[IMP] attachment_indexation: improved PDF text extraction

PyPDF performs badly on many types of PDF documents.
We add a text extraction with pdfminer, which is designed for this task.
Because pdf content extraction was so flaky, it was completely
deactivated by 1b753b0d53744a6e. We revert that :-)

closes odoo/odoo#38508

Task: 2152494
Signed-off-by: Sébastien Theys (seb) <seb@odoo.com>

3 files changed, 19 insertions, 0 deletions

diff --git a/addons/attachment_indexation/tests/__init__.py b/addons/attachment_indexation/tests/__init__.py
new file mode 100644
index 00000000000..d76ba91978d
--- /dev/null
+++ b/addons/attachment_indexation/tests/__init__.py
@@ -0,0 +1,3 @@
+# -*- coding: utf-8 -*-
+
+from . import test_indexation
diff --git a/addons/attachment_indexation/tests/files/test_content.pdf b/addons/attachment_indexation/tests/files/test_content.pdf
new file mode 100644
index 00000000000..062e1e6e2f6
--- /dev/null
+++ b/addons/attachment_indexation/tests/files/test_content.pdf
diff --git a/addons/attachment_indexation/tests/test_indexation.py b/addons/attachment_indexation/tests/test_indexation.py
new file mode 100644
index 00000000000..9f5fa0eca15
--- /dev/null
+++ b/addons/attachment_indexation/tests/test_indexation.py
@@ -0,0 +1,16 @@
+# -*- coding: utf-8 -*-
+
+from odoo.tests.common import TransactionCase, tagged
+import os
+
+directory = os.path.dirname(__file__)
+
+
+@tagged('post_install', '-at_install')
+class TestCaseIndexation(TransactionCase):
+
+    def test_attachment_pdf_indexation(self):
+        with open(os.path.join(directory, 'files', 'test_content.pdf'), 'rb') as file:
+            pdf = file.read()
+            text = self.env['ir.attachment']._index(pdf, 'application/pdf')
+            self.assertEqual(text, 'TestContent!!\x0c', 'the index content should be correct')