[16.0][IMP] mail_embed_image: improve embedding type:

 - CIDs are not working in some email managers (gmail and office365)
   To allow this behavior we need to change the content type to
   multipart/related
 - A new option to embed images is to include the base64 content inside
   the src.
 - Then to select these options, a selected field has been added to company

mail_embed_image/__manifest__.py

@@ -1,4 +1,5 @@
 # Copyright 2019 Therp BV <https://therp.nl>
+# Copyright 2024 Camptocamp SA
 # License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl.html).
 {
     "name": "Mail Embed Image",
@@ -9,8 +10,12 @@
     "summary": "Replace img.src's which start with http with inline cids",
     "website": "https://github.com/OCA/social",
     "depends": [
+        "mail",
         "web",
     ],
+    "data": [
+        "views/res_config_settings_views.xml",
+    ],
     "installable": True,
     "application": False,
 }

mail_embed_image/models/__init__.py

@@ -1,3 +1,6 @@
 # Copyright 2019 Therp BV <https://therp.nl>
+# Copyright 2024 Camptocamp SA
 # License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl.html).
 from . import ir_mail_server
+from . import company
+from . import res_config_settings

mail_embed_image/models/company.py

@@ -0,0 +1,17 @@
+# Copyright 2024 Camptocamp SA
+# License AGPL-3.0 or later (http://www.gnu.org/licenses/agpl).
+from odoo import fields, models
+
+
+class ResCompany(models.Model):
+    _inherit = "res.company"
+
+    image_embedding_method = fields.Selection(
+        selection=[
+            ("none", "No attachment"),
+            ("cid", "CIDs attachment"),
+            ("data", "Data SRC"),
+        ],
+        default="cid",  # previous module version only supported CID
+        required=True,
+    )

mail_embed_image/models/ir_mail_server.py

@@ -1,3 +1,7 @@
+# Copyright 2019 Therp BV <https://therp.nl>
+# Copyright 2024 Camptocamp SA
+# License AGPL-3.0 or later (http://www.gnu.org/licenses/agpl).
+
 import logging
 import uuid
 from base64 import b64encode
@@ -32,8 +36,9 @@ def build_email(
         body_alternative=None,
         subtype_alternative="plain",
     ):
+        image_embedding_method = self.env.company.image_embedding_method
         fileparts = None
-        if subtype == "html":
+        if subtype == "html" and image_embedding_method != "none":
             body, fileparts = self._build_email_replace_img_src(body)
         result = super(IrMailServer, self).build_email(
             email_from=email_from,
@@ -55,30 +60,45 @@ def build_email(
         if fileparts:
             for fpart in fileparts:
                 result.attach(fpart)
+            # Multipart method MUST be multipart/related for CIDs embedding
+            # Gmail and Office won't process the images otherwise
+            if image_embedding_method == "cid":
+                result.set_type("multipart/related")
         return result
 
     def _build_email_replace_img_src(self, html_body):
         """Replace img src with base64 encoded image."""
         if not html_body:
             return html_body
 
