Currently we reject all questions for oral answer, and questions which were originally for oral answer, etc, as a temporary fix to get us past launch.
Most of the code to spot and store questions for oral answer is already written, but there is a bit left to do. In particular we need to make sure we know exactly when the question number for a question for oral answer is unique. It looks to me as though there is a sequence for every combination of
calendar year, house, cluster
Where 'cluster' is something we don't yet store. Cluster should be taken to include 'questions to the deputy president' and 'questions to the president.
Steps to fix:
1) Add 'cluster' to the Question model, with appropriate uniqueness.
2) Modify question scraper to store the cluster.
3) Remove the code to ignore oral questions and rescrape.
4) Adjust the match_answers code.
Currently we reject all questions for oral answer, and questions which were originally for oral answer, etc, as a temporary fix to get us past launch.
Most of the code to spot and store questions for oral answer is already written, but there is a bit left to do. In particular we need to make sure we know exactly when the question number for a question for oral answer is unique. It looks to me as though there is a sequence for every combination of
calendar year, house, cluster
Where 'cluster' is something we don't yet store. Cluster should be taken to include 'questions to the deputy president' and 'questions to the president.
Steps to fix:
1) Add 'cluster' to the Question model, with appropriate uniqueness. 2) Modify question scraper to store the cluster. 3) Remove the code to ignore oral questions and rescrape. 4) Adjust the match_answers code.