Awesome
ph-diver
ph-diver - PH Digitally Versioned Resources - in collaboration with
The modules contained in this repository provide access to versioned resources that reside on several external resource types like HTTP servers, local disks or in-memory data structures.
This library consists of the following submodules:
ph-diver-api
- contains the basic API like version structuredph-diver-repo
- contains the data structures for a generic repository of version objects that can read, write and delete data. Also contains an in-memory based repository and afile system based repository.ph-diver-repo-http
- contains specific support for HTTP based repositoriesph-diver-repo-s3
- contains specific support for AWS S3 based repositories
The reason why the several types of repositories are separated, is mainly because of specific runtime dependencies needed, and to avoid that your dependencies are bloated if you only need a specific kind of repository.
DVR Coordinate
The DVR Coordinate, short for Digitally Version Resource Coordinate, is an identifier for any technical artefact (file) very similar to Maven coordinates.
Hint: The original term was "VESID" which was very much focused on validation artefacts. Each VESID is a DVR Coordinate, but not vice versa. DVR Coordinate defines the syntax constraints required to be adhered to by all applications. The terminology was changed for version 2 (DVRID) and version 3 of the library.
DVR Coordinate Contents
Each DVR Coordinate consists of a combination of:
- Mandatory Group ID
- Represents an organisation or group that provides a set of artefacts. That must be using the reverse domain name notation (as in
com.helger
) - It MUST NOT be empty and follow the regular expression
[a-zA-Z0-9_\-\.]{1,64}
- The usage of dot (
.
) in a Group ID represents the separation of different hierarchy levels (e.g. directory and sub-directory). - The Group ID MUST be treated case sensitive
- Represents an organisation or group that provides a set of artefacts. That must be using the reverse domain name notation (as in
- Mandatory Artefact ID
- Uniquely represents an artefact offered by a specific group. Artefact IDs must be unique per Group ID in which they are used.
- It MUST NOT be empty and follow the regular expression
[a-zA-Z0-9_\-\.]{1,64}
- The Artefact ID MUST be treated case sensitive
- Mandatory Version Number that enforces strict ordering
- Each Version Number must be unique per combination of Group ID and Artefact ID
- The usage of semantic version supports the strict ordering of elements
- Each version must follow either the form
major[.minor[.micro[-classifier]]]
wheremajor
,minor
andmicro
must be unsigned integer values (like 1 or 2023) or the formclassifier
which is interpreted as0.0.0-classifier
. - The version classifier
SNAPSHOT
is a special case and identifies "work in progress" artefacts that are not final yet - The Version Number MUST be treated case sensitive
- Optional Classifier
- It MAY be empty and follow the regular expression
[a-zA-Z0-9_\-\.]{0,64}
- The Classifier MUST be treated case sensitive
- It MAY be empty and follow the regular expression
The limitations in the allowed characters for the different parts are meant to allow an easy representation on file systems.
DVR Coordinate string representation
Each DVR Coordinate can be represented in a single string in the form groupID:artifactID:version[:classifier]
.
The string representation of version numbers is a bit tricky, because 1
, 1.0
and 1.0.0
are all semantically equivalent.
Thats why it was decided, that trailing zeroes for minor and micro versions are NOT contained in the string representation, to be as brief as possible
So e.g., for version 1.0.0
the string representation must be 1
; for version 3.2.0
, the string representation must be 3.2
.
Versions using a classifier like 3.0.0-SNAPSHOT
are represented as 3-SNAPSHOT
.
Versions that only consist of a classifier like 0.0.0-XYZ
are represented only as the classifier XYZ
.
That is a work around to be able to handle all kind of versions, but they are treated with a major version of 0, a minor version of 0 and a micro version of 0.
Repository
A repository is an abstract tree like structure to act as the source for artefacts (files).
Each repository item is uniquely addressed with a RepoStorageKey
that basically is a path structure.
The content of a repository item is represented via class RepoStorageItem
.