+        base_url = self.env["ir.config_parameter"].get_param("web.base.url")
+        image_embedding_method = self.env.company.image_embedding_method
         root = fromstring(html_body)
-        images = root.xpath("//img")
         fileparts = []
-        for img in images:
-            src = img.get("src")
-            if src and not src.startswith("data:") and not src.startswith("base64:"):
-                try:
-                    response = requests.get(src, timeout=10)
-                    _logger.debug("Fetching image from %s", src)
-                    if response.status_code == 200:
+        # Limit results to only internal resources to avoid malicious external
+        # image injections
+        for img in root.xpath(
+            ".//img[starts-with(@src, '%s')]"
+            "| .//img[starts-with(@src, '/web/image')]" % (base_url)
+        ):
+            image_path = img.get("src")
+            try:
+                response = requests.get(image_path, timeout=10)
+                _logger.debug("Fetching image from %s", image_path)
+                if response.status_code == 200:
+                    image_content = response.content
+                    filepart = MIMEImage(image_content)
+                    if image_embedding_method == "data":
+                        raw_content = filepart.get_payload(decode=True)
+                        base_64_content = b64encode(raw_content).decode("utf-8")
+                        mimetype = filepart.get_content_type()
+                        img.set("src", f"data:{mimetype};base64,{base_64_content}")
+                    elif image_embedding_method == "cid":
                         cid = uuid.uuid4().hex
                         # convert cid to rfc2047 encoding
                         filename_encoded = "=?utf-8?b?%s?=" % b64encode(
                             cid.encode("utf-8")
                         ).decode("utf-8")
-                        image_content = response.content
-                        filepart = MIMEImage(image_content)
                         filepart.add_header("Content-ID", f"<{cid}>")
                         filepart.add_header(
                             "Content-Disposition",
@@ -87,6 +107,12 @@ def _build_email_replace_img_src(self, html_body):
                         )
                         img.set("src", f"cid:{cid}")
                         fileparts.append(filepart)
-                except Exception as e:
-                    _logger.warning("Could not get %s: %s", img.get("src"), str(e))
+                else:
+                    _logger.warning(
+                        "Could not get %s: HTTP status code %s",
+                        img.get("src"),
+                        response.status_code,
+                    )
+            except Exception as e:
+                _logger.warning("Could not get %s: %s", img.get("src"), str(e))
         return tostring(root, encoding="unicode"), fileparts

mail_embed_image/models/res_config_settings.py

@@ -0,0 +1,12 @@
+# Copyright 2024 Camptocamp SA
+# License AGPL-3.0 or later (http://www.gnu.org/licenses/agpl).
+from odoo import fields, models
+
+
+class ResConfigSettings(models.TransientModel):
+    _inherit = "res.config.settings"
+
+    image_embedding_method = fields.Selection(
+        related="company_id.image_embedding_method",
+        readonly=False,
+    )

mail_embed_image/tests/test_mail_embed_image.py

@@ -1,6 +1,6 @@
 # Copyright 2019 Therp BV <https://therp.nl>
 # License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl.html).
-from base64 import b64encode
+import base64
 
