Extra <i character in archived MUR ORC text causing display issue

[lbeaufort]

@lbeaufort commented on Wed Mar 14 2018

• Take an elasticsearch backup snapshot
• Manually remove <i characters from archived MUR Investigate and fix 500 errors for unexpected parameters passed to /legal/search/ endpoint #2934
• Take another snapshot
• Sync up environments

Some MURs aren't being displayed properly in the datatable results. They should all line up in a grid.

Example: https://www.fec.gov/data/legal/search/murs/?search=&mur_respondents=Nevada%20NOW&offset=20#results-murs

Screenshot:

I think this might have to do with https://github.com/fecgov/fec-cms/blob/develop/fec/legal/templates/partials/legal-search-results-mur.jinja

---

@lbeaufort commented on Thu Apr 26 2018

Update: The OCR text for one archived MUR contained the characters <i which the browser was interpreting as an <i> HTML tag:

<i ~="" 3xp.uditiae.="" <em="">
</i>

(Search results show this MUR with the "3xp.uditiae" text)

Thank you @vrajmohan for your great detective work on this one!

Because archived MURs are static and OCR'd by our team, the plan for now is to manually update the archived MUR text with an Elasticsearch PUT request: curl -XPUT $URI/.../mur_2934 -d @file_name but if this continues to be a problem we might want to explore a systematic approach..

[lbeaufort]

Update: still waiting on outstanding issues in the dev environment to sync up environments by making the manual fix

[lbeaufort]

Updated all environments, made backups