New workflow needed for incremental updates

[eseiver]

[From old repo]

right now, the download_check_and_move function has three methods of looking for updated article XML after entirely new articles have been downloaded to the temporary download directory:

1. check for new corrections articles in the temp directory & see if the accompanying corrected articles have updated XML, downloading any corrected articles with new versions
2. check Solr for version-of-record (VOR) updates to uncorrected proofs (status: Not working)
3. check all XML directly for updated uncorrected proofs in uncorrected_proofs_list.txt.

If an article's XML is updated for any reason other than corrections or VOR, it currently cannot be detected by searching Solr. The only way to be sure is to check every article's XML manually, as in the revisiondate_sanity_check function in corpus_analysis.py, which is time-consuming and hits journals.plos.org pretty inefficiently. JIRA no-NOR ticket labels can help with this to some degree (see #20), but that doesn't work outside of PLOS.

One solution would be using a hashtable, as in https://github.com/PLOS/allofplos_upload/issues/6. Is there any other way @sbassi?