Show user details of the file imports in the statistics section

[julianharty]

I've added a summary of the number of files imported in the Statistics view (see #14).

I'd like to provide details of each file that's been imported. To do so, the GUI will need to be revamped somewhat. Rather than overload issue 14 I'd prefer to close it and open this enhancement. Provisionally we may be able to use the same presentation in both the statistics view and when importing files. Something to consider, not to commit to at this stage.

[julianharty]

I noticed a bug in the statistics, or at least something that might be unclear from a user's perspective. When a file's contents are rejected e.g. because the records are deemed duplicates via the UNIQUE column clause for the GOOGLE_PLAY_REVIEW table the count of the number of files imported increases. Where the entire file's contents are rejected, perhaps we shouldn't increment the files imported counter? However, once the GUI has been enhanced I hope this nuance will be dealt with.

[julianharty]

BTW: I'd like to try using a CardView, possibly in a RecyclerView e.g. see https://www.sitepoint.com/mastering-complex-lists-with-the-android-recyclerview/ for ideas.

[julianharty]

I'm also keen to track which fields have values in, and perhaps calculate various properties of each field (this seems mainly relevant for the free text fields).

There are a couple of techniques to track which fields have a value in them e.g. using bits in an integer/long
BigInteger has a setBit(int position) method and Integer can provide the number of trailing zeroes as one of the answers here shows http://stackoverflow.com/questions/15430050/getting-position-of-a-set-bit-in-an-integer

I've also looked at options of running the queries in SQL e.g. http://stackoverflow.com/questions/8360264/sql-sqlite-count-for-null-fields-over-all-columns

Perhaps we can also calculate the tendency for relationships between fields e.g. if this field is set to 5 then field B is likely to contain (C, D, or E). This might be useful for calculating a score of the probability a review will be deemed of interest. (Although there's lots of research on parsing reviews and they may well have more elegant and/or useful ways of scoring reviews...).