From fc2204a0fdd9a1bc70ee1bef3261c849497dd951 Mon Sep 17 00:00:00 2001
From: eric <eric@hellman.net>
Date: Sat, 1 Aug 2020 22:12:25 -0400
Subject: [PATCH 1/5] make sure cmp is not just chapters

---
 core/loaders/harvest.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/core/loaders/harvest.py b/core/loaders/harvest.py
index 9c4d3a7d..2b03326c 100644
--- a/core/loaders/harvest.py
+++ b/core/loaders/harvest.py
@@ -489,6 +489,9 @@ def harvest_fahce(ebook):
 
 def harvest_cmp(ebook):
     def selector(doc):
+    objs = doc.select('.tab-content a.cmp_download_link[href]')
+        if (len({obj['href'] for obj in objs})) > 1:
+            return []
         return doc.select('a.cmp_download_link[href]')
     def dl(url):
         return url.replace('view', 'download') + '?inline=1'

From 4839dc549a54590ede603037f43088de409dc544 Mon Sep 17 00:00:00 2001
From: eric <eric@hellman.net>
Date: Sat, 1 Aug 2020 22:12:43 -0400
Subject: [PATCH 2/5] fix cmp

---
 core/management/commands/fix_cmp.py | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
 create mode 100644 core/management/commands/fix_cmp.py

diff --git a/core/management/commands/fix_cmp.py b/core/management/commands/fix_cmp.py
new file mode 100644
index 00000000..9cbe9969
--- /dev/null
+++ b/core/management/commands/fix_cmp.py
@@ -0,0 +1,22 @@
+from django.core.management.base import BaseCommand
+
+from regluit.core.models import EbookFile
+
+class Command(BaseCommand):
+
+    def handle(self, **options):
+        prov = 'editorial.uniagustiniana.edu.co'
+        for ebook in Ebook.objects.filter(provider=prov, format='online'):
+            print(ebook.url)
+            doc = get_soup(ebook.url)
+            if doc:
+                objs = doc.select('.tab-content a.cmp_download_link[href]')
+                for obj in objs:
+                    for ebf in EbookFile.objects.filter(source=ebook.url):
+                        bad_ebook = ebf.ebook
+                        try:
+                            ebf.file.delete()
+                        except:
+                            pass
+                        ebf.delete()
+                        bad_ebook.delete()

From c5787bdd6e37091e3ce5089278136eef08d86a68 Mon Sep 17 00:00:00 2001
From: eric <eric@hellman.net>
Date: Sat, 1 Aug 2020 22:22:07 -0400
Subject: [PATCH 3/5] fix import

---
 core/management/commands/fix_cmp.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/core/management/commands/fix_cmp.py b/core/management/commands/fix_cmp.py
index 9cbe9969..2eb3c835 100644
--- a/core/management/commands/fix_cmp.py
+++ b/core/management/commands/fix_cmp.py
@@ -1,6 +1,6 @@
 from django.core.management.base import BaseCommand
 
-from regluit.core.models import EbookFile
+from regluit.core.models import EbookFile, Ebookfile
 
 class Command(BaseCommand):
 

From 7981b08abe5caabfa890daf80a0cc8d20f9b0bff Mon Sep 17 00:00:00 2001
From: eric <eric@hellman.net>
Date: Sat, 1 Aug 2020 22:43:31 -0400
Subject: [PATCH 4/5] fix

---
 core/management/commands/fix_cmp.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/core/management/commands/fix_cmp.py b/core/management/commands/fix_cmp.py
index 2eb3c835..9c8684f5 100644
--- a/core/management/commands/fix_cmp.py
+++ b/core/management/commands/fix_cmp.py
@@ -1,6 +1,7 @@
 from django.core.management.base import BaseCommand
 
-from regluit.core.models import EbookFile, Ebookfile
+from regluit.core.models import EbookFile, Ebook
+from regluit.core.loaders.utils import get_soup
 
 class Command(BaseCommand):
 
@@ -12,7 +13,7 @@ class Command(BaseCommand):
             if doc:
                 objs = doc.select('.tab-content a.cmp_download_link[href]')
                 for obj in objs:
-                    for ebf in EbookFile.objects.filter(source=ebook.url):
+                    for ebf in EbookFile.objects.filter(source=obj['href']):
                         bad_ebook = ebf.ebook
                         try:
                             ebf.file.delete()

From c63e20c3fe3b37d7fd355b1431cb038e78eb2b2c Mon Sep 17 00:00:00 2001
From: eric <eric@hellman.net>
Date: Sat, 1 Aug 2020 22:50:55 -0400
Subject: [PATCH 5/5] fix

---
 core/loaders/harvest.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/core/loaders/harvest.py b/core/loaders/harvest.py
index 2b03326c..49c8396a 100644
--- a/core/loaders/harvest.py
+++ b/core/loaders/harvest.py
@@ -489,7 +489,7 @@ def harvest_fahce(ebook):
 
 def harvest_cmp(ebook):
     def selector(doc):
-    objs = doc.select('.tab-content a.cmp_download_link[href]')
+        objs = doc.select('.tab-content a.cmp_download_link[href]')
         if (len({obj['href'] for obj in objs})) > 1:
             return []
         return doc.select('a.cmp_download_link[href]')