ghost-tree output files are in a folder now; provide alternative option to name files

[JTFouquier]

ghost-tree output files should be in a folder (this might be a better option than the user selecting a name for all of their files...).

[JTFouquier]

Three output files are: ghost-tree.nwk, log/error file, and accession ID list for use with Qiime's filter_otus_from_otu_table.py. Ask for name of folder and put files there instead. Any thoughts?

[jairideout]

@ebolyen and I discussed this and recommend the following:

• Instead of creating a log/error file, have logging info output to stdout and error/warning info output to stderr. Users can then redirect this as they see fit (or not).
• Have an output directory option (similar to how QIIME scripts work) and two output filepath options (tree, ID list). These options are mutually exclusive: the user must either supply an output directory or the two output filepaths but cannot mix the two types of options.

How does this sound?

[JTFouquier]

@gregcaporaso, @jairideout, @ebolyen @wasade @johnchase, for your reference only, don't need to read.

Some of the updates I made 08/15/15 include:

Users input a folder name (#49) so that they can get the ghost-tree.nwk file
and the new ghost_tree_extension_accession_ids.txt (addresses #48) for use with
filter_otus_from_otu_table.py in QIIME. This file was something I overlooked, but it
would be annoying for a user to make since you must filter your .biom table
prior to using it for diversity analyses with ghost-tree.

These are the two default files that are saved, the rest are options for advanced users.

@jairideout & @ebolyen, thank you, I like your suggestion about "either entering file names OR a folder" but I wasn't sure how to do it in Click so this will have to do for now (I decreased priority unless you feel very strongly about it). It made sense to have a folder since there could be a lot of output files if people want to experiment with ghost-tree via the new flag options I added.

There are now more options:

"--stderr", is_flag=True, help="'saves error log file"
"--foundation_alignment", is_flag=True, help="saves non-redundant foundation alignment file"
"--foundation_tree", is_flag=True, help="saves foundation phylogenetic tree file"
"--exclude_id_list", is_flag=True, help="will not save accession id file"

-ghost-tree lets you know that it is currently running and when it is finished (addresses #43)

-added demo file folders small_ghost_tree_demo_files & large_ghost_tree_demo_files
into the workflow folder

[JTFouquier]

Note to self: when I have time, reread Jai and Evan's suggestion and try to figure out how to do this via Click.

[jairideout]

@JTFouquier your solution sounds fine for now, I think this can be taken care of during review.

stdout and stderr aren't specific to Click or even Python, they are some of the standard streams made available by the operating system (e.g., Unix) to your program. These are files that you can write to: stdout is a file for your program's output and stderr is a file for error messages, warnings, etc.

You can interact with stdout/stderr like this:

import sys
sys.stdout.write("message for stdout")
sys.stderr.write("message for stderr")

Here's a resource for learning more about stdout/stderr in the context of Python.

When you start working on this again let @ebolyen and I know, we're happy to help with designing the CLI!

[JTFouquier]

Thanks @jairideout for the tips/reading! Yeaaaaah.....my unix-ness is extremely spotty. Go 👻-tree team!