2nd script of category review

[MrOrz]

This PR implements script 2 in category review.

• $ node build/scripts/genBERTInputArticles.js -s <Google spreadsheet ID> -o <Output directory>
• Validation part is skipped to save time :P
• Generates JSON files with articles that has article categories satisfying all of:
  • created before latestArticleCategoryCreatedAt (sheet1 generation)
  • positive feedback > negative feedback
  • is not deleted nor blocked

Test run: around 10 minutes

node_modules/.bin/babel-node -- src/scripts/genBERTInputArticles.js -s 1Y9FrI01in2hz5eiveGknH0HE081sr7gVuk0a7hqqKuc -o 20211128 Writing feedbacks to database progress [========================================] 100% | ETA: 0s | 2989/2989 1418 positive feedbacks & 115 negative feedbacks have been written. Scanning through 15584 matching articles progress [========================================] 100% | ETA: 0s | 15584/15584

[coveralls]

Coverage increased (+0.3%) to 87.149% when pulling 6fd962b8b31eaccb0243397e42951c9dc9b934ed on genBERTInputArticles into 0eee4811f5cce5511b353da00b2565785ecf55e8 on master.

[MrOrz]

Production execution result

node_modules/.bin/babel-node -- src/scripts/genBERTInputArticles.js -s 1Y9FrI01in2hz5eiveGknH0HE081sr7gVuk0a7hqqKuc -o 20211128-production Writing feedbacks to database progress [========================================] 100% | ETA: 0s | 2989/2989 2211 positive feedbacks & 131 negative feedbacks have been written. Scanning through 14910 matching articles progress [========================================] 100% | ETA: 0s | 14910/14910

[MrOrz]

Run again on production DB & new review category to include Flow's ground truth labels added in 12/06.

$ node_modules/.bin/babel-node -- src/scripts/genBERTInputArticles.js -s 1E6S_wYkRIEgItKKnZyaXVBejiwYCS67qvtJ4u3YoFq8 -o ../groundtruth/20211209-production Writing feedbacks to database progress [========================================] 100% | ETA: 0s | 476/476 318 positive feedbacks & 61 negative feedbacks have been written. Scanning through 17111 matching articles progress [========================================] 100% | ETA: 0s | 17111/17111