In some cases, working on a large omnibus compilation involves working with poems whose HTML transcriptions are crude or nonexistent. Transcribing poetry, while an advanced production technique, is somewhat more feasible than transcribing prose because of the much smaller amount of text compared to prose. However, poetry formatting (e.g. indentation) poses some challenges for transcription. Manually inserting the whole array of <span> elements and breaks would be unwieldy at best. Instead, it can be easier to use a shell script to convert temporary markup into the final desired HTML:
+ #!/bin/bash
+
+ # Repeat this descending pattern for more indentation levels
+ se interactive-replace 'v####(.*)' '<span class="i3">\1</span><br/>' $1
+ se interactive-replace 'v###(.*)' '<span class="i2">\1</span><br/>' $1
+ se interactive-replace 'v##(.*)' '<span class="i1">\1</span><br/>' $1
+ se interactive-replace 'v#(.*)' '<span>\1</span><br/>' $1
+ se interactive-replace '<br/>\s+</p>' '</p>' $1
+ se clean $1
+
+ Here is an example of the temporary markup:
+<p>
+v#This line is not indented.
+v##This line is at i1.
+v###This line is at i2.
+v####This line is at i3.
+</p>
+ Running the script on this file produces this HTML:
+<p>
+ <span>This line is not indented.</span>
+ <br/>
+ <span class="i1">This line is at i1.</span>
+ <br/>
+ <span class="i2">This line is at i2.</span>
+ <br/>
+ <span class="i3">This line is at i3.</span>
+</p>
+ Do not run se modernize-spelling . on poetry projects. Because changes can affect the rhythm of poetry, we generally do much less modernization than for other projects, so words like “to-day” and “some one” wouldn’t be modernized as they usually would be.