[FIX] attachement_indexation: make pdfminer optional

As pdfminer does not have a Debian package in Ubuntu Bionic, it cannot
be declared as a strong requirement.

With this commit, a warning is logged if the library is not installed.
It does not prevent to index other types of documents.

closes odoo/odoo#44327

Signed-off-by: Christophe Monniez (moc) <moc@odoo.com>

addons/attachment_indexation/__manifest__.py

@@ -8,7 +8,9 @@
 Attachments list and document indexation
 ========================================
 * Show attachment on the top of the forms
-* Document Indexation: odt
+* Document Indexation: odt, pdf, xlsx, docx
+
+The `pdfminer.six` Python library has to be installed in order to index PDF files
 """,
     'depends': ['web'],
     'installable': True,

addons/attachment_indexation/models/ir_attachment.py

@@ -4,15 +4,23 @@ import io
 import logging
 import xml.dom.minidom
 import zipfile
-from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
-from pdfminer.converter import TextConverter
-from pdfminer.pdfpage import PDFPage
 
 from odoo import api, models
 
 _logger = logging.getLogger(__name__)
+
+try:
+    from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
+    from pdfminer.converter import TextConverter
+    from pdfminer.pdfpage import PDFPage
+except ImportError:
+    PDFResourceManager = PDFPageInterpreter = TextConverter = PDFPage = None
+    _logger.warning("Attachment indexation of PDF documents is unavailable because the 'pdfminer' Python library cannot be found on the system. "
+                    "You may install it from https://pypi.org/project/pdfminer.six/ (e.g. `pip3 install pdfminer.six`)")
+
 FTYPES = ['docx', 'pptx', 'xlsx', 'opendoc', 'pdf']
 
+
 def textToString(element):
     buff = u""
     for node in element.childNodes:
@@ -93,6 +101,8 @@ class IrAttachment(models.Model):
 
     def _index_pdf(self, bin_data):
         '''Index PDF documents'''
+        if PDFResourceManager is None:
+            return
         buf = u""
         if bin_data.startswith(b'%PDF-'):
             f = io.BytesIO(bin_data)

addons/attachment_indexation/tests/test_indexation.py

@@ -1,14 +1,21 @@
 # -*- coding: utf-8 -*-
 
 from odoo.tests.common import TransactionCase, tagged
+from unittest import skipIf
 import os
 
 directory = os.path.dirname(__file__)
 
+try:
+    from pdfminer.pdfinterp import PDFResourceManager
+except ImportError:
+    PDFResourceManager = None
+
 
 @tagged('post_install', '-at_install')
 class TestCaseIndexation(TransactionCase):
 
+    @skipIf(PDFResourceManager is None, "pdfminer not installed")
     def test_attachment_pdf_indexation(self):
         with open(os.path.join(directory, 'files', 'test_content.pdf'), 'rb') as file:
             pdf = file.read()

requirements.txt

@@ -21,7 +21,6 @@ mock==2.0.0
 num2words==0.5.6
 ofxparse==0.19
 passlib==1.7.1
-pdfminer.six==20181108
 Pillow==5.4.1 ; python_version < '3.7' or sys_platform != 'win32'
 Pillow==6.1.0 ; sys_platform == 'win32' and python_version >= '3.7'
 polib==1.1.0