Hello,
I have just install PmagPy. I would like to launch QuickMagIC.py from the PmagPy Prompt but I get this message:

C:\Users\user\Documents>QuickMagIC.py
Traceback (most recent call last):
File "C:\Users\user\Documents\PmagPy\QuickMagIC.py", line 14, in
import pmag_basic_dialogs
File "C:\Users\user\Documents\PmagPy\pmag_basic_dialogs.py", line 13, in
import pmag_widgets as pw
File "C:\Users\user\Documents\PmagPy\pmag_widgets.py", line 10, in
import controlled_vocabularies as vocabulary
File "C:\Users\user\Documents\PmagPy\controlled_vocabularies.py", line 129, in

all_codes, code_types = get_meth_codes()
File "C:\Users\user\Documents\PmagPy\controlled_vocabularies.py", line 21, in
get_meth_codes
raw_codes = pd.io.json.read_json('http://api.earthref.org/MAGIC/method_codes
.json')
AttributeError: 'module' object has no attribute 'json'

I do not know how to fix this attribute error. Is anyone could help me?

Lori,

Are you planning to do more with the pmag2.py file? It seems a bit confusing to have a file with that name given that so much functionality of the overall package resides in the main pmag.py file. Although I am sure there is a good reason you did it this way.

Both @CourtneySprain and I are having this issue with the demag_gui:

(python2) fixed-mcc-393-1:~ Laurentia$ demag_gui.py 
/Users/Laurentia/anaconda/envs/python2/lib/python2.7/site-packages/matplotlib/font_manager.py:273: UserWarning: Matplotlib is building the font cache using fc-list. This may take a moment.
  warnings.warn('Matplotlib is building the font cache using fc-list. This may take a moment.')
2016-09-14 11:35:54.176 python[5412:1033584] This application is trying to draw a very large combo box, 29 points tall.  Vertically resizable combo boxes are not supported, but it happens that 10.4 and previous drew something that looked kind of sort of okay.  The art in 10.5 does not break up in a way that supports that drawing.  This application should be revised to stop using large combo boxes.  This warning will appear once per app launch.

The top menu boxes are odd with weird white space around the pull down menus and associated values:

This will be helpful for evaluation of data points determining if they are "bad" or "good" especially when repeat measurements are made.

Would be nice to have an option to view data as VGPs that comes up as a window when selected in the Tools menu of demag_gui

The GUI showing the current directory flashes up, but then disappears and the GUI proceeds as if you have selected your current directory.

Laurentia$ demag_gui.py
/Users/Laurentia/anaconda/envs/python2/lib/python2.7/site-packages/matplotlib/font_manager.py:273: UserWarning: Matplotlib is building the font cache using fc-list. This may take a moment.
  warnings.warn('Matplotlib is building the font cache using fc-list. This may take a moment.')
icon_path /Users/Laurentia/PmagPy/images/PmagPy.ico
-I- done sorting meas data
------------------------------ no magic_measurements.txt found---------------------------------------

Now that the backend issue is resolved by @moonshoes87 (I think?), should ipmagplotlib be deleted from the repository so that we are solely using pmagplotlib?

This error was thrown when switching between coordinate systems and is being thrown in other use cases as well.

Traceback (most recent call last):
  File "/Users/Laurentia/PmagPy/demag_gui.py", line 1548, in onSelect_coordinates
    fit.put(specimen,self.COORDINATE_SYSTEM,self.get_PCA_parameters(specimen,fit.tmin,fit.tmax,self.COORDINATE_SYSTEM,fit.PCA_type))
  File "/Users/Laurentia/PmagPy/demag_gui.py", line 1942, in get_PCA_parameters
    beg_pca,end_pca = self.get_temp_indices(None, tmin, tmax, specimen)
  File "/Users/Laurentia/PmagPy/demag_gui.py", line 4730, in get_temp_indices
    if tmin in self.Data[specimen]['zijdblock_steps']:
KeyError: 'PI47-3'

The new Demag GUI tests seem to require interaction to run. When I run them, I get stuck with screens like this:

and then I have to click through for the testing to continue. This won't work for automated testing (Travis CI), and it isn't ideal for running through all PmagPy tests in a batch.

Is this the behavior that other people get, as well? Or is it a local problem?

We haven't really established a system for determining when to make a new pip release. So, this issue has two parts: 1. does anyone have a reason I shouldn't go ahead and make a new release now? 2. How should we time/coordinate new releases in general?

