[feature]: Include cell reference details in error summary

[deeonwuli]

📌 References

• Issue: Closes https://app.clickup.com/t/4528615/869dg8vmf

📝 Implementation

• Added ImportRowLookup, a class that maps every metadata/org unit/TEI/attribute id present in an imported Excel template to its sheet and cell location ({ sheet, row, column }) before the import is sent to DHIS2
• Column letter is extracted from cell.ref in all three Excel reading paths: readByRow (event programs), readTeiEvents (tracker program events), and readTeiRows (tracker program TEI attributes)
• Import errors now show "Found in sheet X, cell B6 of the Excel file" in the Details column instead of just row number; row-level ids (org unit, TEI, program stage) that have no column association continue to show "row N"
• When both a cell-level and row-only location exist for the same row, the cell ref is preferred
• Tracker import errors that DHIS2 reports at multiple hierarchy levels (e.g. both ENROLLMENT and TRACKED_ENTITY) are deduplicated by message — the copy with location details is kept, the rest dropped

🔥 Notes for the reviewer

📹 Screenshots/Screen capture

📑 Others

#869dg8vmf

[bundlemon]

BundleMon

No change in files bundle size

Groups updated (1)

Status	Path	Size	Limits
✅	Build Folder ./**/*	1.74MB (+1.18KB +0.07%)	+20%

Final result: ✅

View report in BundleMon website ➡️

---

Current branch size history | Target branch size history

[gqcorneby]

Thanks @deeonwuli! Nice work 🎉 the tests are thorough. A few in-line and general comments

1. Translations - Can we check and remove fuzzies in translation files?
2. Cell based data source - As stated in the description, only rows, teis, and events are mapped back to the template. Cell based data sources (usually datasets), are not supported. @MiquelAdell Do we want to include this in this PR or as a separate backlog?

Heads-up: Need to double check but I think we can't just set group = buildRowNumber(cell.ref) in readByCell. Group is the grouping key at ExcelReader.ts:89, so distinct row numbers would re-split the aggregated payload. It probably needs sheet/row/column carried on the data value itself, with the lookup indexing by cell.

file you can use to test:
NHWA Module 1_v1 (1).xlsx

[gqcorneby]

This works for row based data sources but follows an imperative pattern. The same output can be built as id → location pairs via flatMap, then a single groupBy. What do you think?

[gqcorneby]

Grouping by message alone seems correct for the intended case, maybe we can add it as a comment to add some context?

There's a theoretical hole: two distinct errors sharing an identical generic message where only one resolved a location, the un-located would get dropped as if it were a duplicate. I think it's not a real scenario but opening it up for discussion. Cheap fix would be to update the grouping to .groupBy(e => ${e.message}@@${e.details}`

[MiquelAdell]

Cell based data source - As stated in the description, only rows, teis, and events are mapped back to the template. Cell based data sources (usually datasets), are not supported. @MiquelAdell Do we want to include this in this PR or as a separate backlog?

No, since this is paid by the client and the client only uses tracker programs let's not do that in this one.
Thanks!

@deeonwuli you can address the rest of the comments whenever you have a moment.
Thanks!