[TIKA-1722] Tika methods that accept a File needlessly convert it to a URL

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 1.11
• Component/s: core
• Flags: Patch
• Assignee: Jukka Zitting
• Reporter: Yaniv Kunda
• Created: 27/Aug/15 00:16
• Updated: 27/Aug/15 09:52
• Resolved: 27/Aug/15 09:25

Description

The following methods:

• Tika.detect(File)
• Tika.parse(File)
• Tika.parseToString(File)

Convert the given File to a URL and use the corresponding overloaded method that accepts a URL.
This seems like a shortcut, but essentially does the following:

1. Converts the file to a URI
2. Converts the URI to a URL
3. Calls TikaInputStream.get(URL, Metadata), which then performs the following special handling:
4. Checks if the protocol is "file"
5. Tries to convert the URL (back) to a URI
6. Creates a File around the URI
7. Checks if file.isFile()
8. Calls TikaInputStream.get(File, Metadata)

The special handling in TikaInputStream.get(URL/URI) is a good optimization for in-the-wild file resources, but for internal uses it can be skipped - making Tika call TikaInputStream.get(File, Metadata) directly.