Regarding item two, it seems obvious that we should make sure all tests are passing, but maybe we should have other safeguards as well?

Recent changes to demag_gui are resulting in this fatal error:

Traceback (most recent call last):
  File "/Users/Laurentia/PmagPy/programs/pmag_gui.py", line 332, in on_run_demag_gui
    import demag_gui
  File "/Users/Laurentia/PmagPy/programs/demag_gui.py", line 73, in <module>
    from dialogs.demag_interpretation_editor import InterpretationEditorFrame
ImportError: No module named demag_interpretation_editor

The good/bad measurement highlighting behavior still seems a little unintuitive.

First of all, it doesn’t seem like I can simply right click on a measurement — I actually have to select it with a left click and then right click. Additionally, I was confused when I accidentally double-clicked on a measurement (trying to mark it as good/bad) and the whole fit was recalculated. Last thing, measurements outside of the fit cannot be marked — is this the desired behavior?

A possible UI solution: maybe we could introduce a 6th column in the measurements list with a checkbox the denotes a ‘bad’ measurement (or ‘good’). This might be a slightly irritating to implement with wxPython, but I think it would make a lot of sense from a user perspective.

When running aniso_magic.py with a bootstrap I was getting weird error ellipses that were rotated 90º from the data distribution they were meant to be representing.

Datafile:
WT1_anisotropy.txt
Command:
aniso_magic.py -f WT1_anisotropy.txt -par -crd g
Result:


I eventually worked out that somehow the zeta angle was being given the eta angles D,I parameters and vice-versa, which I traced back to pmag.dokent():

In the first two chunks of highlighted code below there appears to be a mismatch with the third chunk: sigma1at [i][1] in an array is associated with eta and sigma2 at [i][0] with zeta, yet at 3221-3222 the zeta directions are calculated from [i][1] and eta from [i][0]. If you instead have:
3305 sigma1=sigma1+xg[i][0]**2
3306 sigma2=sigma2+xg[i][1]**2
Then you get the right result.

Also: in aniso_magic.py lines 322-325, which print the bootstrap output, the zeta parameters are printed in the eta columns and vice versa.

Make it so that this warning is given when needed (it already is, but needs to provide more information):
"There are multiple 'good' XXX steps. The first one will be used for fit XX."

Make it so that this warning is given when needed:
"Within Fit XX, there are multiple good measurements at the XXX step. Both measurements are included in the fit."

Here is an example of the equal area plot. There is only one current fit, but a bunch of fits that I was making for the specimen while playing around with different possibilities (and just making a bunch of bad fits to make this point) are being shown:

Running the test data
find_EI.py -f find_EI_example.dat
returns

Io Inc  I_lower, I_upper, Elon, E_lower, E_upper 
38.9  =>     58.8 _     0.0 ^    67.0:  1.4679 _ 0.0000 ^ 1.7189

when it should return something like this

Io Inc  I_lower, I_upper, Elon, E_lower, E_upper
38.9  =>     58.8 _    48.4 ^    67.0:  1.4679 _ 1.2912 ^ 1.7337

For some reason, both E_lower and I_lower are zero. I get the expected results with pmag-3.3.2, but not the current version.

I'm getting some errors when I run the demag gui tests.
These three tests produce the same error: test_read_write_redo, test_read_write_pmag_tables, test_ie_buttons.
AttributeError: 'Demag_GUI' object has no attribute 'get_ie_open'

I had a merge conflict with the test_demag_gui.py, but I thought I fixed that... Could one of you have a look at this, @Swanson-Hysell or @Caoimhinmg?

The help PDF for Demag GUI appears to be missing. Either the "Help --> Usage & Tips" option should be taken out of the menu, or we should put demag_gui_doc.pdf back in.

I noticed this because I'm trying to make sure all image files etc. are available in the pmagpy-cli edition of the GUIs.

I am still running into errors when trying to import pmag.py due to code in check_updates.py. Here is the current error:

/Users/polarwander/PmagPy/check_updates.py in main()
     15     global pmagpy_path,local_path
     16     local_version=get_version()
---> 17     last_path   = os.path.join(pmagpy_path, 'version_last_checked.txt')
     18     # Get the version of the local PmagPy installation
     19     # from the version.txt file in the current directory.

