Bug .GCT written by cmapPy on Windows have inconsistent line endings

[KarlClauser]

Hi Lev,

#Bug .GCT files written with cmapPy on Windows, show alternating blank lines after the top 3 lines when opened in Excel, though fine in code editor Spyder v5.12.3

Fix: The line below writes the 1st 2 lines of a .GCT file and would otherwise default to OS line_terminator of \r\n which conflicts with all other lines terminated by \n

Inconsistent line endings probably tricks Excels auto line ending recognition

C:\ProgramData\Anaconda3\Lib\site-packages\cmapPy\pandasGEXpress\write_gct.py #line 102

Write top_half_df to file

#top_half_df.to_csv(f, header=False, index=False, sep="\t")
top_half_df.to_csv(f, header=False, index=False, sep="\t", line_terminator='\n')

Please incorporate into next version. Screenshots attached.

Thanks,

--Karl
cmapPybug_inconsistentLineEndings.docx

[levlitichev]

Good catch. I think the better change would be to replace \n with os.linesep in write_version_and_dims:

cmapPy/cmapPy/pandasGEXpress/write_gct.py

Lines 64 to 65 in d1652c3

	f.write(("#" + version + "\n"))
	f.write((dims[0] + "\t" + dims[1] + "\t" + dims[2] + "\t" + dims[3] + "\n"))

[KarlClauser]

Won't that lead to \r\n line endings on Windows? We should be striving to get the entire file to be \n line endings. I seek to have a file that is identical, no matter whether it is written in linux or windows. That is how I encountered this bug.

--Karl

[levlitichev]

I understand your point, but I feel that it would be wise to follow the convention chosen by pandas to use system-specific line terminators. I confirmed (on my Mac) that the file looks the same when opened in Excel if all the terminators are either all \n or all \r\n.

f = open("A.txt", "w")
f.write(("A" + "\n"))
f.write(("B" + "\n"))
f.close()

g = open("B.txt", "w")
g.write(("A" + "\r\n"))
g.write(("B" + "\r\n"))
g.close()

[KarlClauser]

Hi Lev,

I'm a Windows guy and now-a-days Windows programs can routinely handle '\n' line terminators. For reading/writing .GCT files people upstream/downstream of me use Macs. Consequently it is important to be able to read/write and get the same result. If you force windows generated files to be '\r\n' then I'm going to have fix every one I produce with cmaPy to get the desired '\n'.

Would you please at least provide an option to specify the line terminator to be written? So long as I have a means to get '\n' then I don't care which you choose as a default.

Thanks,

--Karl