agnes-the-ai-analyst/docs/superpowers/plans/2026-04-29-issue-134-bq-access-unify-plan.md
ZdenekSrotyr 83adf01bde
fix(v2): #134 BigQuery cross-project errors return structured 502/400 + BqAccess facade (#138)
* docs(spec): #134 unify BigQuery access behind BqAccess facade

Brainstorm output for issue #134. Captures:
- root cause (incl. correction of the issue's hypothesis about commit 33a9964)
- BqAccess facade API + project resolution rules
- error contract — typed BqAccessError mapped to HTTP 502 for upstream
  BQ failures, 500 for deployment/config bugs
- migration plan for v2_scan, v2_sample, RemoteQueryEngine
- test rewrite eliminating _bq_client_factory injection point
- E2E verification protocol on agnes-development as success criterion

* docs(spec): #134 revise after first review

Incorporates code-reviewer findings:

Must-fix:
- Add v2_schema (2 copies of INSTALL/LOAD/SECRET dance) to migration scope.
- Reframe v2_scan headline: missing try/except around BQ calls is the
  actual cause of bare 500s, not project resolution (which 33a9964 fixed).
- List two more deferred call sites (extractor.py, register_bq_table)
  with explicit rationale.

Important:
- Drop billing != data clause from cross_project_forbidden heuristic;
  rely only on 'serviceusage' substring. billing != data is normal
  for cross-project setup, was over-classifying.
- Split bq_bad_request into _user (400) and _server (502) variants;
  add sql_origin parameter to translate_bq_error so call sites declare
  whether SQL contains user input.
- Add @functools.cache to BqAccess.from_config; document tests bypass
  via dependency_overrides.
- Replace monkey-patched-classmethod test pattern with
  BqAccess(client_factory=...) injection at construction time. Cleaner
  than today's _bq_client_factory and 1:1 migration shape.
- Keep BqProjects.data (reviewer assumed registry has source_project;
  it doesn't). Multi-project explicitly listed as non-goal with note.

Nice-to-have:
- Add 'Implementation strategy' section: 2 staged commits (bug fix
  alone is revertable; refactor follows).
- Extend E2E protocol to cover all three endpoints, not just /sample.
- Note removal of stale docstring at src/remote_query.py:204.

* docs(spec): #134 revision 3 — incorporates second-round review

Must-fix from second review:
- v2_schema split into two migration cases: _fetch_bq_schema translates
  errors via translate_bq_error; _fetch_bq_table_options preserves its
  swallow-all 'except Exception → return {}' so /schema doesn't 502 on
  partition-info failures.
- RemoteQueryEngine.__init__ now resolves BqAccess lazily (in
  _get_bq_client, not in __init__). Without this, ~7 DuckDB-only tests
  in test_remote_query.py would suddenly fail with not_configured.
- translate_bq_error pass-through for BqAccessError is now load-bearing
  (clause 1, before any Google-API branch). bq.client() raises BqAccessError
  for bq_lib_missing/auth_failed; without explicit pass-through those
  fall to 'unknown' and re-raise as bare 500.
- Commit 1 now emits the SAME structured response shape as commit 2 to
  avoid contract churn between commits.
- BIGQUERY_PROJECT env-var precedence is BREAKING for env-only deployments
  — flagged in CHANGELOG ### Changed.

Editorial:
- sql_origin renamed to bad_request_status with values 'client_error' /
  'upstream_error' (clearer about what the parameter actually decides).
  bq_bad_request_user/_server kinds collapsed to bq_bad_request (400)
  and bq_upstream_error (502).
- CLI (cli/commands/query.py) noted as external RemoteQueryEngine caller;
  unaffected because new bq_access kwarg has default None.
- Added unit/integration tests for the new contracts:
  test_translate_passes_through_BqAccessError,
  test_v2_scan_returns_500_on_bq_lib_missing,
  test_v2_schema_returns_200_with_empty_partition_on_bq_failure,
  test_resolve_succeeds_after_config_set.
- E2E protocol now covers /schema as the fourth endpoint.
- Documented functools.cache-doesn't-cache-exceptions semantics and
  fixture nullcontext-doesn't-close caveat for nested sessions.

* docs(spec): #134 revision 4 — incorporates third-round review

Third reviewer verdict: 'implementation-ready with two trivial edits';
explicitly noted prior rounds did the heavy lifting.

Edits:
1. get_bq_access() module-level function instead of @classmethod
   @functools.cache from_config. Removes the classmethod-cache stacking
   footgun (different Python versions wrap differently) and gives FastAPI's
   dependency introspection a clean function signature. Drops the
   'Do not subclass BqAccess' caveat that no longer applies.

2. Commit 1 strategy explicitly: wrap _fetch_bq_sample (v2_sample),
   _bq_dry_run_bytes + _run_bq_scan (v2_scan), and _fetch_bq_schema
   (v2_schema strict block). Do NOT touch _fetch_bq_table_options swallow-all
   in commit 1 — preserved as-is, then migrated (still preserved) in commit 2.
   All three endpoints emit the same structured body shape so client parsers
   see one consistent contract throughout the staged rollout. No more
   half-rolled-out window where /sample is bare 500 while /scan is
   structured 502.