NameError: global name 'pmagpy_path' is not defined

I think the issue is that while I have pmagpy in the path it isn't a global variable in an IPython session. And I think one of the real advantages to the notebook is that it isn't a requirement to have anything added to the path as long as one can import: pmag.py, pmagplotlib.py and IPmag.py

This problem goes away if all the code below def main(): in check_updates.py is pulled out and figures still send up being tagged with pmagpy-2.236. One solution could be to simplify the versioning system by putting this code into pmag.py instead of needing to import check_updates.py but perhaps this would have unintended consequences:

version = 'pmagpy-2.236'

def get_version():
        return version

Box tops are in different vertical positions and need to be standardized:

As we start contemplating the switch to Python 3, maybe we can revisit the two issues with Anaconda. It is unclear when Canopy will support Python 3, so I'd like to have other options.

To recap:

Problem 1:
I’m also getting some segmentation faults with the GUIs. This seems to happen consistently with Pmag GUI step 3 (adding er_* information).

/anaconda/bin/pythonw: line 3: 2366 Segmentation fault: 11 //anaconda/python.app/Contents/MacOS/python "$@"

Problem 2:
nebula$ python pmag_gui.py
-I- Starting Pmag GUI - please be patient
This program needs access to the screen.
Please run with a Framework build of python, and only when you are
logged in on the main display of your Mac.

@Swanson-Hysell or @Caoimhinmg, if one of you could check and see if problem 1 is still occurring (I no longer have Anaconda installed), that would be awesome! Regarding problem 2, I can't remember if the framework build issue affected the pip install. With Anaconda, have you been able to do the pip install and run the command line programs?

pmag.py:4779 for k in range(1,66):

number of terms in IGRF varies.
Instead of using a fixed value of 66, should calculate from (n+1)(n+2)/2 or n(n+3)/2 + 1

One of the Demag GUI tests is failing:

When I run:

python -m unittest pmagpy_tests.test_demag_gui.TestMainFrame

I see this error:

ERROR: test_read_write_redo (pmagpy_tests.test_demag_gui.TestMainFrame)

Traceback (most recent call last):
File "pmagpy_tests/test_demag_gui.py", line 233, in test_read_write_redo
self.assertFalse(self.frame.interpretation_editor_open)
AttributeError: 'Demag_GUI' object has no attribute 'interpretation_editor_open'

It seems that there is now a requirement within the PmagPy/programs/init.py file for the 'pmagpy-cli' distribution to be installed. This error gets thrown when trying to convert files via cit_magic.py through pmag_gui. We would like to be able to run straight from the master without doing a pip installation (given that the pip installed version is not working with Anaconda) and simply having the path set.

Traceback (most recent call last):
  File "/Users/Laurentia/PmagPy/dialogs/pmag_basic_dialogs.py", line 1018, in on_okButton
    import programs.cit_magic as cit_magic
  File "/Users/Laurentia/PmagPy/programs/__init__.py", line 21, in <module>
    print pkg_resources.get_distribution('pmagpy-cli')
  File "/Users/Laurentia/anaconda/envs/python2/lib/python2.7/site-packages/setuptools-20.1.1-py2.7.egg/pkg_resources/__init__.py", line 540, in get_distribution

  File "/Users/Laurentia/anaconda/envs/python2/lib/python2.7/site-packages/setuptools-20.1.1-py2.7.egg/pkg_resources/__init__.py", line 420, in get_provider

  File "/Users/Laurentia/anaconda/envs/python2/lib/python2.7/site-packages/setuptools-20.1.1-py2.7.egg/pkg_resources/__init__.py", line 941, in require

  File "/Users/Laurentia/anaconda/envs/python2/lib/python2.7/site-packages/setuptools-20.1.1-py2.7.egg/pkg_resources/__init__.py", line 828, in resolve

pkg_resources.DistributionNotFound: The 'pmagpy-cli' distribution was not found and is required by the application

Given that all of our documentation has people pip install, it is confusing to have the install scripts in the repo. As evidenced in this issue: #80

install_Mac_OSX
install_Windows.vbs

Should we go ahead and delete these from the repo @ltauxe @moonshoes87?

To demag_gui edit menu add choices to mark measurements "good" or "bad"

To do this, split the on right click function into three.

• remove test autointerpret
• remove Convert and Combine MagFiles (test .inp data structure)
• fix or remove Import Interpretations from LSQ File

Make upper hemisphere fits open on specimen plot and colored by fit as they are on the higher level means plot.

