You’re right, same happens for me too.
Some testing
Summary
- Navigate to the image on your target site.
- Right click → Copy image.
xclip -selection clipboard -o -t TARGETS shows a list of available targets in the clipboard:
The list of tagets in clipboard.
TIMESTAMP
TARGETS
MULTIPLE
SAVE_TARGETS
text/html
text/_moz_htmlinfo
text/_moz_htmlcontext
image/png
image/bmp
image/x-bmp
image/x-MS-bmp
image/x-icon
image/x-ico
image/x-win-bitmap
image/vnd.microsoft.icon
application/ico
image/ico
image/icon
text/ico
image/jpeg
image/tiff
image/webp
audio/x-riff
xclip -selection clipboard -t text/html -o > /tmp/test.txt gives
<meta http-equiv="content-type" content="text/html; charset=utf-8"><img src="https://www.nejm.org/cms/10.1056/NEJMicm1801693/asset/849a05b6-eeb6-4f12-9b5e-11be7a704005/assets/images/large/nejmicm1801693_f1.jpg" alt="https://www.nejm.org/cms/10.1056/NEJMicm1801693/asset/849a05b6-eeb6-4f12-9b5e-11be7a704005/assets/images/large/nejmicm1801693_f1.jpg" width="848" height="638" class="shrinkToFit fullZoomIn">
whereas xclip -selection clipboard -t image/jpeg -o > /tmp/test.jpeg gives the picture.
6. Building from source an removing the comments in anki/qt/aqt/editor.py (l. 1591 – 1595), I can see the following output when pasting the image from above link:
use clipboard
html=True image=True urls=False txt=False
html <meta http-equiv="content-type" content="text/html; charset=utf-8"><img src="https://www.nejm.org/cms/10.1056/NEJMicm1801693/asset/849a05b6-eeb6-4f12-9b5e-11be7a704005/assets/images/large/nejmicm1801693_f1.jpg" alt="https://www.nejm.org/cms/10.1056/NEJMicm1801693/asset/849a05b6-eeb6-4f12-9b5e-11be7a704005/assets/images/large/nejmicm1801693_f1.jpg" width="1205" height="907" class="shrinkToFit">
urls []
text
2025-02-11 15:53:10,881:DEBUG:urllib3.connectionpool: Starting new HTTPS connection (1): www.nejm.org:443
2025-02-11 15:53:11,020:DEBUG:urllib3.connectionpool: https://www.nejm.org:443 "GET /cms/10.1056/NEJMicm1801693/asset/849a05b6-eeb6-4f12-9b5e-11be7a704005/assets/images/large/nejmicm1801693_f1.jpg HTTP/11" 403 None
[17467:17519:0211/155313.661525:ERROR:nss_util.cc(357)] After loading Root Certs, loaded==false: NSS error code: -8018
meaning it does seem to find both, html and the image.
8. A normal local screenshot gives this:
use clipboard
html=False image=True urls=False txt=False
html
urls []
text
2025-02-11 15:59:27,383:DEBUG:aqt.mediasrv: GET /paste-e7f953805215a3b51b11a0ff206a748c3c6b2e73.jpg
Attempted solution
So I changed the code slightly (qt/aqt/editor.py, line 1600+).
# try various content types in turn
if mime.hasHtml() and not mime.hasImage():
html_content = mime.html()[11:] if internal else mime.html()
return html_content, internal
# given _processUrls' extra allowed_suffixes kwarg, placate the typechecker
def process_url(mime: QMimeData, extended: bool = False) -> str | None:
return self._processUrls(mime, extended)
# the clipboard stores the image and html containing a link to the remote
# server. There's no need to fetch the image though if the clipboard
# already contains it.
if mime.hasImage() and mime.hasHtml:
types = (self._processImage, self._processText)
With that change the image from the clipboard (if available) will be used. There is no network connection and anki stores the picture in the media folder, using something like this:
<img src="paste-dec03436e8a74f25918b1854a73ef8dde6f7eb8b.jpg">
Limitations
File names and alt tags are lost.
I have no idea how not remove the tags and file names though. Well, I mean I suppose one could do something like that:
html_content = mime.html()
start = html_content.find("<img")
html_content = mime.html()[start:]
which would give
<img src="https://www.nejm.org/cms/10.1056/NEJMicm1801693/asset/849a05b6-eeb6-4f12-9b5e-11be7a704005/assets/images/large/nejmicm1801693_f1.jpg" alt="https://www.nejm.org/cms/10.1056/NEJMicm1801693/asset/849a05b6-eeb6-4f12-9b5e-11be7a704005/assets/images/large/nejmicm1801693_f1.jpg" width="1205" height="907" class="shrinkToFit">
And I guess we could use the same logic to get the alt fields and image name out of the html.
But then we’d need to modify _read_pasted_image() and _addMedia() too. I do not think I’m skilled enough to make these kind of changes. But maybe this helps someone.
Yeah I’m a bit self-conscious over that. What’s Anki’s policy on storage space? I’m not trying to abuse and my major deck just has really rich audio and image data.