* docs(plan): #134 implementation plan — Phase 1 (atomic bug fix) + Phase 2 (BqAccess refactor) + Phase 3 (verification)

Bite-sized TDD tasks. 3 phases, 16 tasks total:

Phase 1 (Commit 1) — atomic bug fix across all four v2 endpoints:
  Tasks 1.1-1.5 wrap _fetch_bq_sample, _bq_dry_run_bytes, _run_bq_scan,
  _fetch_bq_schema with structured 502/400 try/except. _fetch_bq_table_options
  preserved untouched. CHANGELOG Fixed entries.

Phase 2 (Commit 2) — BqAccess facade extraction + migration:
  Tasks 2.1-2.5 build connectors/bigquery/access.py bottom-up
  (BqProjects, BqAccessError, translate_bq_error, default factories,
  BqAccess class, get_bq_access module-level cached). Task 2.6 adds
  conftest.py fixture. Tasks 2.7-2.9 migrate v2_scan, v2_sample, v2_schema
  to BqAccess. Tasks 2.10-2.11 migrate RemoteQueryEngine + tests
  (lazy bq_access, drop _bq_client_factory). Task 2.12 CHANGELOG
  Changed BREAKING + Internal.

Phase 3 — Verification:
  3.1 full pytest. 3.2 squash into two PR-shape commits. 3.3 manual
  E2E on agnes-development per spec protocol → close #134.

Self-review table maps spec sections to implementing tasks; no gaps.

* fix(v2): #134 structured 502/400 on BQ errors across /scan, /scan/estimate, /sample, /schema

Wraps the BigQuery call sites in v2_scan, v2_sample, and v2_schema (strict
block only) with try/except for google.api_core exceptions, translating to
HTTPException with a structured body shape: {error, message, details}.

Fixes Pavel's report (#134) where these endpoints returned bare HTTP 500
with no body when the SA on agnes-development hit cross-project Forbidden
on serviceusage.services.use.

Also fixes /sample's missing billing_project fallback (the bug 33a9964
fixed for /scan never landed here).

Status code split:
  - /scan, /scan/estimate: BadRequest -> 400 (bq_bad_request) since SQL is
    user-derived from req.select/where/order_by.
  - /sample, /schema: BadRequest -> 502 (bq_upstream_error) since SQL is
    server-constructed from validated identifiers.
  - All Forbidden -> 502 with cross_project_forbidden if 'serviceusage' in
    error message (with hint pointing at data_source.bigquery.billing_project),
    else bq_forbidden.

Body shape matches what the upcoming BqAccess refactor (next commit) will
produce, so client-side parsers see one consistent contract throughout
the staged rollout.

_fetch_bq_table_options preserved exactly as-is — its swallow-all-and-return-empty
contract is intentional and survives into the refactor; /schema continues to
return 200 with empty partition info when partition queries fail.

Outer wraps in scan_endpoint, scan_estimate_endpoint, sample, and schema
endpoints exist only to make the test pattern (monkeypatching whole
_fetch_* functions) work, and are tagged TODO(#134 Phase 2) for removal
once BqAccess centralizes translation.

* refactor(bq): #134 BqAccess facade — unify v2_scan, v2_sample, v2_schema, RemoteQueryEngine

Extracts the duplicated BigQuery-access pattern (project resolution +
client construction + DuckDB-extension session + Google-API error
translation) into connectors/bigquery/access.py. Migrates four
call sites to use it:

- app/api/v2_scan.py — _bq_dry_run_bytes, _run_bq_scan
- app/api/v2_sample.py — _fetch_bq_sample
- app/api/v2_schema.py — _fetch_bq_schema (strict translation),
  _fetch_bq_table_options (preserves swallow-all best-effort contract)
- src/remote_query.py — RemoteQueryEngine, lazy bq_access kwarg

The new module exposes:
- BqProjects (frozen dataclass: billing + data project IDs)
- BqAccessError (typed exception with HTTP_STATUS class mapping)
- BqAccess (facade with injectable client_factory/duckdb_session_factory
  for tests; defaults call the real google-cloud-bigquery + DuckDB extension)
- get_bq_access (module-level @functools.cache; FastAPI Depends target)
- translate_bq_error (Google API exception → BqAccessError mapper, with
  BqAccessError pass-through, 'serviceusage'-substring heuristic for
  cross_project_forbidden, and bad_request_status param distinguishing
  user-derived (400) from server-constructed (502) SQL)
- _default_client_factory, _default_duckdb_session_factory

RemoteQueryEngine.__init__ no longer accepts _bq_client_factory; tests
migrate to bq_access=BqAccess(projects, client_factory=...). DuckDB-only
RemoteQueryEngine tests need no changes — bq_access defaults to None and
get_bq_access() is only invoked on first BQ call (lazy resolution).
BqAccessError raised internally is translated to RemoteQueryError(
error_type="bq_error") in _get_bq_client to preserve the engine's
existing public contract — CLI and /api/query/hybrid callers see no change.

Endpoint tests (test_v2_scan, test_v2_scan_estimate, test_v2_sample,
test_v2_schema) migrate from monkey-patching whole _fetch_* functions
to using the new bq_access fixture in tests/conftest.py — which
exercises the REAL translation path through BqAccess + translate_bq_error,
closing the test gap flagged in Task 1.1's review.

Side-effect behavior change: v2_sample's FROM clause now uses the data
project (instance.yaml data_source.bigquery.project), not the conflated
billing_project from Phase 1. Documented in CHANGELOG ### Internal.

BREAKING for deployments combining BIGQUERY_PROJECT env var with
data_source.bigquery.project in instance.yaml — env var now overrides
data project too. See CHANGELOG ### Changed.

