce3c025f2ed Mon Sep 17 00:00:00 2001
From: cy_altern <cy.altern@gmail.com>
Date: Mon, 29 May 2023 18:43:34 +0200
Subject: [PATCH] Feat: pas de balises li et br pour la version CSV

---
 corpus_web_fonctions.php      | 17 ++++++++++++++++-
 json_affiche_corpus.json.html | 12 ++++++------
 2 files changed, 22 insertions(+), 7 deletions(-)

diff --git a/corpus_web_fonctions.php b/corpus_web_fonctions.php
index 395ea17..f81d832 100644
--- a/corpus_web_fonctions.php
+++ b/corpus_web_fonctions.php
@@ -330,7 +330,7 @@ function recup_liens_int($html) {
 		foreach($match as $m) {
 			$m[7] = preg_replace('/<(img|(\/)?span)[^>]*>/si', '', $m[7]);
 			// liens exclus
-			if(in_array($m[3], $url_exclus) || str_contains($m[3], '#comment') || str_contains($m[3], '/user/') || in_array($m[7], $txt_exclus) || str_replace([' ', '&nbsp;'], '', $m[7]) === '') {
+			if(in_array($m[3], $url_exclus) || str_contains($m[3], '#comment') || str_contains($m[3], '/user/') || str_contains($m[3], '/members/') || str_contains($m[3], '/subscription/') || in_array($m[7], $txt_exclus) || str_replace([' ', '&nbsp;'], '', $m[7]) === '') {
 				continue;
 			}
 			if (!array_key_exists($m[3], $liens_int)) {
@@ -347,3 +347,18 @@ function recup_liens_int($html) {
 
 	return count($l_int) ? "<ul><li>".join("</li><li>", $l_int)."</li></ul>" : '';
 }
+
+function sans_li($html) {
+	$html = str_replace(['<ul>', '</ul>', '<li>'], '', $html);
+	$html = str_replace('</li>', "\r\n", $html);
+	$html = str_replace('&nbsp;', " ", $html);
+		
+	return $html;
+}
+
+function sans_br($html) {
+	$html = str_replace('<br>', "\r\n", $html);
+	$html = str_replace('&nbsp;', " ", $html);
+
+	return $html;
+}
\ No newline at end of file
diff --git a/json_affiche_corpus.json.html b/json_affiche_corpus.json.html
index cf8bee7..147468a 100644
--- a/json_affiche_corpus.json.html
+++ b/json_affiche_corpus.json.html
@@ -94,12 +94,12 @@
 				"title": [(#GET{title}|json_encode)],
 				"desc": [(#GET{description}|json_encode)],
 				"keywords": [(#GET{keywords}|json_encode)],
-				"trackers": [(#GET{trackers}|json_encode)],
-				"rezos": [(#GET{rezos}|json_encode)],
-				"titres": [(#GET{titres}|json_encode)],
-				"liens_ext": [(#GET{liens_ext}|json_encode)],
-				"liens_int": [(#GET{liens_int}|json_encode)],
-				"text": [(#GET{text}|json_encode)]
+				"trackers": [(#GET{trackers}|sans_li|json_encode)],
+				"rezos": [(#GET{rezos}|sans_li|json_encode)],
+				"titres": [(#GET{titres}|sans_li|json_encode)],
+				"liens_ext": [(#GET{liens_ext}|sans_li|json_encode)],
+				"liens_int": [(#GET{liens_int}|sans_li|json_encode)],
+				"text": [(#GET{text}sans_br|json_encode)]
 			}
 		} [(#SET{id_jsonl, #GET{id_jsonl}|plus{1}})]
 	</BOUCLE_2lignes>