release 10.0
 - update the database if the data already in the database, or delete the old data before inserting the new data
 - add unique index to qcal table and set end value as 0 if the qcal value is for single run or merged paired runs
 - modules to load qc data from other directories, especially for fastqcheck files from the repository
 - handle different cycle numbers of the paired runs for cache_query table
 - IVC base call plot all in the same scale 100%
 - stop read Q value from qcalreport file if fastqcheck files created for all lanes
 - store first indexing cycle number into database for multiplex run
 - query run recipe table for cycle count of a run
 - display cycle count for both ends of a pair of runs, which may be different
 - display qcal value for each lane

release 9.0
 - add api to delete qc data for one run
 - add graph view of cluster number against tile number as an alternative of heatmap
 - add view to show error rates of the latest runs per instrument
 - bugfix about getting the control lane data from lane_qc table to populate run_graph table
 - increase the x, y field length of table offset in database schema
 - display statistical data by week with the graph
 - store cycle number, tile number per lane in database from recipe and tile layout xml files
 - back populate run_recipe table by counting current QC or runlog data
 - display phasing information on summary page
 - retrieve project name and library name for each lane from sequence scape and display in the run summary page
 - QC data complete, check number of cycles, tiles and lanes in table lane_qc, signal_mean, errors_by_cycle and qcal against values in table run_recipe, not just check lane_qc table with startard number of tiles, lanes
 - only cache lane_summary data into cache_query table if data complete
 - get a list of runs from cahce_query table with lane_summary infomation to cache run_graph and instrument_statistics tables, and for npg_qc data loader and main page runlist
 - check whether lane summary cached in cache_query table to report data completely loaded for XML web API
 - check runlog data fully loaded against the actual number of tiles and lanes from run_recipe

release 8.0
 - analysis and analysis_lane api modules, initially set for ability to obtain data for analysis_report script, which is under npg-tracking
 - store frequency response matrix and offset data into database and simple web interface
 - add avgerage error per run by instrument statistics graph
 - get cycle number from move_z table if can't get it from errors_by_cycle table
 - store cycle number for a run into run_graph table
 - add cycle length filter to the statistics graphs
 - bug fix about getting the run end number from a single run qcalreport file name
 - bug fix to get correct lane number from fastqcheck file name of single end run
 - cope with relocation of score, rescore files of the pipeline software 1.3rc1
 - add caching scripts to the runlog loading script, to add them in the cron job list
 - get different levels of run folder directly from Latest Summary symbolic link if not given

release 7.0
 - api direct loading now gets database credentials from config.ini
 - heatmaps for pf_perc_error_rate now scales upt to 10+ (anything 10+ is considered to high, and is therefore all very hot)
 - all thumbnails for error plot view
 - store the run start, complete and end time in the database
 - add run statistics by time graph

release 6.0
 - Change to json output format for run_tile list for a run (default - single run, data structure streamlined)
 - IVC thumbnail plots now horizontal, merged image with a area map using DHTML.
 - IVC large plots now loaded using a lightbox to sit over the rotated IVC thumbnails
 - image_store table to store png that are regularly acessed for speed improvement
 - lane_qc table - change to column definitions for control lane only columns - from text to float (unknown stored as NULL)
 - indexes added to many tables, in order to improve speed efficiency.
 - Uniqueness checked for rows where applicable.

release 5.0
 - API modules to obtain summary, run list, run_tile and signal mean data
 - bugfix to ensure only single loading of rows into database
 - yield per run per instrument graph
 - blank base error rate per tile by cycle
 - bugfix to ensure most common words to be loaded into database
 - graph image now uses gray lines instead of black for clarity
 - first cycle intensities heatmap now scaled rather than linear

release 4.0
 - JSON service - run_tiles obtained but now in order, and with qc data already packaged in for the tile
 - JSON list of run ids that have data
 - caching for instrument statistical graphs
 - Error graph scale to 5% on Y-axis
 - XML feed to query if run is loaded /run/.xml 

release 3.0
 - Link to NPG run pages from the id_run/id_run_pair in the chip summary
 - Y Scale max for IVC plots set to default to 1500 if 100 tiles per lane (i.e. GA2)
 - JSON services - summary for run, obtain tiles, some stats for tile
 - different graph views possible on statistical graphs
 - run tile view - linked to when you select a tile on a heatmap or run alert
 - statistical graphs for instruments - tile count of errors by run

release 2.0
 - api to process Runlogs for move z data
 - display for move z data
 - hoverable heatmaps with clicking to run tile viewer
 - move to own database instance
 - improvements to Bustard processing
 - caching of big queries to improve performance
 - statistical graph by run

release 1.0
 - api to process Bustard files into xml, and post them to webservice
 - webservice to process bustard xml and save to database
 - set-up of database schema
 - tests
 - views showing graphs and tables
 - swift summary processed
 - heatmaps of tile data
 - alerts for out of scope metrics