Two known-duplicate BQ-access sites (connectors/bigquery/extractor.py,
scripts/duckdb_manager.register_bq_table) explicitly out of scope;
tracked as follow-up.

Removed stale docstring at the previous src/remote_query.py:204
that referenced scripts.duckdb_manager._create_bq_client as the default
BQ client factory (RemoteQueryEngine never actually used that function).

Test counts: tests/test_bq_access.py +27 (new), tests/test_v2_*.py +
tests/test_remote_query.py migrated to bq_access fixture (counts unchanged
or +1-2 per file). Full suite: 2086 passed, 8 pre-existing failures
(DB migration tests with unrelated internal_roles DependencyException —
not introduced by this PR).

* fix(bq_access): translate DefaultCredentialsError to BqAccessError(auth_failed)

CI on PR #138 caught: bigquery.Client(...) resolves Application Default
Credentials at construction time; without ADC (CI without SA key, dev
laptop without 'gcloud auth application-default login') it raises
google.auth.exceptions.DefaultCredentialsError synchronously.

Pre-fix _default_client_factory only caught ImportError, so DefaultCredentialsError
propagated as raw exception — and from production endpoints would surface
as bare 500 (the exact failure mode #134 sets out to fix).

Now translates to BqAccessError(kind='auth_failed', details.hint='Run
gcloud auth application-default login...'). Endpoint catch chain returns
HTTP 502 with structured body. Adds unit test
test_raises_auth_failed_on_default_credentials_error.

Third-round spec review flagged this case in passing; the fix didn't land.
CI's auth-less environment surfaced it.

* fix(bq_access): get_bq_access() returns sentinel instead of raising when not configured

Devin BUG_0001 on PR #138 review: 'get_bq_access() as FastAPI Depends
breaks all v2 endpoints for non-BigQuery instances'.

Pre-fix: get_bq_access() raised BqAccessError(not_configured) when
neither BIGQUERY_PROJECT env nor data_source.bigquery.project was set.
Because FastAPI resolves Depends() BEFORE the endpoint body runs, this
exception fires during dep-injection — the endpoint's try/except
BqAccessError clause never gets a chance to catch it. Result: every
v2 request on Keboola-only or CSV-only instances returned bare HTTP
500, even for local-source tables that never touch BigQuery.

Fix: get_bq_access() now returns a sentinel BqAccess with empty
BqProjects and factories that raise BqAccessError(not_configured)
on actual use. Construction succeeds, FastAPI's dep-injection cleanly
yields the sentinel, the endpoint runs. The local-source code path
in build_sample / build_schema / etc. never calls bq.client() or
bq.duckdb_session() (it reads parquet directly), so non-BQ tables
return 200 as before. Only when an endpoint actually tries to query
BQ (source_type == 'bigquery') does the sentinel raise — and the
endpoint's existing except BqAccessError catches it normally,
returning structured 502 with hint.

Test get_bq_access::test_raises_not_configured_when_neither_set
renamed and rewritten to test_returns_sentinel_when_neither_set:
asserts BqAccess is returned, then asserts client() and
duckdb_session() each raise BqAccessError(not_configured) on call.

Test test_does_not_cache_exceptions removed (no longer applicable)
and replaced with test_sentinel_is_cached_per_process documenting
the operator-restart-on-config-change contract.

* docs(spec+plan): #134 genericize customer-specific tokens (CLAUDE.md OSS rule)

Devin BUG_0001/0002 round 3 on PR #138: spec and plan docs contained
customer-specific deployment hostnames, deployment names, and a GCP
project ID that violated CLAUDE.md's vendor-agnostic OSS rule
('Nothing customer-specific belongs in code, configuration defaults,
comments, docs, commit messages, PR titles, or PR bodies').

Replacements:
  agnes-development.groupondev.com -> <your-agnes-host>
  agnes-development                -> <your-dev-instance>
  prj-grp-dataview-prod-1ff9       -> <your-data-project>
  s1_session_landings              -> <bq_table_id>

E2E verification semantics unchanged — operators still run the same
four curls + config flip + retry, just substituting their own host /
deployment name / project / table.

* fix(bq_access): hook get_bq_access.cache_clear into instance_config.reset_cache

Devin ANALYSIS_0004 on PR #138: get_bq_access is @functools.cache'd at
process level, so it captures BigQuery project IDs at first call and
ignores subsequent instance.yaml changes. Pre-Phase-2 the v2 endpoints
re-read get_value() on every request, so admin /api/admin/server-config
saves (which call instance_config.reset_cache()) hot-reloaded the BQ
project. Without this fix, my refactor silently regresses that contract
— operators editing instance.yaml via the admin UI would see no effect
on v2 endpoints until container restart.

instance_config.reset_cache() now also calls
connectors.bigquery.access.get_bq_access.cache_clear() (lazy import,
swallowed if connectors module isn't loaded — keeps instance_config
usable in isolated unit tests).

Adds test_instance_config_reset_cache_invalidates_get_bq_access as
regression guard. Updates CHANGELOG Internal entry to mention the
hot-reload contract + the not-configured sentinel behavior (round-3
fix from Devin BUG_0001 was previously only in commit message).

* fix(bq_access): surface not_configured before identifier validation + plan path genericize

Devin BUG_0001 + BUG_0002 round 5 on PR #138.

BUG_0001 (plan doc): personal filesystem path violated CLAUDE.md
vendor-agnostic rule. Replaced with '<worktree-root>' placeholder.

