Energy Wars Display

Details the infrastructure we're using for the Energy Wars eDisplay.

links

These are the eDisplays.

• [Buildings]
• [College Avenue]
• [National Road]

RPNL

Transferring

I (ghcrows13) believe that a HIP account should be used instead of my accounts. To transfer my work to another account you'd have to do a couple of things.

On the email server do the following:

1. Copy the "rpnl" directory from my home directory.
2. Copy my ".forward" file and change the paths.
3. Create rsa keys and share them with the database server.
4. Ensure that owners, groups, and permissions are appropriate.

On the database server (proto) do the following:

1. Copy

Methods

Currently this is how we're getting data from RPNL.

1. RPNL sends an email containing CSV to testcs128@cs.earlham.edu
2. testcs128 forwards it to ghcrows13@cs.earlham.edu from its .forward file
3. ghcrows13@cs.earlham.edu receives the email and does the following:
   1. appends it to mail_longterm.csv
   2. mail.store overwrites mail.tmp
   3. mail.py does the lionshare of the work

mail.py does the following:

1. scrapes off CSV attachments
2. scrapes data from the CSV
3. generates formatted data for long-term storage
4. copies formatted data to ghcrows13@proto.cs.earlham.edu:/home/users/energy/contest/daily/tmp.csv
   • uses RSA keys to bypass passwords
   • usernames must be the same for RSA keys to work
5. executes ghcrows13@proto.cs.earlham.edu:/home/users/energy/contest/daily/rpnl_store.bash
   • uses RSA keys to bypass passwords
   • the script is executed by ghcrows13 on Proto
   • this user must have the correct permissions (must be in the energy group)
6. rpnl_store.bash copies the tmp.csv into the electrical_energy_rpnl database
   • we do it this way so that permissions are less of a hassle (local permissions are simpler than remote permissions)

Display

The display is written using html, javascript, and google charts. It displays average daily consumption when compared to a baseline. It's intended to represent smaller datasets.

Data

Overview

• Data is generated by ~energy/contest/daily/execute.bash
  • executed daily by energy's crontab
• Temporary csv files are stored in ~energy/contest/daily/tmp_data/
  • useful for desktop testing
• final csv files are stored in ~energy/public_html/development/EnergyWars/
• execute.bash has some options that you can change, including:
  • start and end dates for the baseline and the contest
  • the target database
  • name -- the prefix for CSV files
  • street -- that's the string that RPNL uses

Specifics

• execute.bash generates the daily, baseline, cumulative, and cumulative baseline
• each of these are a different .bash script
• it does this for buildings, national road, and college avenue
• an overview of what each script does:
  1. generates appropriate file names
  2. executes a dynamic SQL statement that's based on arguments
  3. stores the SQL response in a file
  4. reformats the data using ~energy/contest/daily/reformat.py
  5. stores the reformatted data in ~energy/contest/daily/tmp_data/
  6. copies the reformatted data to the destination

JavaScript

• the javascript isn't as organized as it could be
• visualization.js contains functions that create objects (charts)
• css.js contains functions for manipulating the classes of HTML objects
• chart.js, NationalRoad.js, and CollegeAvenue.js are very similar
  1. download daily power readings
  2. download baseline power readings
  3. create container object hip.data.building for building names, baselines, and power
  4. create a new dashboard with appropriate options
  5. download cumulative power readings
  6. download cumulative baseline power readings
  7. create container object hip.data.building for building names, baselines, and power
  8. create a new dashboard with appropriate options
  9. set the window resize function to redraw the chart
  10. draw the chart and intialize buttons
  • setDisplay() changes which buildings are enabled
  • setViewInterval() is for previous week / next week
  • flipViewZoom() is for zooming in / out
• gc_zoom.js changes the interval to switch between zoomed in / zoomed out
• global.js is used to intialize global namespaces and variables
• html_helper has a bunch of functions for getting and modifying the html
  • used to emulate buttons
• interval.js controls what interval is displayed
  • it's just an iterator
  • provides functions for changing instantly or gradually
  • can have its state stored/restored
  • can setView() to the interval
• hip.js contains:
  • get csv / csv to array functions
  • the animation class (for setting a chart's animation to a preset)
  • dashboard creator for this project
  • building container class

HTML

• the order in which javascript files are included is important
• make_chart() is executed on start up
• column 1
  • the display_buttons are generated in code
  • names are from the buildings / houses
• column 2
  • contains dashboard and chart
    • date_slider has to exist; it's a bug
    • not sure about cumulative
  • menubar contains the previous/zoom/next buttons
• column 3
  • contains the cumulative dashboard and chart

CSS

It's magic.