Include a gif which shows how to use the tool in README file.

Index of last segment will change when using the split functionality. However, the last segment index is not updated in divide_segment method.

alguerre@alguerre-VirtualBox:~/Desktop/TrackEditor$ python3 src/track_editor.py
Traceback (most recent call last):
File "src/track_editor.py", line 112, in
MainApplication(root).pack(side='top', fill='both', expand=True)
File "src/track_editor.py", line 57, in init
toolbar.children['!button6'].pack_forget() # configure subplots
KeyError: '!button6'

The script issue_14_16.sh has a workaround for Linux launch.

• Unselected
  

• Selected
  

• Used file trail-20200821T164304.zip

When the option "save session" is used the next warning is displayed, besides this warning the created .h5 file can be properly loaded:

C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\tkinter_init_.py:1420: PerformanceWarning:
your performance may suffer as PyTables will pickle object types that it cannot
map directly to c-types [inferred_type->integer,key->block1_values] [items->Index(['segment'], dtype='object')]

self.tk.mainloop(n)

The click_distance constant should be a variable whose value is assign according to the map zoom and the side distance of the tile, which varies with latitude.

An exception is raised when trying to save a gpx file:

ValueError: NaTType does not support utcoffset

It is solved by including a timestamp in the file.

Note: gpxpy packages does not solve the problem

Use pyinstaller to create track_editor executable files for windows and linux

This warning raises when using the fix_elevation functionality

A TypeError is rised. This may be affected by the 'time' column, which has a time type.

Traceback (most recent call last):
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\tkinter_init_.py", line 1883, in call
return self.func(*args)
File "C:\Users\Alonso\Desktop\proyectos\TrackEditor\src\edit_menu.py", line 54, in reverse_segment
self.controller.shared_data.obj_track.reverse_segment(segment_idx)
File "C:\Users\Alonso\Desktop\proyectos\TrackEditor\src\track.py", line 64, in reverse_segment
rev_segment = pd.DataFrame(segment.values[::-1].astype('float32'),
TypeError: float() argument must be a string or a number, not 'NaTType'

Can all the methods in edit_menu and file_menu be called by using a decorator?
This would be great to generate a warning/error window when an exception is caught.

Digital Elevation Models (DEM) are huge files which contains a full characterization of Earth surface elevation. For Europe all these data can be downloaded from https://land.copernicus.eu/imagery-in-situ/eu-dem/eu-dem-v1.1

It is possible to create a customized docker container to run the application. This way we can be sure to run it using, exactly, the same environment every time. No python package problem would be generated.

Possible Steps:

1. Create an image. This should contain all the instalation process.
2. Configure x-server to process the GUI.
3. Create a folder which can be common for host and docker, as input/output directory.
4. Run application on a docker which is auto-deleted after its use.

When trying a second split the result is not shown.
Test description:

1. Load Innacessible_Island_part1.gpx -> segment 1
2. Load Innacessible_Island_part2.gpx -> segment 2
3. Load Innacessible_Island_part3.gpx -> segment 3
4. Split segment 2 -> into 2a and 2b
5. Split segment 3 -> nothing happens

A new warning is raised when updating matplotlib:
UserWarning: FixedFormatter should only be used together with FixedLocator

C:\Users\Alonso\Desktop\proyectos\TrackEditor\src\plots.py:469: UserWarning: FixedFormatter should only be used together with FixedLocator
ax.set_xticklabels(dist_label)
C:\Users\Alonso\Desktop\proyectos\TrackEditor\src\plots.py:470: UserWarning: FixedFormatter should only be used together with FixedLocator
ax.set_yticklabels(ele_label)

The figure which contains the initial world map has not the same size than after loading the first track.

When using "Fix elevation" in the track Innacessible_Island_Full.gpx the following exception is reached:

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

test_gpx.py

• There are some skipped test cases

test_db_handler.py

• remove mark.dependency, it does not work

test_iosm.py

• download_tiles_by_num is not tested but used in the code
• test_download_tiles_nominal should be renamed
• test_deg2num and test_num2deg are set to True and do not contribute to coverage metric, should we do anything?

Coverage Analysis
Name Stmts Miss Cover
src/db_handler.py 97 16 84%
src/gpx.py 49 3 94%
src/iosm.py 75 16 79%

A loading message can be useful in time consuming operation such as:

• fix elevation
• load gpx
• load session

This may can be shown after waiting more than 0.5seconds and using a decorator.

When a "New session" is created two weird behaviours have been observed:

1. Color of the plot is modified
2. Last shown elevation remains

Problematic sequence:

1. Upload segment
2. Remove segment
3. Upload segment -> list indices must be integers or slices, not numpy.floag64

In segments with few points they must be forced to use the moving_average algorithm. Otherwise, we will see random exceptions from any possible part of the functionality.

Create a dockerfile to run all unit tests automatically.

A good color management policy must be defined.

• What color must be assigned when re-ordering segments?
• New colors after split
• Colors after segment removal
• ...

It would be great if a line join different segments. Therefore, it would be more clear the order of the segments.

Two errors emerge when the first segment is removed:

• 'Last segment has been removed.' warning box is raised
• ValueError: Axis limits cannot be NaN or Inf
  

Utils module has no unit test. To be implemented.

The following sequence is followed:

1. Load track
2. Split
3. Split
4. Remove medium segment
5. Split last segment
6. The split operation generates three segments!