BUG_0002 (sentinel error path): when get_bq_access() returns the sentinel
BqAccess (BQ not configured), the empty bq.projects.data was reaching
validate_quoted_identifier first and raising ValueError -> endpoint
mapped to HTTP 400 'unsafe_identifier' instead of structured 500
'not_configured' with hint.

Each fetch helper now checks 'if not bq.projects.data: bq.client()' as
the first step, which triggers the sentinel's BqAccessError(not_configured).
Endpoint catches the typed error and returns HTTP 500 with hint pointing
at data_source.bigquery.project. Best-effort _fetch_bq_table_options
returns {} silently in this case (preserves the swallow-all contract).

* fix(bq_access): classify DuckDB-native exceptions from bigquery_query() via string match

Devin ANALYSIS on PR #138 review (latest round). The DuckDB bigquery
extension is a C++ plugin making its own HTTP calls — when BQ returns
403, it throws duckdb.IOException with the BQ error embedded as text,
not gax.Forbidden. translate_bq_error's isinstance checks would miss
these, falling to case 7 → bare 500 in production for v2_scan, v2_sample,
and v2_schema (the bigquery_query() paths).

Fix: last-resort string-match heuristic before the re-raise. 'Forbidden'
/ '403' / 'Bad Request' / '400' in the lowercased message classifies via
the same kind hierarchy. The 'serviceusage' substring still distinguishes
cross_project_forbidden from bq_forbidden. Specific enough that random
exceptions without HTTP-error keywords still re-raise.

Adds 4 unit tests covering the new heuristic + the 'don't swallow random
exceptions' invariant.

* chore(release): cut 0.22.0

PR #138 contains issue #134 user-visible behavior changes:
- BREAKING: BIGQUERY_PROJECT env var now overrides instance.yaml
  data_source.bigquery.project for v2 endpoints (previously
  RemoteQueryEngine billing only).
- Fixed: structured 502/400 on /api/v2/sample, /scan, /scan/estimate,
  /schema when BigQuery raises Forbidden/BadRequest (was bare 500).
- Internal: BqAccess facade refactor unifying four duplicate BQ-access
  call sites; instance_config.reset_cache() now invalidates BqAccess
  cache too so admin server-config saves hot-reload BQ project IDs.

Bumps to 0.22.0 because PR #137 merged first and took 0.21.0.
2026-04-30 10:11:20 +02:00