A repository itself is represented as implementations of class IRepoStorage
.
Each repository is always readable, and optionally writable and optionally allows for deletion.
Several repositories may be chained together for reading. E.g., first the local file system is queried for a resource - if the artefact is not found locally, another remote repository might be used instead. The local caching of remote resources is also supported, to limit the necessity for external access.
Table of Contents per Group ID and Artefact ID
Since v1.0.1 a special "table of contents" (ToC) is supported per Group ID and Artefact ID.
It contains all the versions of that combination and allows for easy access of the latest version, without iterating any directory structure.
The filename used is toc-diver.xml
per default.
An extended API is available for repository storage implementations via the IRepoStorageWithToc
interface.
Maven usage
Add the following to your pom.xml
to use e.g. the HTTP repository artifact, replacing x.y.z
with the latest version:
<dependency>
<groupId>com.helger.diver</groupId>
<artifactId>ph-diver-repo-http</artifactId>
<version>x.y.z</version>
</dependency>
Alternate usage as a Maven BOM:
<dependency>
<groupId>com.helger.diver</groupId>
<artifactId>ph-diver-parent-pom</artifactId>
<version>x.y.z</version>
<type>pom</type>
<scope>import</scope>
</dependency>
News and Noteworthy
- v3.0.1 - 2024-09-15
- Moved classes
RepoToc1Marshaller
andRepoTopToc1Marshaller
to sub-packagejaxb
- Added new method
DVRVersion.getStaticVersionAcceptor
- Moved classes
- v3.0.0 - 2024-09-13
- Renamed
DVRID
toDVRCoordinate
- Made a lot of API changes and extension on the API part. Now it is stable.
- Renamed
- v2.0.0 - 2024-09-12
- Renamed
*VESID*
to*DVRID*
- Renamed
IVES*
toIDVR*
- Renamed
IPseudoVersionComparable
toIDVRPseudoVersionComparable
- Removed all deprecated APIs marked for removal
- Moved
DVRID
related classes in packagecom.helger.diver.api.id
- Added class
DVRException
- Renamed
- v1.2.0 - 2024-04-25
- Extended the API of
IRepoStorageWithToc
withgetLatest(Release)Version
- Extended the API of
RepoToc
- Replaced the enum
EVESPseudoVersion
withIVESPseudoVersion
andVESPseudoVersionRegistry
- Now supporting the following pseudo versions:
oldest
,latest
andlatest-release
- Extended the API of
- v1.1.1 - 2024-03-29
- Updated to ph-commons 11.1.5
- Ensured Java 21 compatibility
- v1.1.0 - 2024-03-09
- Extracted
RepoStorageKeyOfArtefact
fromRepoStorageKey
[backwards incompatible change] - Class
RepoStorageHttp
got an API extension, so that the used HTTP requests can be customized - Added a top-level table of contents (ToC) service that contains all groups and the artefacts of all groups (via
IRepoTopTocService
) and an XML based implementation (classRepoTopTocServiceRepoBasedXML
) - Added a new interface
IRepoStorageAuditor
to be able to handle accesses to the repository - Extended
RepoToc
API - Renamed
RepoTopToc
toRepoTopTocXML
- Reworked
RepoStorageItem
toRepoStorageContentByteArray
andRepoStorageReadItem
and extracted interfaces for both of them - Changed the writable repo API to use
IRepoStorageContent
instead ofbyte[]
for stream based activities - Extracted
IRepoStorageType
interface
- Extracted
- v1.0.2 - 2023-12-12
- Restricted VESID part maximum lengths - defaults to 64 but customizable via
VESIDSettings
.
- Restricted VESID part maximum lengths - defaults to 64 but customizable via
- v1.0.1 - 2023-11-07
- Added support for a "Table of contents" per Group ID and Artefact ID.
- v1.0.0 - 2023-09-13
- Initial version with support for in-memory, file system, HTTP and S3 repositories
My personal Coding Styleguide | It is appreciated if you star the GitHub project if you like it.