Seems to be an issue with the recent changes related to font_type. See Traceback below:

Traceback (most recent call last):
  File "/Users/Laurentia/PmagPy/demag_gui.py", line 3969, in on_menu_edit_interpretations
    self.interpretation_editor = EditFitFrame(self)
  File "/Users/Laurentia/PmagPy/demag_gui.py", line 5039, in __init__
    self.init_UI()
  File "/Users/Laurentia/PmagPy/demag_gui.py", line 5049, in init_UI
    font1 = wx.Font(10, wx.SWISS, wx.NORMAL, wx.NORMAL, False, self.font_type)
AttributeError: 'EditFitFrame' object has no attribute 'font_type'

The good/bad measurement highlighting behavior still seems a little unintuitive.

First of all, it doesn’t seem like I can simply right click on a measurement — I actually have to select it with a left click and then right click. Additionally, I was confused when I accidentally double-clicked on a measurement (trying to mark it as good/bad) and the whole fit was recalculated. Last thing, measurements outside of the fit cannot be marked — is this the desired behavior?

A possible UI solution: maybe we could introduce a 6th column in the measurements list with a checkbox the denotes a ‘bad’ measurement (or ‘good’). This might be a slightly irritating to implement with wxPython, but I think it would make a lot of sense from a user perspective.

Thoughts on this? @Swanson-Hysell, @Caoimhinmg , @ronshaar

One thing that I think would be good to strategize about would be how the directory structure of the repository could be updated to make things more approachable. It would be nice if when people landed on the Github page that the README wasn't below such a long list of files. One solution could be to put the command line programs into a subfolder and the function modules into a subfolder just leaving the GUI programs in the main folder. Have you you thought about some ways to do this @moonshoes87 ?

Here is a good piece with a simple discussion of repository architecture: http://www.kennethreitz.org/essays/repository-structure-and-python

Project is misspelled as "Prpject" in QuickMagIC dialogue box when unpacking downloaded txt file.

When demag gui is first loaded, and there is no previous interpretation for the given specimen the specimen equal area plot appears with tick marks

Debug: Unknown accel modifier: 'x'

get rid of pdb import too.

I am just starting to use the MagicDataFrame object in some code and am bumping into the problem that the values that should be floats and integers within the DataFrame are all in there as strings.

I can change them myself by doing something like:

Halls1974_sites['dir_dec'] = pd.to_numeric(Halls1974_sites['dir_dec'])

but given that most of the columns are always floats or always integers it seems like having the correct dtype should be a feature of the MagicDataFrame object.

Thoughts, @moonshoes87?

It would be a great addition to the Pmag_GUI program if whenever a measurement was selected on the measurement list the color of it would change in the Zijderveld/Orthographic diagram and in the Equal Area/Stereonet diagram. This would allow us to know which measurement we are looking at quickly even if the labels are written on the other portion of the vector (DEC v. INC).

Attached is a GIF showing how this works in the Craig Jones PaleoMag software

There is currently hidden functionality to add a fit to all specimens in the interpretation editor if now specimen is selected. It would be nice to make this functionality be more accessible and transparent to the user by having a button that says "make new fit for all specimens" in the demag_gui interpretation editor.

Summary:
PmagPy throws an error when I try to run the post-install test-script on Mac OS X.

How to Recreate:

1. On Mac OS X: v10.9.5 (Maverick), download .Zip file of the PmagPy GitHub repository:
   https://github.com/PmagPy/PmagPy/

2. Extract all files and then open and run the file: ../PmagPy-master/install_Mac_OSX.app

3. Choose to have the files installed onto the Mac HD. (The option to install under the the current user was grayed out. The installer said this option was not allowed).

4. Run the installer. Installs all PmagPy files under: /Users/{my username}/PmagPy/

5. At finish, the installer runs the test script:
   ":~ {my username}$ eqarea.py -h" in Terminal

This command fails with the following error:
Traceback (most recent call last):
File "/Users/lab/PmagPy/programs/eqarea.py", line 8, in
import pmagpy.pmagplotlib as pmagplotlib
ImportError: No module named pmagpy.pmagplotlib

I checked the install directory and found the relevant file:
/Users/{my username}/PmagPy/pmagpy/pmagplotlib.py

6. I tried rerunning the command prompt as super user:
   ":~ {my username}$ sudo eqarea.py -h"

and got the same error.