2174 lines
No EOL
84 KiB
Markdown

# Issue #134 — Unify BigQuery access behind `BqAccess` (implementation plan)
> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
**Goal:** Fix issue #134 (v2 BigQuery endpoints returning bare HTTP 500) by (a) translating Google API errors into structured 502/400 responses and (b) extracting the duplicated BQ-access pattern into a single `BqAccess` facade in `connectors/bigquery/access.py`.
**Architecture:** Two-phase rollout. Phase 1 = inline `try/except` in three v2 endpoint files emitting the final structured body shape — fixes Pavel's bug atomically. Phase 2 = extract `BqAccess` facade + migrate all four call sites (`v2_scan`, `v2_sample`, `v2_schema` two blocks, `RemoteQueryEngine`) to it; delete `_bq_client_factory` test injection point. Final PR has two reviewable commits.
**Tech Stack:** Python 3.11, FastAPI, DuckDB, google-cloud-bigquery, pytest. See `docs/superpowers/specs/2026-04-29-issue-134-bq-access-unify-design.md` for the full design.
---
## File structure
**Created (Phase 2):**
- `connectors/bigquery/access.py``BqProjects`, `BqAccessError`, `BqAccess`, `get_bq_access`, `translate_bq_error`, default factories.
- `tests/test_bq_access.py` — unit tests for the new module.
**Modified:**
- `app/api/v2_sample.py` — Phase 1: billing_project + try/except; Phase 2: replace with `BqAccess`.
- `app/api/v2_scan.py` — Phase 1: try/except around BQ calls; Phase 2: replace with `BqAccess`.
- `app/api/v2_schema.py` — Phase 1: try/except around `_fetch_bq_schema` only; Phase 2: replace with `BqAccess` (both blocks, preserve `_fetch_bq_table_options` swallow-all).
- `src/remote_query.py` — Phase 2: lazy `bq_access` kwarg; drop `_bq_client_factory` + fallback chain + stale docstring.
- `tests/test_v2_scan.py`, `tests/test_v2_scan_estimate.py`, `tests/test_v2_sample.py`, `tests/test_v2_schema.py` — add structured-error tests; in Phase 2 update the `_fetch_*` lambda signatures.
- `tests/test_remote_query.py` — Phase 2: migrate `_bq_client_factory=` injection to `bq_access=BqAccess(..., client_factory=...)`.
- `tests/conftest.py` — Phase 2: add `bq_access` fixture.
- `CHANGELOG.md` — Phase 1 (Fixed); Phase 2 (Changed BREAKING + Internal).
---
# Phase 1 — Commit 1: Atomic bug fix (no `BqAccess` yet)
Goal: structured-body responses for cross-project Forbidden / BadRequest from BQ on all three v2 endpoints. Same body shape that Phase 2 will produce, so client-side parsers (CLI, UI) see one consistent contract throughout the rollout.
**Body shape (used in all Phase 1 tasks):**
```python
# 502 case (Forbidden / upstream BadRequest)
{"error": "<kind>", "message": "<bq error text>", "details": {<contextual>}}
# 400 case (user-derived BadRequest in v2_scan only)
{"error": "bq_bad_request", "message": "<bq error text>", "details": {}}
```
`<kind>` for Forbidden = `"cross_project_forbidden"` if `"serviceusage" in str(e).lower()` else `"bq_forbidden"`.
---
### Task 1.1: v2_sample — add billing_project fallback + wrap `_fetch_bq_sample`
**Files:**
- Modify: `app/api/v2_sample.py:21-48` (`_fetch_bq_sample`), `:97-110` (`sample` endpoint).
- Test: `tests/test_v2_sample.py` (add new tests).
- [ ] **Step 1: Read current state**
Run: `head -120 app/api/v2_sample.py` to confirm baseline (today: reads only `data_source.bigquery.project`; `_fetch_bq_sample` has no try/except).
- [ ] **Step 2: Write failing test for cross-project Forbidden → 502 + structured body**
Add to `tests/test_v2_sample.py`:
```python
def test_sample_returns_502_on_bq_forbidden_serviceusage(reload_db, monkeypatch):
from app.api import v2_sample
from google.api_core.exceptions import Forbidden
def _raise_forbidden(*args, **kwargs):
raise Forbidden("Permission denied: serviceusage.services.use on project foo")
monkeypatch.setattr(v2_sample, "_fetch_bq_sample", _raise_forbidden)
# ... seed registry with a bigquery table, then call the endpoint
# via TestClient (match the pattern used by other tests in this file)
# Assert response.status_code == 502
# Assert response.json()["detail"]["error"] == "cross_project_forbidden"
# Assert "billing_project" in response.json()["detail"]["details"]["hint"].lower()
```
NOTE: match the test-client construction style of existing tests in this file. If they use a `TestClient` fixture, reuse it; if they call helpers directly, follow that.
- [ ] **Step 3: Write failing test for billing_project fallback being read**
```python
def test_sample_reads_billing_project_from_instance_yaml(reload_db, monkeypatch):
"""Regression guard for the original bug: the project passed to _fetch_bq_sample
must come from billing_project when set, not from project."""
captured = {}
def _capture(project, dataset, table, n):
captured["project"] = project
return []
monkeypatch.setattr("app.api.v2_sample._fetch_bq_sample", _capture)
monkeypatch.setattr("app.instance_config.get_value", lambda *keys, **kw: {
("data_source", "bigquery", "project"): "data-proj",
("data_source", "bigquery", "billing_project"): "billing-proj",
}.get(keys, kw.get("default", "")))
# ... seed bigquery table, call /api/v2/sample/<id>, assert 200
assert captured["project"] == "billing-proj"
```
- [ ] **Step 4: Run tests — expect FAIL**
Run: `pytest tests/test_v2_sample.py::test_sample_returns_502_on_bq_forbidden_serviceusage tests/test_v2_sample.py::test_sample_reads_billing_project_from_instance_yaml -v`
Expected: both FAIL — sample endpoint today returns 500 on Forbidden and reads only `project`.
- [ ] **Step 5: Implement billing_project fallback**
