North American Sprint Scoring

This is software to help score the North American Sprint ham radio competition. The process begins with a collection of Cabrillo log files that contain records of ham radio contacts that took place during the running of the NA Sprint contest.

The purpose of this software is to verify the information in the Cabrillo log files and apply the scoring rules to calculate a score for each entry. Verification can take many forms such as is the date & time in the period of the contest, and it can include cross checking logs against each other.

The main difficulty of scoring ham radio contests is bad data. People will submit logs that do not match the specification, and they will also incorrect incorrect data in the various fields.

The overall approach of this software is to put all the relevent log and contact data into a SQL database. The verification and cross checking is processed using SQL queries and additional matches in the program.

Getting the logs into shape and into a database

0. Edit cabrillo.rb and set CONTEST_START and CONTEST_END appropriately.
1. Review logs for matching system and other issues ruby testcab.rb --checkonly *.log. Use a powerful editor like GNU Emacs to edit the logs. The rectangle-orient commands in GNU Emacs are particularly useful.
2. Send the list of "missing logs": ruby testcab.rb --checkonly --missing *.log to the contest chairman. The chairman may choose to send email to try to get more logs. If successful, return to the previous step and check the new logs.
3. Read the logs into the MySQL database: ruby testcab.rb --checkonly --new --name "Fall SSB Sprint" --year 2016 --populate *.log
4. If you need to start over (reload all the logs), it's ruby testcab.rb --checkonly --new --name "Fall SSB Sprint" --year 2016 --restart --populate *.log
5. If there are callsigns whose multiplier or entity ID come up wrong, you can make an overrides.yaml file and add --overrides overrides.yaml to the command line when you upload the logs into the database.

Cross matching the QSOs in the database

Now the logs and QSOs are in the database. You don't need to mess with the log files again unless you discover issues that need to be fixed or more logs arrive.

1. Make some directories the program expects to already exist: mkdir xml_db output
2. Start the cross matching by running ruby testcross.rb --name "Fall SSB Sprint" --year 2016 --qrzuser callsign --qrzpwd password where callsign and password are a subscribers callsign and password. To redo the cross matching, add --restart.
3. Review the name mismatches, grep "name mismatch" output/*_cab.txt. You may want to add new homophones to the homophones.csv file in the src directory. They will get added to the database the next time you run. If you change the homophones.csv file, you will need to run again with --restart.
4. Set start and end in the Contest table of the MySQL database for this contest running. (Go into MySql)
5. Generate the report ruby asciireport.rb --name "Fall SSB Sprint" --year 2016
6. Load up scores_Fall_SSB_Sprint2016.csv and toxic_Fall_SSB_Sprint2016.csv in a spreadsheet program and review. In particular, format the % Toxic column of the toxic report as a percentage, and then sort the whole toxic worksheet by % Toxic decreasing. This will cause the most toxic logs to appear at the top of the list. A toxic log is one that gives lots of partials, NILs or other bad QSOs to other stations. Any log with a toxic percent higher than 10% deserves review and maybe some less than 10% too.
7. For a toxic log, look at what the other logs got dinged for. For example, if the toxic log was NS6T I use, awk '$1=="QSO:" && $10=="NS6T" { print}' output/*_cab.txt to get a list of all the QSOs from the logs who worked NS6T. This assumes you have the Linux/Unix tool awk. Any consistent error across multiple logs may indicate a mistake in the sent information of NS6T's log. For example, NS6T's log might have a sent name of THOMAS, but if everything logged him as TOM giving the all partial matches, chances are that NS6T actually transmitted TOM instead of THOMAS. Use your judgement. You may end up editing NS6T's log and going back to loading the QSOs from the log file. Ultimately, you may also decide that a log is so unreliable that it's better to leave it out of the scoring. For example, it is taking QSO credit away from too many other stations because of errors most like in the toxic logs logging or operating.
8. Once we've got all the logs and you've come to terms with the toxic logs, it's time to load in the team information. Generally, someone from the SSB Sprint committee will be able to send you a team-registration-2017-02-05.csv file or something similar. It's a 15 column comma separated value file. Column 1 is the manager's call and column c in the manager's email. Column 3 is the team name. Column 4 is the number of members. The next ten columns are two per team member. The first is the team operator's callsign, and the second is the callsign used. The last column is the date and time the team was registered. The command to load team information is: ruby loadteam.rb --name "Fall SSB Sprint" --year 2016 team-registration-2017-02-05.csv