IBX-7149: Refactored content type-based indexing to rely on a dedicated strategy

[mateuszdebinski]

Question	Answer
JIRA issue	IBX-7149
Type	bug
Target Ibexa version	v4.5+
BC breaks	no

Maintainer update:

Given how complex and error-prone content id batch generating logic is, the fix includes the first portion of refactoring, introducing internal, command-dedicated \Ibexa\Bundle\Core\Command\Indexer\ContentIdListGeneratorStrategyInterface which produces \Ibexa\Core\Search\Indexer\ContentIdBatchList result wrapper object containing iterable batches of content Ids (meaning content id list chunks/slices) and total count of content ids.

This is the first in a series of refactoring steps, making the command more stable, as reindexing is mission-critical part of content delivery system. The refactoring allows for better test coverage of crucial parts of the logic, aiming towards more single-responsibility-oriented approach.

QA

Sanity check of ibexa:reindex command for various argument lists of the command, including different iteration count (meaning really different batch sizes) on an instance with test fixtures data.
There are foremost 2 edge cases:

• list of content items of a given content type, which exceeds the batch size (let's say are all 12 folders in the system indexed when using --content-type=folder option? - see that before the fix they weren't)
• empty list of content items for given criteria (can be e.g. --subtree, but should be tested against --content-type as well)

Checklist:

• Provided PR description.
• Tested the solution manually.
• Provided automated test coverage.
• Checked that target branch is set correctly (main for features, the oldest supported for bugs).
• Ran PHP CS Fixer for new PHP code (use $ composer fix-cs).
• Asked for a review (ping @ibexa/engineering).

[sonarcloud]

Quality Gate passed

Kudos, no new issues were introduced!

0 New issues
0 Security Hotspots
No data about Coverage
0.0% Duplication on New Code

See analysis details on SonarCloud

[sonarcloud]

Quality Gate passed

Issues
1 New issue

Measures
0 Security Hotspots
No data about Coverage
0.0% Duplication on New Code

See analysis details on SonarCloud

[alongosz]

Update:
Resolved #296 (comment) via 4aef887 + be2af07 + tests & PHPStan alignment.

Introducing ContentIdBatchList result object which is iterable and wraps both the batch list and total count, so the strategy-implementing service itself is stateless and returns that object instead. It will make more sense once other parts of if-elseif-... expression are refactored into its corresponding strategies, but that's out of scope for the bugfix.

To achieve uniform handling of iterable arrays, Generators, and Traversable objects, the ReindexCommand::runParallelProcess method now accepts iterable<int, array<int>> $batchList and repacks it into a \Generator internally.

[Steveb-p]

Personally, I'd leave the mapping function as is, and use self::assertContainsOnlyInstancesOf() to prevent test framework from increasing it's assertion count. But that is only a detail 😅

[sonarcloud]

Quality Gate passed

Issues
1 New issue
0 Accepted issues

Measures
0 Security Hotspots
No data about Coverage
0.0% Duplication on New Code

See analysis details on SonarCloud

[tomaszszopinski]

QA approved on IbexaDXP 4.5 exp.