Edit `app/api/v2_sample.py:104` (in the `sample` endpoint):
```python
project_id = (
get_value("data_source", "bigquery", "billing_project", default="")
or get_value("data_source", "bigquery", "project", default="")
or ""
)
```
(Replace today's `project_id = get_value("data_source", "bigquery", "project", default="") or ""`.)
- [ ] **Step 6: Wrap `_fetch_bq_sample` with structured-error translation**
Edit `app/api/v2_sample.py:21-48`. Replace the function body's interior with:
```python
def _fetch_bq_sample(project: str, dataset: str, table: str, n: int) -> list[dict]:
import duckdb
from google.api_core import exceptions as gax
from connectors.bigquery.auth import get_metadata_token
from src.identifier_validation import validate_quoted_identifier
from fastapi import HTTPException
if not (validate_quoted_identifier(project, "BQ project")
and validate_quoted_identifier(dataset, "BQ dataset")
and validate_quoted_identifier(table, "BQ source_table")):
raise ValueError("unsafe BQ identifier in registry — refusing to query")
token = get_metadata_token()
conn = duckdb.connect(":memory:")
try:
conn.execute("INSTALL bigquery FROM community; LOAD bigquery;")
escaped = token.replace("'", "''")
conn.execute(f"CREATE OR REPLACE SECRET bq_s (TYPE bigquery, ACCESS_TOKEN '{escaped}')")
bq_sql = f"SELECT * FROM `{project}.{dataset}.{table}` LIMIT {int(n)}"
try:
df = conn.execute(
"SELECT * FROM bigquery_query(?, ?)",
[project, bq_sql],
).fetchdf()
except gax.Forbidden as e:
kind = "cross_project_forbidden" if "serviceusage" in str(e).lower() else "bq_forbidden"
raise HTTPException(
status_code=502,
detail={
"error": kind,
"message": str(e),
"details": {
"billing_project": project,
"hint": (
"Set data_source.bigquery.billing_project in instance.yaml to a project "
"where the SA has serviceusage.services.use, or grant the SA that role "
"on the data project."
) if kind == "cross_project_forbidden" else "",
},
},
)
except gax.BadRequest as e:
# /sample SQL is server-constructed (validated identifiers + LIMIT n);
# a BadRequest here means registry corruption → upstream error, not user fault.
raise HTTPException(
status_code=502,
detail={"error": "bq_upstream_error", "message": str(e), "details": {}},
)
except gax.GoogleAPICallError as e:
raise HTTPException(
status_code=502,
detail={"error": "bq_upstream_error", "message": str(e), "details": {}},
)
return df.to_dict(orient="records")
finally:
conn.close()
```
- [ ] **Step 7: Run the two new tests — expect PASS**
Run: `pytest tests/test_v2_sample.py::test_sample_returns_502_on_bq_forbidden_serviceusage tests/test_v2_sample.py::test_sample_reads_billing_project_from_instance_yaml -v`
Expected: both PASS.
- [ ] **Step 8: Run full v2_sample test file — expect no regressions**
Run: `pytest tests/test_v2_sample.py -v`
Expected: all green (existing tests unaffected — the change is additive).
- [ ] **Step 9: Commit (intermediate, will be squashed before PR)**
```bash
cd <worktree-root>
git add app/api/v2_sample.py tests/test_v2_sample.py
git commit -m "fix(v2_sample): #134 add billing_project fallback + structured 502 on BQ Forbidden"
```
---
### Task 1.2: v2_scan — wrap `_bq_dry_run_bytes`
**Files:**
- Modify: `app/api/v2_scan.py:43-55` (`_bq_dry_run_bytes`).
- Test: `tests/test_v2_scan_estimate.py`.
- [ ] **Step 1: Read current state**
Run: `sed -n '43,55p' app/api/v2_scan.py` to confirm: today no try/except.
- [ ] **Step 2: Write failing test for /scan/estimate 502 on Forbidden**
Add to `tests/test_v2_scan_estimate.py` (match the file's existing test style):
```python
def test_scan_estimate_returns_502_on_bq_forbidden_serviceusage(reload_db, monkeypatch):
from app.api import v2_scan
from google.api_core.exceptions import Forbidden
def _raise_forbidden(project, sql):
raise Forbidden("Permission denied: serviceusage.services.use on project foo")
monkeypatch.setattr(v2_scan, "_bq_dry_run_bytes", _raise_forbidden)
# ... seed bigquery table, call POST /api/v2/scan/estimate
# Assert response.status_code == 502
# Assert response.json()["detail"]["error"] == "cross_project_forbidden"
# Assert "hint" in response.json()["detail"]["details"]
```
```python
def test_scan_estimate_returns_400_on_bq_bad_request(reload_db, monkeypatch):
from app.api import v2_scan
from google.api_core.exceptions import BadRequest
def _raise_bad_request(project, sql):
raise BadRequest("Syntax error: unexpected token at line 1, column 5")
monkeypatch.setattr(v2_scan, "_bq_dry_run_bytes", _raise_bad_request)
# ... call POST /api/v2/scan/estimate
# Assert response.status_code == 400
# Assert response.json()["detail"]["error"] == "bq_bad_request"
# Assert "Syntax error" in response.json()["detail"]["message"]
```
- [ ] **Step 3: Run tests — expect FAIL**
Run: `pytest tests/test_v2_scan_estimate.py::test_scan_estimate_returns_502_on_bq_forbidden_serviceusage tests/test_v2_scan_estimate.py::test_scan_estimate_returns_400_on_bq_bad_request -v`
Expected: both FAIL with bare 500.
- [ ] **Step 4: Wrap `_bq_dry_run_bytes`**
Replace `app/api/v2_scan.py:43-55` with:
```python
def _bq_dry_run_bytes(project: str, sql: str) -> int:
"""Run a BQ dry-run via the google-cloud-bigquery client and return totalBytesProcessed.
Errors translated to HTTPException with the structured-body shape used across
all v2 endpoints. SQL here is user-derived (built from req.select/where/order_by),
so BadRequest → 400.
"""
from google.cloud import bigquery
from google.api_core import exceptions as gax
from google.api_core.client_options import ClientOptions
