Tutorial: "Invalid index to scalar variable" during "gamtools process_nps"

[JediKnightChan]

Hi, I'm going through tutorial in the docs. The installation of gamtools was pretty non trivial from the start, it could install successfully and then trigger an error in one of the commands (eg --help), due to bad dependency handling.

Honestly, a docker image would be nice.

I managed to get to the step of "gamtools process_nps" after creating conda environment with python 3.9, installing samtools and bowtie2. However, I got this error:

/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/segregation.py:79: FutureWarning: The 'delim_whitespace' keyword in pd.read_csv is deprecated and will be removed in a future version. Use ``sep='\s+'`` instead
  return pd.read_csv(path_or_buffer,
.  Calling positive windows:50kb
/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py:379: FutureWarning: The 'delim_whitespace' keyword in pd.read_csv is deprecated and will be removed in a future version. Use ``sep='\s+'`` instead
  coverage_data = pd.read_csv(input_file,
/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py:124: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
  if (positive_windows[0] == 1) and (positive_windows[1] == 0):
/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py:127: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
  if (positive_windows[-1] == 1) and (positive_windows[-2] == 0):
########################################
TaskError - taskid:Calling positive windows:50kb
PythonAction Error
Traceback (most recent call last):
  File "/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/doit/action.py", line 403, in execute
    returned_value = self.py_callable(*self.args, **kwargs)
  File "/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py", line 382, in threshold_file
    segregation_matrix, fitting_data = do_coverage_thresholding(
  File "/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py", line 322, in do_coverage_thresholding
    sample_fitting_data = fitting_function(sample_coverage_data)
  File "/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py", line 205, in orphan_windows_fitting
    single_read_estimate = mode(nonzero_coverage)[0][0]
IndexError: invalid index to scalar variable.


/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py:379: FutureWarning: The 'delim_whitespace' keyword in pd.read_csv is deprecated and will be removed in a future version. Use ``sep='\s+'`` instead
  coverage_data = pd.read_csv(input_file,
/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py:124: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
  if (positive_windows[0] == 1) and (positive_windows[1] == 0):
/home/cat/.conda/envs/gamtools3.9/lib/python3.9/site-packages/gamtools/call_windows.py:127: FutureWarning: Series.__getitem__ treating keys as positions is deprecated. In a future version, integer keys will always be treated as labels (consistent with DataFrame behavior). To access a value by position, use `ser.iloc[pos]`
  if (positive_windows[-1] == 1) and (positive_windows[-2] == 0):

[JediKnightChan]

The issue probably arised due to fact that I had to downgrade numpy version to '1.23.5', because with the numpy version coming with pandas 1.3, mentioned in tox.ini, the following error occured when I did 'pip install gamtools':

error: ‘PyArray_Descr’ {aka ‘struct _PyArray_Descr’} has no member named ‘subarray’

After I downgraded numpy, I could install gamtools successfully, but it probably broke some code in the call_windows.py, which caused the error I mentioned firstly. Now after changing line

single_read_estimate = mode(nonzero_coverage)[0][0]

to

single_read_estimate = mode(nonzero_coverage)[0]

and rerunning process nps command, I finally got my cosegregation table and a matrix, hurray!