[TIKA-2815] Priority of processing EML file should be TEXT_PLAIN instead of TEXT_HTML

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 1.17, 1.18
• Component/s: parser
• Labels: eml extraction parser

Environment

Source code MailContentHandler.java, function handleInlineBodyPart()

• Assignee: Unassigned
• Reporter: Edwin Yeo Zheng Lin
• Created: 15/Jan/19 11:48
• Updated: 15/Jan/19 11:48

Description

From the source code MailContentHandler.java, handleInlineBodyPart() function, we notice that in the processing of the EML files, the priority is to get the TEXT_HTML, followed by application/rtf, and finally TEXT_PLAIN.

However, as per my explanation in TIKA-2814, the content in TEXT_HTML is not cleaned, whereas the content in TEXT_PLAIN is clean and readable. 

As such, we should set the priority to be getting from TEXT_PLAIN as the first priority. This will prevent all the unwanted words like "FONT-SIZE: 9pt; FONT-FAMILY: arial" to be extracted out, and it could even lead to faster processing speed,

I have uploaded a sample EML file here: https://drive.google.com/file/d/1z1gujv4SiacFeganLkdb0DhfZsNeGD2a/view?usp=sharing

It has both the text/html section and text/plain section, and you can see that the text/plain section is way much cleaner and readable, as compared to text/html.