 from lxml import html
 from requests import get
@@ -9,14 +9,23 @@
 
 
 class TestMailEmbedImage(common.TransactionCase):
-    def test_mail_embed_image(self):
-        """We pass a mail with <img src="..." /> tags to build_email,
-        and then look into the result, check there were attachments
-        created and you find xpaths like //img[src] have a cid"""
-        # DATA
-        base_url = self.env["ir.config_parameter"].get_param("web.base.url")
-        image_url = base_url + "/mail_embed_image/static/description/icon.png"
-        image = get(image_url, timeout=10).content
+    @classmethod
+    def setUpClass(cls):
+        super(TestMailEmbedImage, cls).setUpClass()
+        cls.company = cls.env.ref("base.main_company")
+        base_url = cls.env["ir.config_parameter"].get_param("web.base.url")
+        cls.image_url = base_url + "/mail_embed_image/static/description/icon.png"
+        cls.image_content = get(cls.image_url, timeout=10).content
+        cls.email_from = "test@example.com"
+        cls.email_to = "test@example.com"
+        cls.subject = "test mail"
+
+    def build_email(self, option="cid"):
+        """Build an email with a given embedding option
+
+        option -- the embedding option to use according to the company setting
+        """
+        self.company.image_embedding_method = option
         body = html.tostring(
             html.fromstring(
                 """
@@ -27,24 +36,61 @@ def test_mail_embed_image(self):
             </div>"""
                 % (
                     # won't be hit because we ignore embedded images
-                    b64encode(image),
+                    base64.b64encode(self.image_content).decode("utf-8"),
                     # dito, not uploaded content
-                    image_url,
+                    self.image_url,
                 )
             )
         )
-        email_from = "test@example.com"
-        email_to = "test@example.com"
-        subject = "test mail"
-        # END DATA
-        res = self.env["ir.mail_server"].build_email(
-            email_from,
-            [email_to],
-            subject,
+        return self.env["ir.mail_server"].build_email(
+            self.email_from,
+            [self.email_to],
+            self.subject,
             body,
             subtype="html",
             subtype_alternative="plain",
         )
+
+    def test_mail_embed_image_option_none(self):
+        """No embedding option
+
+        We pass a mail with <img src="..." /> tags to build_email,
+        and then look into the result, check there no changes were made"""
+        res = self.build_email("none")
+        images_in_mail = 0
+        for part in res.walk():
+            if part.get_content_type() == "text/html":
+                # we do not search in text, just in case that texts exists in
+                # the text elsewhere (not probable, but this is better)
+                images_in_mail += len(
+                    html.fromstring(part.get_payload(decode=True)).xpath(
+                        "//img[starts-with(@src, 'data:image/png;base64,')]"
+                    )
+                )
+                images_in_mail += len(
+                    html.fromstring(part.get_payload(decode=True)).xpath(
+                        "//img[starts-with(@src, 'cid:')]"
+                    )
+                )
+        # verify 0 replaced images
+        self.assertEqual(images_in_mail, 0)
+        # verify 0 attachment present
+        self.assertEqual(
+            [
+                x.get_content_type()
+                for x in res.walk()
+                if x.get_content_type().startswith("image/")
+            ],
+            [],
+        )
+
+    def test_mail_embed_image_option_cids(self):
+        """CIDs attachement option
+
+        We pass a mail with <img src="..." /> tags to build_email,
+        and then look into the result, check there were attachments
+        created and you find xpaths like //img[src] have a cid"""
+        res = self.build_email("cid")
         images_in_mail = 0
         for part in res.walk():
             if part.get_content_type() == "text/html":
@@ -66,3 +112,32 @@ def test_mail_embed_image(self):
             ],
             ["image/png"],
         )
+
+    def test_mail_embed_image_option_data(self):
+        """Data URL option
+
+        We pass a mail with <img src="..." /> tags to build_email,
+        and then look into the result, check there were attachments
+        created and you find xpaths like //img[src] have a data URL"""
+        res = self.build_email("data")
+        images_in_mail = 0
+        for part in res.walk():
+            if part.get_content_type() == "text/html":
+                # we do not search in text, just in case that texts exists in
+                # the text elsewhere (not probable, but this is better)
+                images_in_mail += len(
+                    html.fromstring(part.get_payload(decode=True)).xpath(
+                        "//img[starts-with(@src, 'data:image/png;base64,')]"
+                    )
+                )
+        # verify 2 replaced image
+        self.assertEqual(images_in_mail, 1)
+        # verify 0 attachment present
+        self.assertEqual(
+            [
+                x.get_content_type()
+                for x in res.walk()
+                if x.get_content_type().startswith("image/")
+            ],
+            [],
+        )

mail_embed_image/views/res_config_settings_views.xml

@@ -0,0 +1,26 @@
+<?xml version="1.0" encoding="utf-8" ?>
+<odoo>
+
+    <record id="res_config_settings_view_form" model="ir.ui.view">
+        <field name="name">res.config.settings.view.form.inherit.mail</field>
+        <field name="model">res.config.settings</field>
+        <field name="inherit_id" ref="mail.res_config_settings_view_form" />
+        <field name="arch" type="xml">
+            <div id="companies_setting" position="inside">
+                <br />
+                <div
+                    class="o_setting_right_pane"
+                    id="mail_templates_setting"
+                    groups="mail.group_mail_template_editor,base.group_system"
+                >
+                    <span class="o_form_label">Email Preprocessing</span>
+                    <div class="text-muted">
+                        Method used to embed images in HTML emails. CIDs attachment does not work with all email clients. Data SRC is more reliable.
+                    </div>
+                    <field name="image_embedding_method" />
+                </div>
+            </div>
+        </field>
+    </record>
+
+</odoo>