get_channel("Height") issue

[nihtmusic]

One more issue I am running into that not being able to view height channel. I am able to get .list_channels() to work, but .get_channel("Height") is not working. Below is a screen shot of my jupyter notebook showing the error:

This file version is v5.15b7 if that helps.

Thank you,

Kurt

[nihtmusic]

One more thing...if I just do a vanilla get_channel() I get the following errors:

[dineshpinto]

If you send me the Bruker file (annual.fallout_0z@icloud.com), I can take a look

[scholi]

You should do a afm_image.list_channels() then you will know what is available

[nihtmusic]

If you send me the Bruker file (annual.fallout_0z@icloud.com), I can take a look

Sorry for the delay. I will send the file to you shortly. Thanks for taking a look. -Kurt

[nihtmusic]

You should do a afm_image.list_channels() then you will know what is available

I did that as shown in the first post in this thread. That works...and "Height" is there. But .get_channel("Height") doesn't work:

[dineshpinto]

So the file defines "Line direction" not "Line Direction", which is a trivial bugfix.

However, it looks like the byte packing of the file differs from a standard Bruker file. From the file you sent:

>>> afm = pySPM.Bruker("junk.002")
>>> afm.list_channels()
Channels
========
S [Height] "Height"
>>> byte_length = int(afm.layers[0][b"Data length"][0])
32768
>>> cols = int(afm.layers[0][b"Number of lines"][0])
512
>>> rows = int(afm.layers[0][b"Samps/line"][0])
512

This is causing a problem as Bruker._get_raw_layer() calculates bytes per pixel as bpp = byte_length // (rows * cols), which here is 32768//(512*512) = 0, which is obviously incorrect.

Querying bpp directly:

>>> int(self.layers[i][b"Bytes/pixel"][0])
2

Could the file be corrupted or in a different format than expected, @scholi do you have any idea?

[scholi]

The bpp info is confusing and this is confirmed by the source code of Gwyddion: https://github.com/christian-sahlmann/gwyddion/blob/master/modules/file/nanoscope.c

The importand part is:

/* Bytes/pixel determines the scaling factor, not actual raw data type.
     * This is what Bruker people say. */
    val = g_hash_table_lookup(hash, "Bytes/pixel");
    qbpp = val ? GWY_ROUND(val->hard_value) : 2;

    bpp = 2;
    if (file_type == NANOSCOPE_FILE_TYPE_BIN32)
        bpp = 4;

and

typedef enum {
    NANOSCOPE_FILE_TYPE_NONE = 0,
    NANOSCOPE_FILE_TYPE_BIN,
    NANOSCOPE_FILE_TYPE_BIN32,
    NANOSCOPE_FILE_TYPE_TXT,
    NANOSCOPE_FILE_TYPE_FORCE_BIN,
    NANOSCOPE_FILE_TYPE_FORCE_BIN32,
    NANOSCOPE_FILE_TYPE_FORCE_VOLUME,
    NANOSCOPE_FILE_TYPE_FORCE_VOLUME32,
    NANOSCOPE_FILE_TYPE_BROKEN
} NanoscopeFileType;

Now something is definitely strange with the data length and the col/rows. If you open it with gwyddion. What size do you get? Maybe you can send me the file. I'll be on holidays, but I can have a look at it in few weeks

[dineshpinto]

@scholi I've forwarded you the file

[scholi]

I have prepared a fix. Could you try it. I have tried it on 4 different files. Hopeful it will be more stable, but Bruker is the worst format ever inverted. The header contains a lot of incorrect values....

Note: it seems that your file has 32 rows and pySPM read the 32 rows while gwyddion only reads 16. Do you know how many lines you recorded?

[scholi]

@dineshpinto I have added the unittest and the header of some Bruker data for the unittest. the data contains only the header to save space and I have added a mock_data argument in Bruker.get_channel() to skip the missing data in the file

[dineshpinto]

@scholi the fixes look good, all tests are passing. Did some minor restructuring and formatting, and will publish a new release at v0.6.1

cc @nihtmusic

[nihtmusic]

Awesome! I will hit it a try…much appreciated!

Kurt

On Fri, Feb 9, 2024 at 10:46 AM Dinesh Pinto @.***> wrote:

@scholi https://github.com/scholi the fixes look good, all tests are passing. Did some minor restructuring and formatting, and will publish a new release at v0.6.1

cc @nihtmusic https://github.com/nihtmusic

— Reply to this email directly, view it on GitHub https://github.com/scholi/pySPM/issues/41#issuecomment-1936254356, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANZLN3RB2W4OETQMGDX3MI3YSZHGHAVCNFSM6AAAAABCCI7QSCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZWGI2TIMZVGY . You are receiving this because you were mentioned.Message ID: @.***>