underchemist / nanonispy Goto Github PK

Thanks a lot for this module.

Are there any plans for a Python 3.6 support?

Hi,

I've got a guy trying to install nanonispy from a git clone/ running install, and it throws the following error:

$python setup.py install
running install
running bdist_egg
running egg_info
creating nanonispy.egg-info
writing nanonispy.egg-info\PKG-INFO
writing dependency_links to nanonispy.egg-info\dependency_links.txt
writing requirements to nanonispy.egg-info\requires.txt
writing top-level names to nanonispy.egg-info\top_level.txt
writing manifest file 'nanonispy.egg-info\SOURCES.txt'
error: package directory 'nanonispy\tests' does not exist

I've been scratching my head looking at the setup.py file but can't offer you any further insight into why its hanging.

Cheers,

Jack

Hi,

I managed to take some data with duplicitous header entries (lol nanonis).

So the nap.read.Spec chokes on line 405, because the data footprint of the header is longer than the length of the parsed header, e.g., the return of _parse_dat_header (which doesn't duplicate the entries apparently).

If I replace 405 with:

        f = open(self.fname, 'r')
        ff = f.readlines()
        header_lines = ff.index('[DATA]\n') + 2
        f.close()

It appears to get the right value for header_lines.

Commit fc91606 is not included in the latest PyPI release, so using Spec is rather annoying because this line is missing.

Could you please push a new release to the PyPI? Thanks in advance!

For example

from pathlib import Path
import nanonispy as nap

path_to_datafile = Path('/path/to/datafile.3ds')
grid = nap.read.Grid(path_to_datafile)

will fail with a TypeError

TypeError: 'PosixPath' object is not subscriptable

The reason is, well the a PosixPath object is not subscriptable like this

if fname[-3:] != ext:

I will open a pull request to fix that.

//anaconda/lib/python3.5/site-packages/nanonispy/read.py in start_byte(self)
112 for line in f:
113 # Convert from bytes to str
--> 114 print(line)
115 entry = line.strip().decode()
116 if tag in entry:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0xec in position 10: invalid continuation byte

Hi,

I encounter this ~semi-regularly, but can't reliably reproduce it unfortunately. Sometimes adding an 'ignore' in the decode() works, sometimes not.

np.int, np.float, etc have been deprecated in newer versions of numpy. There is an open pull request to rectify this

from david.wander at web dot de

Dear Yann-Sebastien,

first of all, I would like to thank you for sharing your nanonispy library.
It makes the data treatment of my measurements much easier!

However, I just found a major bug in the _load_data function of the Grid class.
Line 230 of read.py (nanonispy Version 1.0.7) reads:

griddata_shaped = griddata.reshape((nx, ny, exp_size_per_pix))

Here you are confusing x and y. The correct code would be:

griddata_shaped = griddata.reshape((ny, nx, exp_size_per_pix))

Of course this then implies that after loading the data, you first have to address y and then x, so for example like this:

grid.signals['params'][index_y][index_x][param_index]

I admit that it is easy to miss this error because it seems to give the right array shapes but if you take a close look, you will realize that when going from
signals['params'][0][0] to signals['params'][1][0] (i.e. increasing the x index by 1), it is actually the y coordinate that changes (signals['params'][1][0][3])
For square shaped grids, this might not be a desaster, as it only leads to a grid rotated by 90° but as soon as you work with other shapes, things get totally messy.

I hope you will correct this in the next version of the package!

Best regards,

David Wander

Currently, the header information for comment of .sxm-files do not support entries with more than one line; only the first line is taken and the rest is cut off.
This is because the header-entries are split with \n as the delimiter.

EDIT: Saw PR #13 that addresses the multipass entries.