Turn a block of hypothesis document annotations into a plaintext list of detailed comments for a peer review

HypothesisToReview.rsc

module HypothesisToReview

import lang::html::IO;
import IO;
import List;

// this is a simple HTML scraper against an untyped HTML document model (`node`)
str detailedComments(loc hypothesisCards) {
   content = readHTMLFile(annos);
   review = "";

   for (/card:"li"(_,class="annotation-card") := content) {
       if (/"blockquote"(["text"(str quote)]) := card) {
           if (/"div"(["p"(["text"(str comment)])],class="annotation-card__text") := card) {
              review += "\> <quote>
                        '  <comment>\n\n";
           }
           else {
              review += "\> <quote> - highlighted\n\n";
           }
       }
       else {
           throw "no quote? <card>";
       }
   }

   return review;
}

• we use an untyped HTML model here, just because. We are matching against nodes like "li" and "div" that do not have an ADT declaration.
• the deep match / operator allows us to skip all kinds of uninteresting structure and matching class attributes using "keyword field matching" allows us to anchor where we want to find information again.
• += concatenates string templates lazily for maximal efficiency (it uses an internally balanced tree to avoid stack overflow)

👍