Skip to content

Conversation

@github-actions
Copy link
Contributor

No description provided.

thoniTUB and others added 30 commits June 2, 2025 09:50
improve error message/stack trace on MigrateCommand storage
update codeowners file and dependabot config
… is actually contained. (#3750)

This avoids cases, where dates are improperly merged for multiple connectors
* adds warning to CachedStore, if there are duplicate keys (This can be due to seriliaziation errors - not misues of the underlying Store)

---------

Co-authored-by: MT <12283268+thoniTUB@users.noreply.github.com>
Implements substring logic for SelectFilter(s) and MappableSingleColumnSelect derivates
also merges SingleSelectFilter logic to use MultiSelectFilterNode under the hood with Singleton-Set, as that is probably exactly as fast in our applications.
Reworks all API Endpoints that used resolved Id-References on-the-fly. We now explicitly have to resolve all Ids.

This also introduces Id#getDomain, which is the storage or Storage provider of an Id (i.e. the Domain on which it can be resolved). Ids without domains can exist as a kind of query into the store, but this should be avoided. Instead they should be derived from their corresponding object.  
Ids are strictly separated into the MetaStorage domain and the Namespaced domain:

* MetaStorage containing user dependent information or configuration
* Namespaced relates to Datasets and Concept configuration.

NamespacedIds and -Objects try to always delegate to their DatasetId as root for any Domain related tasks. MetaIds and -Objects usually hold their own domain as there are no obvious roots.

---

We refactored the API away from explicit object usage to reduce implicit conversions, but also to make it easier to use with openapi specs. 

Alongside, we also cleaned up the usage of resolve and getId, where they sometimes would be followed by each other in close succession, making them meaningless. We also tried refactoring BucketManager#create, though it still has some circular dependency with WorkerStorage and does IO at inopportune times in startup.
warn if an encoded string contains unencoded characters
ignores WorkerInformation#entityBucketSize
Remove initial filter value loading upon opening a concept node
Refactors BucketManager to avoid heavy IO in #create:
Now only works on Ids (i.e. Store-keys) which is much faster and avoids circular dependencies with WorkerStorage.
We also create a new Entity2Bucket store, tracking the existence of entities on Shards.
adds validation of inserted Table against SQL Server
thoniTUB and others added 9 commits July 1, 2025 13:26
We cannot rely on the injection of a specific namespaced storage because we have "global" object mappers attached to mina that are not bound to a specific namespace. Where possible, we inject the storage directly (API, storage itself)
fix npe on missing JacksonInject in InterToExternMapper
@thoniTUB thoniTUB merged commit 7860514 into master Jul 15, 2025
35 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants