Simplify deltacode output

[steven-esser]

For the csv output, it would be a better presentation if we simply included a single path value, instead of empty or repeated paths that are redundant.

[steven-esser]

@johnmhoran Plan of attack:

1. Review + merge PR 19 fix to dict #20
2. Modify Delta object to_dict function to return only minimal info (category and path for the most part)
3. Modify CSV ouput function to handle new data structure
4. Modify JSON output to Add header information that has been removed from DeltaCode.to_dict()

[steven-esser]

Feel free to push your changes in a branch without opening a PR. This is easier and less noisy than constantly updating a PR.

If you have questions or would like me to expand in detail on any of the above items, let me know.

[johnmhoran]

@MaJuRG I've modified Delta.to-dict() and the CSV output to handle the new data structure, and fixed the 9 failing tests as well. About to add missing headers to the JSON output. Here's an excerpt from the current JSON output file testing a set of test scans with 1 added file:

{
    "added": [
        {
            "category": "added", 
            "path": "a/a5.py"
        }
    ], 
    "removed": [], 
    "modified": [], 
    "unmodified": [
        { . . .

How do we want to handle the redundant category information? Replace the 4 category keys with a single key named deltas and a value containing a list of category/path key/value pairs? Example:

{
    "deltas": [
        {
            "category": "added", 
            "path": "a/a5.py"
        },
        {
            "category": "unmodified", 
            "path": "a/a1.py"
        } . . .

Also, I assume we want the missing header info added to the top of the JSON output file. If so, I think that means adding values to the top of the incoming OrderedDict created by DeltaCode.to_dict(). Do I have that right??

[johnmhoran]

@MaJuRG I just committed and pushed my work to date so you can vet including in connection with my recent questions.

[steven-esser]

We can keep the redundant categories around for now. If you recall our conversation yesterday, the delta object's category field will no longer match the Deltas dict category keys once we add license/copyright information.

[johnmhoran]

I do recall, e.g., license_change. Thanks @MaJuRG . When you have a chance, can you sketch out an example excerpt of how you see the future JSON structure?

Re the 2nd question, do we want to add the missing header info to the top of the incoming OrderedDict. If yes, how does one do that? I've seen 2 approaches: rewrite the OrderedDict -- said to be slow but relatively straightforward -- or write a function to prepend.

Finally, once I finish this, what shall I tackle next?

[johnmhoran]

@MaJuRG I can add the version header from inside the JSON function by moving the variable from __init__.py to cli.py, but adding the stats from DeltaCode.get_stats() has eluded me so far. Do we need to restore that to the DeltaCode.to_dict() method in order to add it to the JSON output file?

[johnmhoran]

@MaJuRG I'm able to add the stats by passing new and old to generate_json and calling the stats like this: data['deltacode_stats'] = DeltaCode(new, old).get_stats(). The 2 headers (version and stats) have been added to the JSON file, though they appear at the bottom rather than the top (per my question above re adding to the top).

[steven-esser]

a = OrderedDict([
    ('header', header_info),
    ('deltacode_stats' deltacode.get_stats()),
    ('deltas', deltacode.to_dict())
])

pass the deltacode object into generate_json instead of the data dict and just call to_dict inside the csv function at the right place.

[steven-esser]

Shouldnt need to move version to cli.py. Just import deltacode at the top and reference the version like: deltacode.__version__

[johnmhoran]

Very nice. Thanks, @MaJuRG .

My only open issue: the list comprehension you suggested -- still digging into how to include the 3 variable assignments and the tuple_list.append() operation currently inside the 2nd of the nested for loops.