hourly data summarization for current day

There are some data availability / event-driven changes Dave Costantino made on a branch for hourly summarization of the current day data. These need to be put in the spikev2 branch (or trunk after spikev2 is merged) and hourly reporting for the current day should be added to the test plan to ensure the current day pixel data summarization works.

Since hourly summarization for flowers is currently disabled, we don't need the code changes from r28527 to be brought into the playback v2 branch yet. This should definitely go into the backlog because they are required for any client using v2 playback that wants current day data summarized.

1) cron/pixelsumm-hourly-today.sh script: replace reference to pixelsummarizer.sh with lastpaidhitsummarizer.sh, do not set TSA_APP_DIR variable and replace $TSA_APP_DIR with $DIR so it points to the trunk build for v2 whsumm.sh. add "--dim-sync-refresh-time 0" to whsumm command so products pixel playback created on st-tracker will be synced over to the warehouse.

2) cron/run_alerts.sh: alerting won't work with v2 so remove flowers_hourly_alerts.sh from this script.

These shell scripts are checked into trunk under src/pixel_playback.

http://trac.ri.thesearchagency.com/cgi-bin/trac.cgi/changeset/29311

Summary of code changes from r28527:

1) Add new "skip-da-update" parameter to Warehouse Summarizer so we don't update the dataAvailability table.

2) Add new "skip-event" parameter to Warehouse Summarizer so we don't send a data availability message to JBoss.

http://trac.ri.thesearchagency.com/cgi-bin/trac.cgi/changeset/28527

"Hourly reporting for the current day" means running pixel playback for 1-2 hours of the current day, then running pix/whsumm (in the case of v2, just whsumm) and passing it the current day as the date. Run a report to verify the data. Then when after more data gets recorded, running playback and pix/whsumm again (with the current date) to make sure the data in the warehouse is correct. We want to make sure we're replacing the data correctly in the 2nd, 3rd, nth run of whsumm and we're not putting duplicate data in there.