fix(csv): empty CSV must fail fast with clear input-error message

[divyasinghds]

The cascade

`CSVIngestor.read_data` caught `pd.errors.EmptyDataError` and silently returned an empty generator with only a YELLOW warning. From there:

1. Ingestor proceeded with an empty record stream
2. Created an empty MySQL table
3. Called `send_generate_edge_label_meta` against zero rows
4. Backend returned HTTP 400: `{"table_name":"No data found for table X and injestor_id Y"}`
5. PR fix: fail loud when backend dataset registration fails (no silent half-ingest) #187's fail-loud handler surfaced that as: `Backend rejected edge-label metadata; the dataset was NOT registered (its rows are already in the database).`

The user-visible message blames the backend and (worse) falsely tells the user their rows are in the database, when the real cause is an empty input file. Same shape #213 documents for self-supervised + label mismatch — a fault at one layer surfaces with a misleading message several layers down.

Repro

Adversarial test F8 on v0.3.10-rc1: a zero-byte CSV produced:

```
WARNING | csv_ingestor.py:493 | Empty CSV file: /data/shared/divya-adv2/F8_empty.csv
ERROR | api/client.py:393 | Error generating edge label metadata: HTTP 400: {"table_name":"No data found for table name e2etest_adv2_f8_... and injestor_id ..."}
ERROR | base.py:985 | Error during ingestion: Backend rejected edge-label metadata; the dataset was NOT registered (its rows are already in the database).
```

— the empty-file cause is buried in a WARNING that's easy to miss, and the visible ERROR is misdirection.

Fix

Raise `ValueError` at the read layer with a source-truthful message:

```
Empty CSV file: . The file has no header and no rows. Either
stage a non-empty CSV at this path, or check the cluster-side path
(the chart mounts your PVC at /data/shared/ — confirm staging
completed before helm install).
```

DataValidator's existing "No data found to validate" / preflight path catches the ValueError BEFORE any backend round-trip, so the user sees the right reason at the right layer.

Test plan

• Updated `test_read_data_empty_file_raises_with_clear_message` (renamed from `..._returns_nothing` to reflect the new contract) asserts the new fail-fast behavior with `pytest.raises(ValueError, match="Empty CSV file")`
• Existing parser-error path (ragged rows, etc.) still propagates as before
• CI green
• Re-run F8 against rc2 → empty-CSV error surfaces at preflight, no backend call

🤖 Generated with Claude Code

---

Note

Low Risk
Localized CSV read-path behavior change with clearer errors; no auth, security, or broad API surface impact.

Overview
Empty zero-byte CSVs are now treated as hard input errors instead of a successful zero-row read.

CSVIngestor.read_data no longer catches pd.errors.EmptyDataError with a warning and an empty generator. It raises ValueError with a message that names the path and suggests staging / /data/shared/ checks, so users see an input problem at read time rather than a backend HTTP 400 (“No data found for table…”) after an empty table and metadata call.

The test test_read_data_empty_file_raises_with_clear_message replaces the old contract that expected [] and now asserts pytest.raises(ValueError, match="Empty CSV file").

^{Reviewed by Cursor Bugbot for commit 2f37a15. Bugbot is set up for automated code reviews on this repo. Configure here.}

[LukasWodka]

👋 Heads-up — Code review queue is at 14 / 8

Above the WIP limit. The team convention is to review existing PRs before opening new work.

Open PRs currently in Code review (oldest first):

• averaging-service#106 — test(integration): pin XGBoost/LightGBM/CatBoost end-to-end averaging round-trips · author: @LukasWodka · no reviewer assigned
• averaging-service#111 — fix(averaging): correctness round-2 sweep — N5/N7/N8/N9/N10 + load_model except (Sync develop → master for v0.3.0 release #109) · author: @LukasWodka · no reviewer assigned
• client#254 — Sync develop → main for v1.7.1 chart release (egress-enforcement preflight helm test, inert) · author: @saadqbal · no reviewer assigned
• client#256 — fix(egress-proxy): probe reports DNS failure distinctly, not as a TCP connect [image_validator falsely rejects images when target_size matches exactly #104] · author: @saadqbal · no reviewer assigned
• client-runtime#105 — chore(ci): drop unused mysqlclient dep; report test coverage (Phase 0) · author: @LukasWodka · no reviewer assigned
• client-runtime#106 — fix(schema): sync vendored ingest.v1.json — restore token_classification · author: @LukasWodka · no reviewer assigned
• data-ingestors#248 — refactor(P2): extract ingestion-summary renderer into reporting.ConsoleRenderer · author: @LukasWodka · no reviewer assigned
• data-ingestors#249 — fix(database): CHAR(N) must map to a SQLAlchemy type in _get_sqlalchemy_type · author: @divyasinghds · no reviewer assigned
• design-system#47 — feat(table): Table compound organism (Figma 959:6061 + 986:9822) · author: @waqaskhanroghani · no reviewer assigned
• tracebloc-py-package#218 — fix(upload): output-shape validation + hard-fail class-count check for text_classification · author: @shujaatTracebloc · reviewer: @aptracebloc

Pull from review before opening new work. (This is a nudge from the kanban WIP check, not a block.)