Improve paging functionality

Is your feature request related to a problem? Please describe the problem.

In my opinion, the code required for getting all paged results of a collection request is not ideal. In v5 I had a generic method that handled it for all types, but I couldn't achieve the same using v6. Mainly because the classes for the original request use a RequestConfiguration object and the classes for the paginated requests use a RequestInformation object and they seem to be wholly incompatible.

This is the best I could come up with:

public List<User> getAllUsers(List<String> selectAttributes, String filter)
{
    UserCollectionResponse userCollectionResponse = graphClient.users().get(
            requestConfig ->
            {
                if (!selectAttributes.isEmpty())
                {
                    requestConfig.queryParameters.select = selectAttributes.toArray(new String[0]);
                    if (selectAttributes.contains("manager"))
                        requestConfig.queryParameters.expand = new String[]{"manager"};
                }

                if (filter != null)
                    requestConfig.queryParameters.filter = filter;
            });

    UnaryOperator<RequestInformation> requestInformation =
            requestInfo ->
            {
                if (!selectAttributes.isEmpty())
                {
                    requestInfo.addQueryParameter("%24select", selectAttributes.toArray(new String[0]));
                    if (selectAttributes.contains("manager"))
                        requestInfo.addQueryParameter("%24expand", new String[]{"manager"});
                }

                if (filter != null)
                    requestInfo.addQueryParameter("%24filter", filter);

                return requestInfo;
            };

    return loadPagedEntities(userCollectionResponse, UserCollectionResponse::createFromDiscriminatorValue, requestInformation);
}

private <T extends Parsable> List<T> loadPagedEntities(
        BaseCollectionPaginationCountResponse baseCollectionPaginationCountResponse,
        Function<ParseNode, BaseCollectionPaginationCountResponse> collectionPageFactoryFunction,
        UnaryOperator<RequestInformation> requestInformation)
{
    try
    {
        List<T> entities = new ArrayList<>();

        PageIterator<T, BaseCollectionPaginationCountResponse> pageIterator =
                new PageIterator.Builder<T, BaseCollectionPaginationCountResponse>()
                        .client(graphClient)
                        .collectionPage(Objects.requireNonNull(baseCollectionPaginationCountResponse))
                        .collectionPageFactory(collectionPageFactoryFunction::apply)
                        .requestConfigurator(requestInformation)
                        .processPageItemCallback(entities::add)
                        .build();

        pageIterator.iterate();
        return entities;
    }
    catch (ReflectiveOperationException e)
    {
        throw new RuntimeException(e);
    }
}

What irks me most about this, is that I have to specify the query parameters twice and using different syntax! Especially the "%24..." is pretty ugly.

Describe the solution you'd like.

The best solution, of course, would be if the SDK handled this transparently:

List<User> allUsers = graphClient.users().getAll(requestConfig -> ...);

The next best solution would be some kind of compatibility between RequestConfiguration and RequestInformation, so the query parameters would only have to specified once.

What do you think?

Additional context?

No response

Commencez par examiner les Java SDK APIs nommées dans le rapport : RequestConfiguration, RequestInformation et PageIterator. Reproduisez le flux d’exemple de pagination et comparez la façon dont les paramètres de requête sont configurés pour la première requête et les suivantes. Le travail est terminé lorsque les appelants peuvent configurer les paramètres de requête de pagination sans les dupliquer ni utiliser des noms encodés tels que "%24select" et "%24filter".