phuoc/html_cleanup
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fixed cleanup function

Browse Source
This commit is contained in:
Phuoc
2023-12-03 10:38:17 +01:00
parent 82436da5d4
commit 99b78a9ddf
6 changed files with 19 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
html_cleanup/static/temp/clean_Nantes_anthologie_cinema_viet.htm
View File

@@ -337,7 +337,11 @@ Mercredi 29.11, 14:45</p>
<p align="justify"> <p align="justify"><a
href="https://www.3continents.com/fr/programme/2023/anthologie-du-cinema-vietnamien/"
target="_blank">
https://www.3continents.com/fr/programme/2023/anthologie-du-cinema-vietnamien/</a></p> https://www.3continents.com/fr/programme/2023/anthologie-du-cinema-vietnamien/</a></p>
@@ -346,8 +350,9 @@ https://www.3continents.com/fr/programme/2023/anthologie-du-cinema-vietnamien/</
<p align="justify"> <p align="justify"><a
https://www.3continents.com/wp-content/uploads/f3c-prog-2023-40p-net-planche.pdf</a></p>
href="https://www.3continents.com/wp-content/uploads/f3c-prog-2023-40p-net-planche.pdf">https://www.3continents.com/wp-content/uploads/f3c-prog-2023-40p-net-planche.pdf</a></p>

5
html_cleanup/static/temp/clean_proposed_file_name.html
View File

@@ -1,5 +0,0 @@
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>untitled</title>
</head>

5
html_cleanup/static/temp/clean_www.icloud.com.html
View File

@@ -1,5 +0,0 @@
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>untitled</title>
</head>

12
html_cleanup/static/temp/errors_log
View File

@@ -1,11 +1 @@
2023-11-29 17:22:16 [ Nantes_anthologie_cinema_viet.htm ] *** Cleanup réussi. *** 03-12-2023 10:35 [ Nantes_anthologie_cinema_viet.htm ] *** Cleanup réussi. ***
2023-11-29 17:23:56 [ Screenshot 2023-11-20 at 12.07.54.png ] Le format du fichier n'est pas valide. Téléchargement refusé.
30-11-2023 17:29 [ Nantes_anthologie_cinema_viet.htm ] *** Cleanup réussi. ***
30-11-2023 18:32 [ www.icloud.com.html ] *** Cleanup réussi. ***
02-12-2023 10:41 [ Nantes_anthologie_cinema_viet.htm ] Le format du fichier n'est pas valide (absence de tag <body> ou </body>). Téléchargement refusé.
02-12-2023 10:44 [ Nantes_anthologie_cinema_viet.htm ] *** Cleanup réussi. ***
02-12-2023 10:45 [ cleanup_html.vbs.txt ] Le format du fichier n'est pas valide (absence de tag <body> ou </body>). Téléchargement refusé.
02-12-2023 10:49 [ Nantes_anthologie_cinema_viet.htm ] *** Cleanup réussi. ***
02-12-2023 11:02 [ Nantes_anthologie_cinema_viet copy.htm ] *** Cleanup réussi. ***
02-12-2023 11:10 [ Nantes_anthologie_cinema_viet copy.htm ] *** Cleanup réussi. ***
02-12-2023 11:11 [ Nantes_anthologie_cinema_viet.htm ] *** Cleanup réussi. ***

2
html_cleanup/templates/view_log.jinja2
View File

@@ -11,7 +11,7 @@
{% if not lines %} {% if not lines %}
<div class="form-group"> <div class="form-group">
<input type="text" name="username" class="form-control" placeholder="Identifiant Zidol"> <input type="text" name="username" class="form-control" placeholder="Identifiant ZD">
</div> </div>
<div class="form-group"> <div class="form-group">

11
html_cleanup/views/default.py
View File

@@ -136,13 +136,19 @@ def clean_html(input_file, output_file):
done = True done = True
skip_tag = False skip_tag = False
tag = s.split(" ") tag = s.split()
if tag: if tag:
tag[0] = tag[0].lower() tag[0] = tag[0].lower()
if tag[0] in ["table", "div", "img", "a"]: if tag[0] in ["table", "div", "img", "a"]:
line_out += "<{}>".format(s) line_out += "<{}".format(s)
if end_tag > 0:
# si end tag trouvé, fermer
line_out += ">"
else:
# sinon contine de recopier la ligne
skip_tag = False
elif tag[0] in ["p"]: elif tag[0] in ["p"]:
line_out += "<p align=\"justify\">" line_out += "<p align=\"justify\">"
elif tag[0] in ["b", "/b", "em", "/em", "i", "/i", "li", "li/", "/li", elif tag[0] in ["b", "/b", "em", "/em", "i", "/i", "li", "li/", "/li",
@@ -195,6 +201,7 @@ def view_log(request):
lines = file.readlines() lines = file.readlines()
file.close() file.close()
except Exception as e: except Exception as e:
message = "Le fichier Journal est vide !"
pass pass
# lire tous les fichiers du dossier triés par date # lire tous les fichiers du dossier triés par date