This commit splits the feature of Spring Initializr in several modules:
* `initializr-generator` is a standalone library that is responsible for
generating projects based on a `File` directory. It has a minimal set of
dependencies and is not web-related
* `initializr-web` provides the web integration (project generation,
meta-data, etc)
* `initializr-actuator` is an optional module that can be added to
support project-generation-specific statistics
Closes gh-214
This commit replaces the query tokenizer with a "nonword" tokenizer, in
order to mirror the one used for the indexation process.
This allows users to enter dashes in their search queries such as
"data-" and still get meaningful results.
Fixes#209
If an elastic instance is available, publish a document to a
configurable index every time a `ProjectRequest` is handled by the
service.
In practice, this means that every attempt to generate a project leads to
a new document in the index. The document gathers the settings of the
required project, including invalid ones if any. If an exception is
thrown, the message of the cause is made available.
CloudFlare is explicitely supported and the IP and country of the request
is added to the document. If that information is not available and the
request contains a `X-Forwarded-For` header, the value is also associated
with the document. If an IPv4 is detected, it is set in a separate
`requestIpv4` property.
If for some reason the document could not be indexed, we attempt to retry
a configurable amount of times.
Closes gh-185
Parse the `User-Agent` http request header to determine if the client is
known. If that is the case, increment the relevant `client_id` counter.
A `ProjectRequest` had now a generic `parameters` array with all the
request headers by default. Since we don't want to accidentally map any
of those from a form input, `BasicProjectRequest` contains only "public"
fields and is the type exposed at the controller level.
Closes gh-193
`1.3.2.RELEASE` is available now so there is no need to have a version
range on the snapshot. We don't need to copy/paste the version of the
bom in the actual dependency (that's the whole business of the bom
definition (including versions mapping according to the Spring Boot
version).
We want the `cloud-bom` to be automatically added when the data flow bom
is required so it is referenced as "additional BOM" as well.
Closes gh-189
BOMs are structured in such a way than adding one to the project may
require another one(s) to be added to the project.
This commit adds an `additionalBoms` property to `BillOfMaterials` that
can be used to refer the BOM(s) that should be automatically added if
said BOM is added to the project.
Closes gh-190
Previously, Spring Initializr was refering to
`spring-cloud-starter-parent` as the BOM source for any cloud-related
dependencies.
This pom actually inherits from a specific version of Spring Boot and
brings all its dependency management. If the chosen boot version and the
one that the cloud bom brings do not match, inconsistent versions can
appear in the project.
The underlying issue was fixed by providing a new BOM that only provides
dependency management for the cloud-related bits.
Closes gh-116