from fastapi import HTTPException
client = bigquery.Client(
project=project,
client_options=ClientOptions(quota_project_id=project),
)
try:
job = client.query(
sql,
job_config=bigquery.QueryJobConfig(dry_run=True, use_query_cache=False),
)
return int(job.total_bytes_processed or 0)
except gax.Forbidden as e:
kind = "cross_project_forbidden" if "serviceusage" in str(e).lower() else "bq_forbidden"
raise HTTPException(
status_code=502,
detail={
"error": kind,
"message": str(e),
"details": {
"billing_project": project,
"hint": (
"Set data_source.bigquery.billing_project in instance.yaml to a project "
"where the SA has serviceusage.services.use, or grant the SA that role "
"on the data project."
) if kind == "cross_project_forbidden" else "",
},
},
)
except gax.BadRequest as e:
raise HTTPException(
status_code=400,
detail={"error": "bq_bad_request", "message": str(e), "details": {}},
)
except gax.GoogleAPICallError as e:
raise HTTPException(
status_code=502,
detail={"error": "bq_upstream_error", "message": str(e), "details": {}},
)
```
- [ ] **Step 5: Run tests — expect PASS**
Run: `pytest tests/test_v2_scan_estimate.py -v`
Expected: all green.
- [ ] **Step 6: Commit**
```bash
git add app/api/v2_scan.py tests/test_v2_scan_estimate.py
git commit -m "fix(v2_scan): #134 structured 502/400 on BQ errors in dry-run path"
```
---
### Task 1.3: v2_scan — wrap `_run_bq_scan`
**Files:**
- Modify: `app/api/v2_scan.py:266-282` (`_run_bq_scan`).
- Test: `tests/test_v2_scan.py`.
- [ ] **Step 1: Write failing tests (mirror Task 1.2)**
Add to `tests/test_v2_scan.py`:
```python
def test_scan_returns_502_on_bq_forbidden_serviceusage(reload_db, monkeypatch):
from app.api import v2_scan
from google.api_core.exceptions import Forbidden
def _raise(project, sql):
raise Forbidden("Permission denied: serviceusage.services.use on project foo")
monkeypatch.setattr(v2_scan, "_run_bq_scan", _raise)
# ... seed bigquery table, call POST /api/v2/scan
# Assert response.status_code == 502
# Assert response.json()["detail"]["error"] == "cross_project_forbidden"
def test_scan_returns_400_on_bq_bad_request(reload_db, monkeypatch):
from app.api import v2_scan
from google.api_core.exceptions import BadRequest
def _raise(project, sql):
raise BadRequest("Syntax error")
monkeypatch.setattr(v2_scan, "_run_bq_scan", _raise)
# ... call POST /api/v2/scan
# Assert response.status_code == 400
# Assert response.json()["detail"]["error"] == "bq_bad_request"
```
- [ ] **Step 2: Run — expect FAIL**
Run: `pytest tests/test_v2_scan.py::test_scan_returns_502_on_bq_forbidden_serviceusage tests/test_v2_scan.py::test_scan_returns_400_on_bq_bad_request -v`
Expected: FAIL with bare 500.
- [ ] **Step 3: Wrap `_run_bq_scan`**
Replace `app/api/v2_scan.py:266-282` with:
```python
def _run_bq_scan(project: str, sql: str):
"""Run a BQ query via DuckDB BQ extension. Returns Arrow table.
Errors translated to HTTPException with the structured-body shape used across
all v2 endpoints. SQL here is user-derived → BadRequest → 400.
"""
import duckdb
from google.api_core import exceptions as gax
from connectors.bigquery.auth import get_metadata_token
from fastapi import HTTPException
token = get_metadata_token()
conn = duckdb.connect(":memory:")
try:
conn.execute("INSTALL bigquery FROM community; LOAD bigquery;")
escaped = token.replace("'", "''")
conn.execute(f"CREATE OR REPLACE SECRET bq_s (TYPE bigquery, ACCESS_TOKEN '{escaped}')")
try:
return conn.execute(
"SELECT * FROM bigquery_query(?, ?)",
[project, sql],
).arrow()
except gax.Forbidden as e:
kind = "cross_project_forbidden" if "serviceusage" in str(e).lower() else "bq_forbidden"
raise HTTPException(
status_code=502,
detail={
"error": kind,
"message": str(e),
"details": {
"billing_project": project,
"hint": (
"Set data_source.bigquery.billing_project in instance.yaml to a project "
"where the SA has serviceusage.services.use, or grant the SA that role "
"on the data project."
) if kind == "cross_project_forbidden" else "",
},
},
)
except gax.BadRequest as e:
raise HTTPException(
status_code=400,
detail={"error": "bq_bad_request", "message": str(e), "details": {}},
)
except gax.GoogleAPICallError as e:
raise HTTPException(
status_code=502,
detail={"error": "bq_upstream_error", "message": str(e), "details": {}},
)
finally:
conn.close()
```
- [ ] **Step 4: Run — expect PASS**
Run: `pytest tests/test_v2_scan.py -v`
Expected: all green.
- [ ] **Step 5: Commit**
```bash
git add app/api/v2_scan.py tests/test_v2_scan.py
git commit -m "fix(v2_scan): #134 structured 502/400 on BQ errors in scan path"
```
---
### Task 1.4: v2_schema — wrap `_fetch_bq_schema` (strict block only)
**Files:**
- Modify: `app/api/v2_schema.py:36-73` (`_fetch_bq_schema` function — exact line range may vary; the strict block is the FIRST one in the file, NOT `_fetch_bq_table_options`).
- Test: `tests/test_v2_schema.py`.
**IMPORTANT — preserve `_fetch_bq_table_options` exactly as-is in this task.** That function (lines 90-129) wraps everything in `try/except Exception → return {}` and is the best-effort partition-info path. Phase 2 will migrate it to use `BqAccess` while preserving the swallow-all. **Do not touch it in Phase 1.**
- [ ] **Step 1: Write failing test for /schema 502 on _fetch_bq_schema Forbidden**
Add to `tests/test_v2_schema.py`:
```python
def test_schema_returns_502_on_bq_forbidden_serviceusage(reload_db, monkeypatch):
from app.api import v2_schema
from google.api_core.exceptions import Forbidden
def _raise(project, dataset, table):
