Package org.apache.tika.pipes.api

Class ComponentIds

java.lang.Object

org.apache.tika.pipes.api.ComponentIds

public final class ComponentIds
extends Object

Naming rules for pipes component instance ids (fetchers, emitters, iterators).

Ids beginning with "__" are reserved for components the host wires up for its own use -- tika-server's upload fetcher, for instance. Reserving a namespace rather than enumerating names means a new internal component cannot be reached from a request by someone forgetting to add it to a list.

Callers must normalize(String) an id before comparing or storing it, and must validate the normalized form. Validating the raw id and resolving the trimmed one would let " __foo" slip past the prefix check and then bind to __foo.

Field Summary

Fields

Modifier and Type	Field	Description
static final Pattern	LEGAL_ID	Ids are restricted to this set so a normalized id can be used verbatim everywhere.
static final String	SYSTEM_PREFIX	Prefix reserved for host-wired components; not nameable from user config or a request.

Method Summary

Modifier and Type	Method	Description
static boolean	isSystem(String id)	True if the normalized id is in the reserved system namespace.
static String	normalize(String id)	Canonical form of an id: trimmed, or null if null.
static String	requireLegalId(String id, String what, String where)	As requireUserId(java.lang.String, java.lang.String, java.lang.String), but permits the system namespace.
static String	requireUserId(String id, String what, String where)	Normalizes id and rejects it if it is blank, illegal, or system-reserved.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

SYSTEM_PREFIX

public static final String SYSTEM_PREFIX

Prefix reserved for host-wired components; not nameable from user config or a request.

See Also:

• Constant Field Values

LEGAL_ID

public static final Pattern LEGAL_ID

Ids are restricted to this set so a normalized id can be used verbatim everywhere. Trimming alone leaves NBSP and zero-width characters, which produce ids that look reserved but resolve to nothing; the allowlist refuses them outright. It also keeps path separators and quoting characters out of anything derived from an id. System ids satisfy it too -- _ is a member.

Method Details

normalize

public static String normalize(String id)

Canonical form of an id: trimmed, or null if null.

isSystem

public static boolean isSystem(String id)

True if the normalized id is in the reserved system namespace.

requireUserId

public static String requireUserId(String id,
 String what,
 String where)

Normalizes id and rejects it if it is blank, illegal, or system-reserved.

Parameters:

id - the raw, caller-supplied id

what - component kind, for the message (e.g. "fetcher")

where - origin, for the message (e.g. "configuration" or "a request")

Returns:

the normalized id

Throws:

IllegalArgumentException - if the id is blank, illegal, or system-reserved

requireLegalId

public static String requireLegalId(String id,
 String what,
 String where)

As requireUserId(java.lang.String, java.lang.String, java.lang.String), but permits the system namespace. For ids the host itself supplies, where the charset rule still applies but __ is legitimate.

Throws:

IllegalArgumentException - if the id is blank or outside LEGAL_ID