lewislabucsd / glycompare Goto Github PK

Best practices (as far as I understand) say we should not save data to files from within a function by default. Can we make this an optional functionality?

select_profile_name=[] --> columns of the abundance table, otherwise this list

For example, if someone wanted to run an analysis to compare linked with unlinked analysis results, they would need to spin up two separate sessions. Additionally, they won't necessarily know that the init file exists.

I think that these may be behaving like global variables which are not recommended for a package (see #12 & #18)

https://github.com/bkellman/GlyCompare/blob/5682dad4df50b660c01fc141f16a9bd713116d15/src/customize_motif_vec.py#L130

I'd like to suggest an alternative to this function which sets the keys of a dictionary to the glycoct strings. This avoids the costly subtree function and I think speeds things up a lot. What do you think bokan?

def merge_motif_dict_pipe2(glycan_motif_dic):
	# param glyca_motif_dic, glycan motif dictionary output from extract_motif.get_motif_pip
	dic_out = {}
	# iterate over glycans
	for gk in glycan_motif_dic:
		# iterate over glycans of size i
		for m_size in glycan_motif_dic[gk]:
			for m_ij in glycan_motif_dic[gk][m_size]:
				m = str(m_ij)
				if m in dic_out.keys():
					dic_out[m].append(gk)
				else:
					dic_out[m] = [gk]
	return dic_out

This can go in in a vignettes folder. But we need to go through his functions and see if we want to integrate any of his functions.

We need to clean up the src/ directory. It should only contain source code.

num_processors --> used in merge_substructure functions

only_substructures_start_from_root --> boolean value whether to start from core
core='' --> string representation of the root
drop_parellel=False --> used in selectmotif.NodeState.nodes_dropping_pipe() , think is used to check if you can drop parallel node edges???
drop_diff_abund=True --> used in selectmotif.NodeState.nodes_dropping_pipe(), think is used in to see if you need to generate edges????
select_col=[] --> columns passed into the substructure_abd_table for getting substructure weight
remove_core=True --> remove core from the nodes???

unique_glycan_identifier_to_structure_id=False --> maybe new simple_profile????
already_glytoucan_id=False --> maybe new simple_naming parameters????

It seems that there is a breakdown somewhere in this code that results in it ignoring that fuc1 is a substructure of fuc2. Can we fix that?

'glytoucan_database.json' is not in the data package. I believe that causes

Unexpected error: <class 'AssertionError'>
Traceback (most recent call last):
  File "/Users/Leria 1/Library/Mobile Documents/com~apple~CloudDocs/Documents/GlyCoCan_PhD/VU-code/test_glycompare_python_version.py", line 36, in <module>
    structure_loader=structure_loader)
  File "/usr/local/lib/python3.7/site-packages/glycompare-1.0.3-py3.7.egg/glycompare/pipeline_functions.py", line 243, in load_glycans_pip
AssertionError: missing glycan G79412GP