Add output file (--outfile) command line argument

shaarli / python-shaarli-client

Python3 CLI to interact with a Shaarli instance

https://python-shaarli-client.readthedocs.io/

MIT License

44 stars 10 forks source link

Add output file (--outfile) command line argument #23

Closed nodiscc closed 6 years ago

nodiscc commented 6 years ago

add --outfile optional parameter: specify a filename to write the API request result to
 * by default output is redirected to standard output
 * add -f and -o short options (format/output)
 * overwrite output file on each run, don't append (appending will cause a malformed json file)
 * move json import to utils.py
 * move response formatting and output to separate functions
 * styling fixes, tab to spaces, alignment, sort imports
 * use a context manager to handle writing output file
 * Handle OSError execptions

nodiscc commented 6 years ago

https://github.com/shaarli/python-shaarli-client/pull/23/commits/a2a3190212d3feb85fa01f39689779ec61821244 tries to address

printing to console and writing to files should be split over two separate functions, with a bit of logic to process argparse flags in the main() function

Or did you mean something else?

TODO test scenarii: for each supported format:

[ ] check a file is written
[ ] check an existing file is overwritten
[ ] check the file can be read, and its output loaded using the corresponding Python module

nodiscc commented 6 years ago

Hi, sorry for the delay, I think this PR should be in working order now (?).
I will have a look at https://github.com/shaarli/python-shaarli-client/commit/95623f43dbbfbf10d7590cecfc47def9b26b06a8 and try to add the tests you suggested.

nodiscc commented 6 years ago

Fixed, but at this point I have no idea how to implement the tests, because I see nothing that deals with the program output or files on disk in the current tests...

Pardon my weak Python-fu, if you could point me to a draft test/pseudocode/similar example to do this the right way I could probably start from there

virtualtam commented 6 years ago

Sure :)

Basically what such tests do is:

create a temporary folder to serve as a workspace for the test
optional: populate the temporary workspace
call the tested function(s)
run checks on generated files or resources (if any, and if needed)

This is done by using Pytest's tmpdir fixture, which in turn leverages the py.path high-level filesystem API.

More specifically, tests could cover these functionalities:

format_reponse(): ensure the output is properly formatted for the available output formats
write_output(): ensure the correct exception and message are raised upon a file access error

nodiscc commented 6 years ago

Superseded by #31