CLI: Add support to change the encoding (i.e. UTF-8)

[PaulKraut]

Hi,

Creating a qr-code with "German Umlauts":

$ segno -o=fool.png --encoding="UTF-8" --title="Example QR code" "The Fool on [123-äöüÖÄÜ-456] the Hill"

creates the fool.png file. Showing that png on my desktop and taking a picture smartphone (Galaxy Note 10), the smartphone finds the qr-code immediately, but it the German umlauts have been wrongly translated to some special characters (see Photo).

I use my Smartphone often to read qr-codes with Umlauts. I did not observe utf-8 decoding issues yet. Do you have any idea

Interestingly, fool.png can be decoded by zbarimg correctly as:
$ zbarimg fool.png QR-Code:The Fool on [123-äöüÖÄÜ-456] the Hill

Do you have any idea, why the qr-codes with German Umlauts are wrongly decoded by my smartphone, but correctly with zbarimg? Di you see any possibility for a workarounf or finxing?

Thank you & kind regards,

[heuer]

Thanks for your report.

It seems to be a problem with the CLI. The QR code is not encoded with UTF-8 but ISO 8859-1.

The correct output (using UTF-8) would be:

segno.make_qr('The Fool on [123-äöüÖÄÜ-456] the Hill', encoding='utf-8') \
     .save('fool-utf-8.png', scale=3)

ZXing and other decoders are able to decode both QR codes correctly, but I'll check why the command line interface ignores the encoding.

[heuer]

Ah, I got it! --encoding is used by the SVG serializer and is not used to encode the input. :(

I'm thinking of a solution, thanks for the hint!

[PaulKraut]

Hi,

thanks for the immediate response 👍

Maybe it is worth to mention, that the encoding issue also exist when you use the helpers functions for any qr-code with free text:

I met the problem, when I tried to create a vcard qr-code with helpers.make_vcard() containing an Umlaut: I was not allowed to explicitly specify the encoding, yet.

Probably the best way would be, that the app detects itself, whether or not to fall back into UTF-8 encoding. But I have no clue, what this means for coding ..

Kind regards,

[heuer]

I met the problem, when I tried to create a vcard qr-code with helpers.make_vcard() containing an Umlaut: I was not allowed to explicitly specify the encoding, yet.

Well, I think the behaviour (try ISO 8859-1 and if it does not fit, use UTF-8) is correct acc. to ISO/IEC 18004. It seems to be an issue with the decoder app. Which decoder app do you use, btw?

The helpers module provides for each QR code factory function also a _data function which can be used as input for segno.make.

>>> import segno
>>> from segno import helpers
>>> qr = helpers.make_vcard(name='Müller;Änne', displayname='Änne Müller', email='aenne@example.org', phone='+1234567')
>>> qr.save('vcard-iso.png', scale=5)
>>> # Use the _data factory
>>> data = helpers.make_vcard_data(name='Müller;Änne', displayname='Änne Müller', email='aenne@example.org', phone='+1234567')
>>> # Create and save a QR code with an explicit encoding == UTF-8
>>> segno.make(data, encoding='utf-8').save('vcard-utf8.png', scale=5)

Results

ISO 8859-1

Decoded (using ZXing)

UTF-8

Decoded:

[PaulKraut]

Well, I think the behaviour (try ISO 8859-1 and if it does not fit, use UTF-8) is correct acc. to ISO/IEC 18004. It seems to be an issue with the decoder app. Which decoder app do you use, btw?

Ah, now I got it, sorry:

In my first post, was confused why the zbarimg (linux, command line tool) returned the correct string, while the qr-decoder in the smartphone photo-app failed: zbarimg used latin-1 and the Samsung photo-App seems not to be able to handle latin-1.

The helpers module provides for each QR code factory function also a _data function which can be used as input for segno.make.

>>> import segno
>>> from segno import helpers
>>> qr = helpers.make_vcard(name='Müller;Änne', displayname='Änne Müller', email='aenne@example.org', phone='+1234567')
>>> qr.save('vcard-iso.png', scale=5)
>>> # Use the _data factory
>>> data = helpers.make_vcard_data(name='Müller;Änne', displayname='Änne Müller', email='aenne@example.org', phone='+1234567')
>>> # Create and save a QR code with an explicit encoding == UTF-8
>>> segno.make(data, encoding='utf-8').save('vcard-utf8.png', scale=5)

Results

ISO 8859-1

Decoded (using ZXing)

UTF-8

Decoded:

Perfect! Helped a lot, thank you very much!

[heuer]

Ah, now I got it, sorry:

In my first post, was confused why the zbarimg (linux, command line tool) returned the correct string, while the qr-decoder in the smartphone photo-app failed: zbarimg used latin-1 and the Samsung photo-App seems not to be able to handle latin-1.

No problem, thank you for discovering a bug! 👍

Your example

$ segno -o=fool.png --encoding="UTF-8" "The Fool on [123-äöüÖÄÜ-456] the Hill"

should work as expected with release 1.3.0