[
https://issues.apache.org/jira/browse/TIKA-3153?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17175056#comment-17175056
]
Tim Allison commented on TIKA-3153:
-----------------------------------
Could we add a MinShouldMatchClause? Terminology is stolen directly from
Lucene...
This would be "match at least two of the following":
{noformat}
<match minShouldMatch="2">
<match value="\nMessage-ID:" type="stringignorecase" offset="0:1000"/>
<match value="\nReturn-Path:" type="stringignorecase" offset="0:1000"/>
<match value="\nX-Originating-IP:" type="stringignorecase"
offset="0:1000"/>
<match value="\nReceived:" type="stringignorecase" offset="0:1000"/>
<match value="\nDate:" type="stringignorecase" offset="0:1000"/>
<match value="\nFrom:" type="stringignorecase" offset="0:1000"/>
</match>
{noformat}
This _should_ improve precision/recall for rfc822...
The more we can do in the mime-types.xml and the less we have to do with custom
detectors, the better...I think...
> Text File identified as message/rfc822
> --------------------------------------
>
> Key: TIKA-3153
> URL: https://issues.apache.org/jira/browse/TIKA-3153
> Project: Tika
> Issue Type: Bug
> Components: detector
> Affects Versions: 1.24.1
> Reporter: Akash
> Priority: Major
> Attachments: TextFileIdentifiedAsMessage.txt
>
>
> Text file containing the word Received: is identified as message/rfc22.
> We were earlier using version 1.9 and it used to identify file type properly
> as text/plain.
> Even if multiple lines are there, if one line with Received: is present,
> content type is incorrectly identified.
> To check we can run java -jar tika-app-1.24.1.jar
> TextFileIdentifiedAsMessage.txt
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