Thank you,
Isaac

On some data sets, demag_gui is referencing NRM before NRM is assigned:

Traceback (most recent call last):
  File "/Users/Laurentia/PmagPy/programs/pmag_gui.py", line 334, in on_run_demag_gui
    demag_gui.main(self.WD, standalone_app=False, parent=self)
  File "/Users/Laurentia/PmagPy/programs/demag_gui.py", line 6513, in main
    frame = Demag_GUI(WD, parent, write_to_log_file=write_to_log_file)
  File "/Users/Laurentia/PmagPy/programs/demag_gui.py", line 194, in __init__
    self.Data,self.Data_hierarchy=self.get_data() # Get data from magic_measurements and rmag_anistropy if exist.
  File "/Users/Laurentia/PmagPy/programs/demag_gui.py", line 3117, in get_data
    DIR=[dec,inc,intensity/NRM]
UnboundLocalError: local variable 'NRM' referenced before assignment

@moonshoes87's sleuthing suggests this wasn't an issue at the time point of this commit:

This is the oldest commit that works for me (from September 26):

6430242, added documentation and cleaned up code….

The "builtins" module isn't part of the Canopy package, although future is installed. (not sure if this is relevant.... builtins issue)

Because of this, cit_magic3 fails with an import error. Also, pdb is referenced in the code but never imported. I fixed these issues locally so that I can run cit_magic3, but I didn't want to muck around with it without understanding why it was set up that way!

It would be easier to see the highlighted points if they became a lighter color rather than a darker color when selected in the z-plot, equal area and M/Mo plots.

It would be nice if we could have an option in the bootstrapped common mean (and reversal and fold test) to use the parametric bootstrap instead of naive bootstrap in case we have data sets with N<20.

I see that there is an option in bootams.py to call "-par"

Ideally this would be incorporated within the pmagpy functions and then into the ipmag functions for jupyter notebooks "common_mean_bootstrap", "bootstrap_fold_test" and "reversal_test_bootstrap"

It makes sense to add an open source license to the repository.

I was reading this piece: http://www.astrobetter.com/blog/2014/03/10/the-whys-and-hows-of-licensing-scientific-code/ which laid out this advice:

(1) Always license your code. Unlicensed code is closed code, so any open license is better than none (but see #2).
(2) Always use a GPL-compatible license. GPL-compatible licenses ensure broad compatibility for your code, and include GPL, new BSD, MIT, and others (but see #3).
(3) Always use a permissive, BSD-style license. A permissive license such as new BSD or MIT is preferable to a copyleft license such as GPL or LGPL.

Github put a helpful site together in terms of describing available licenses:
http://choosealicense.com

This one is a simple one that could work:
http://choosealicense.com/licenses/bsd-2-clause/

@Caoimhinmg @Swanson-Hysell

this is now working pretty much - but there is problem calculating best fit lines that doesn't occur in zeq_magic.py. I get this error:

with this pmag_specimens.txt file and
pmag_specimens.txt

and this magic_measruements table.
magic_measurements.txt

When I open Demag GUI, I can initially only select bounds by clicking on the steps in the lower left corner. After I've clicked on a step, the drop-down Bounds in the upper left appear and I can edit using those. It seems like the Bounds drop-down should work from the start, or we should remove that way of interacting with the parameters.

I can dig into this, but it will probably be easier for one of you who have been working on Demag GUI more...

site_edit functionality can be added. What that program does is plot expected directions from two common errors - wrong mark on the sample and wrong way for the arrow. By doing so, samples can be excluded if there directions are consistent with such an error.

Going through the manuscript, I realized that, at present, Make_magic.py and QuickMagIC.py have similar names but with different syntax in terms of capitalization and the underscore. It seems like these should be standardized.

Also, is QuickMagIC still the right name for that program? As an executable, this program now accomplishes a lot in terms of data import, data wrangling, PCA fits through demag_gui, and paleointensity data reduction through thellier_gui. I'm not sure what a better name would be. Perhaps if MagIC.py is going to be retired, QuickMagIC could be renamed MagIC.py?

Some tests were set up to use sys.prefix as the working directory because that is where the pip installation puts all the data files. That was so that someone with the pip installation who has not downloaded from Github could run tests and make sure everything was working correctly on their machine, and also so that I could test the pip installation directly. Using cwd + 'data_files' is going to break that functionality. Here's an explanation about how that works with pip: stack overflow

However, using sys.prefix doesn't work correctly if you haven't installed using pip (or copied the data_files folder to sys.prefix + pmagpy_data_files). Perhaps we can discuss how and where we want the tests to be able to run? I like having the ability to directly test the pip installation, so maybe we could figure out how to get it to work both ways?

Also, quite a few tests are now failing for me. 'pmagpy_data_files' was not changed to 'data_files' everywhere, so a bunch of tests can't find test data. Let's decide how we want to do things and get this fixed.

@Caoimhinmg, @Swanson-Hysell ?