- Renamed the articles.txt to something more specific
Changes to both scripts:
- Updated filenames to match the new standard
- Reworked the logging code so that it can write to stderr by
default. Because we can only call logging.basicConfig() once, this
eneded up being a bigger changes.
- Caused scripts to output git commits and export to track which code
produced which dataset.
- Caused programs to take files instead of directories as
output (allows us to run programs more than once a day).
Changes to the wikipedia_views/scripts/fetch_daily_views.py:
- Change output that it outputs a sequence of JSON dictionaries (one
per line) as per the standard we agreed to and which is what
Twitter, Github, and other dumps do. Previous behavior was to create
output a single JSON list object.
- A number of other small changes and tweaks throughout.