Using a gps device with altimeter is a good decission if the gpx file will be uploaded to strava. Since:

"For devices without recognized barometric altimeters, Strava calculates your activity's elevation gain by comparing the GPS data to a known basemap of absolute elevation measurements."

It has been observed that, actually, when uploading an activity with fixed elevation the final result is a kind of mean value between strava data base and your gpx file.

Suggested model to use: Garmin Edge 830

General

• Include file header

track.py

• Organize public and private classes in track.py
• Include a small comment to describe methods, when not fully clear: fix_elevation, change_order
• Review methods which may be short but bulky: reverse_segment

plots.py

• can be this somehow refactor?

db_handler.py

• remove TODO and create issue if needed

constants.py

• remove TODO and create issue if needed

gpx.py

• remove TODO and create issue if needed

FLAKE8

• Review all the output of flake8 from github actions

split_segment.py produces an exception which is, apparently, needed to have a proper behaviour of split segment functionality. Interactivity is broken without this exception.
It is the print of non-existing variable:

  File "C:\Users\Alonso\Desktop\proyectos\TrackEditor\src\split_segment.py", line 80, in connect
    print(f'cidpress: {cidpress}')
NameError: name 'cidpress' is not defined

Implications:

• An exception causing a final expected behaviour is smelly.
• In edit_menu.py the split_segment does not uses the exception_handler decorator. So, other exceptions from split_segment.py module can reach the final user without control.
• Flake8 task in CI process excludes this file.

The docker image creation takes too long (5min aprox), can it be reduced by using a lighter distribution?
Proposal:

• Alpine Linux: not debian, so apt get is not used and Dockerfile must be modified

When world map image is missing the following exception is raised:

FileNotFoundError: [Errno 2] No such file or directory: 'tiles/1/0/0.png'

It should be automatically downloaded.

Look for a free compass icon on the web and use it.

Right now, the split segment operation is done in the main plot. The creation of a new window can drive to a more friendly user experience since:

• The 'special' DONE button is not needed in main GUI
• No clicking segment operations will be available

This could also fix #34 bug

The src.[module] notation is used in import when launching from pycharm. Launching configuration must be reviewed, since it is being problematic when launching in a Linux environment.

How can these things be included?

Different log locations are established in track_editor.py, main function.
It is not clear whether log files must be created at "log/" or "../log/", which causes an exception when using a new environment.

When clicking in the elevation plot, instead of map, a ValueError is reported.

Traceback (most recent call last):
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\matplotlib\cbook_init_.py", line 196, in process
func(*args, **kwargs)
File "C:\Users\Alonso\Desktop\proyectos\TrackEditor\plots.py", line 276, in on_click
get_closest_segment(ob_track.track, (event.ydata, event.xdata))
File "C:\Users\Alonso\Desktop\proyectos\TrackEditor\plots.py", line 220, in get_closest_segment
df_track['point_distance'] = df_track.apply(
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\pandas\core\frame.py", line 6878, in apply
return op.get_result()
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\pandas\core\apply.py", line 186, in get_result
return self.apply_standard()
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\pandas\core\apply.py", line 295, in apply_standard
result = libreduction.compute_reduction(
File "pandas_libs\reduction.pyx", line 620, in pandas._libs.reduction.compute_reduction
File "pandas_libs\reduction.pyx", line 128, in pandas._libs.reduction.Reducer.get_result
File "C:\Users\Alonso\Desktop\proyectos\TrackEditor\plots.py", line 221, in
lambda row: geopy.distance.geodesic((row.lat, row.lon), point).km,
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\geopy\distance.py", line 389, in init
super(geodesic, self).init(*args, **kwargs)
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\geopy\distance.py", line 164, in init
kilometers += self.measure(a, b)
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\geopy\distance.py", line 410, in measure
a, b = Point(a), Point(b)
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\geopy\point.py", line 165, in new
return cls.from_sequence(seq)
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\geopy\point.py", line 405, in from_sequence
return cls(*args)
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\geopy\point.py", line 178, in new
_normalize_coordinates(latitude, longitude, altitude)
File "C:\Users\Alonso\AppData\Local\Programs\Python\Python38\lib\site-packages\geopy\point.py", line 76, in _normalize_coordinates
raise ValueError('Latitude must be in the [-90; 90] range.')
ValueError: Latitude must be in the [-90; 90] range.

Process finished with exit code 0

When using the options "save session" or "save gpx" in Docker, no file is created in host machine.
Also, a check would be needed to notify the user.

This check does not include any extra value to the project. This can be problematic with the following sequence:

1. Load file test.gpx
2. Split segment
3. Remove one of the segments
4. Reload test.gpx to get some removed information

Step 4 would not load any gpx file.

Remove and split operations may be provoking non-consecutive indexes for segments. Debug segments have been included in last commit to help the investigation.
Can be reproduced by splitting several times same segment.

Changes in matplotlib package has been shown to change some plots. Versions will be hardcoded in requirements.txt.
They will be updated periodically in a controlled approach

Segment selection is not working after loading session.
Insert time functionality works.

Too many commits have been done to create github action for pytest. For the next time a debug tool will be needed to do it offline.

Using tmate

• https://www.youtube.com/watch?v=RIeLtfGdB3w
• https://github.com/marketplace/actions/debugging-with-tmate

Using act

• https://mauricius.dev/run-and-debug-github-actions-locally/
• https://github.com/nektos/act