raise Forbidden("Permission denied: serviceusage.services.use on project foo")
monkeypatch.setattr(v2_schema, "_fetch_bq_schema", _raise)
# ... seed bigquery table, call GET /api/v2/schema/<id>
# Assert response.status_code == 502
# Assert response.json()["detail"]["error"] == "cross_project_forbidden"
```
- [ ] **Step 2: Write regression-guard test for /schema 200 with empty partition info on `_fetch_bq_table_options` failure**
```python
def test_schema_returns_200_with_empty_partition_on_table_options_failure(reload_db, monkeypatch):
"""Regression guard: _fetch_bq_table_options is best-effort. /schema must keep
returning successfully even if partition-info fetch fails."""
from app.api import v2_schema
from google.api_core.exceptions import Forbidden
def _ok_schema(project, dataset, table):
return [{"name": "event_date", "type": "DATE"}]
def _fail_options(project, dataset, table):
raise Forbidden("denied")
monkeypatch.setattr(v2_schema, "_fetch_bq_schema", _ok_schema)
monkeypatch.setattr(v2_schema, "_fetch_bq_table_options", _fail_options)
# ... seed bigquery table, call GET /api/v2/schema/<id>
# Assert response.status_code == 200
# Assert response.json()["columns"] == [{"name": "event_date", "type": "DATE"}]
# Assert "partition_by" not in response.json() OR response.json().get("partition_by") is None
```
(NOTE: today `_fetch_bq_table_options` already has the swallow-all `try/except → return {}`, so this test should pass even without changes — it's a regression guard for Phase 2.)
- [ ] **Step 3: Run — expect first test FAIL, second PASS**
Run: `pytest tests/test_v2_schema.py::test_schema_returns_502_on_bq_forbidden_serviceusage tests/test_v2_schema.py::test_schema_returns_200_with_empty_partition_on_table_options_failure -v`
Expected: 502 test FAILS (today returns 500); regression-guard PASSES.
- [ ] **Step 4: Wrap `_fetch_bq_schema` with the same try/except shape**
Edit `app/api/v2_schema.py:36-73` (the FIRST INSTALL/LOAD/SECRET block, in `_fetch_bq_schema`). Wrap the `conn.execute("SELECT * FROM bigquery_query(?, ?, ?)", ...)` line in:
```python
from google.api_core import exceptions as gax
from fastapi import HTTPException
try:
rows = conn.execute(
"SELECT * FROM bigquery_query(?, ?, ?)",
[project, bq_sql, table],
).fetchall()
except gax.Forbidden as e:
kind = "cross_project_forbidden" if "serviceusage" in str(e).lower() else "bq_forbidden"
raise HTTPException(
status_code=502,
detail={
"error": kind,
"message": str(e),
"details": {
"billing_project": project,
"hint": (
"Set data_source.bigquery.billing_project in instance.yaml to a project "
"where the SA has serviceusage.services.use, or grant the SA that role "
"on the data project."
) if kind == "cross_project_forbidden" else "",
},
},
)
except gax.GoogleAPICallError as e:
raise HTTPException(
status_code=502,
detail={"error": "bq_upstream_error", "message": str(e), "details": {}},
)
```
(`/schema` SQL hits INFORMATION_SCHEMA, fully server-constructed → BadRequest folded into upstream_error 502; no separate 400 case.)
**Do not modify `_fetch_bq_table_options`.**
- [ ] **Step 5: Run — expect PASS**
Run: `pytest tests/test_v2_schema.py -v`
Expected: all green.
- [ ] **Step 6: Commit**
```bash
git add app/api/v2_schema.py tests/test_v2_schema.py
git commit -m "fix(v2_schema): #134 structured 502 on BQ errors in strict schema path"
```
---
### Task 1.5: CHANGELOG entry for Phase 1; finalize commit 1
**Files:** `CHANGELOG.md`.
- [ ] **Step 1: Read current state**
Run: `head -20 CHANGELOG.md` to confirm there is an `## [Unreleased]` section. If not, create one above the topmost release version.
- [ ] **Step 2: Add Fixed entries under `## [Unreleased] / ### Fixed`**
```markdown
### Fixed
- v2 `/sample` endpoint: BigQuery cross-project queries now respect `data_source.bigquery.billing_project` from `instance.yaml` (mirrors v2 `/scan` fix from `33a9964`). Closes #134 for `/sample`.
- v2 `/scan`, `/scan/estimate`, `/sample`, `/schema`: BigQuery upstream errors no longer return bare HTTP 500 with empty body. `Forbidden` from BQ now returns HTTP 502 with structured JSON body (`{"error": "cross_project_forbidden", "message": "...", "details": {"hint": "..."}}`); user-derived `BadRequest` on `/scan*` returns HTTP 400 with `kind=bq_bad_request`. Closes #134.
- v2 `/schema`: best-effort partition-info path (`_fetch_bq_table_options`) preserves its swallow-all behavior unchanged; `/schema` still returns 200 with empty partition info when BQ partition queries fail.
```
- [ ] **Step 3: Stage CHANGELOG and verify Phase 1 working tree is clean**
Run: `git add CHANGELOG.md && git status`
Expected: only `CHANGELOG.md` staged; nothing else dirty.
- [ ] **Step 4: Run full test suite to confirm no regressions before sealing Phase 1**
Run: `pytest tests/ -v -x --ignore=tests/test_telegram --ignore=tests/test_ws_gateway 2>&1 | tail -60`
Expected: all green (or only pre-existing failures unrelated to BQ/v2).
If any new failures appear, fix them before committing — Phase 1 must ship green.
- [ ] **Step 5: Squash Phase 1 commits (intermediate WIP) into a single Phase-1 commit**
```bash
# Identify the first commit in this phase (it has "fix(v2_sample): #134 add billing_project")
git log --oneline | grep "#134" | tail -1
# Note the commit BEFORE that one — call it $BASE
# Soft-reset to $BASE, then make one clean commit
git reset --soft $BASE
git commit -m "fix(v2): #134 structured 502/400 on BQ errors across /scan, /scan/estimate, /sample, /schema
Wraps the BigQuery call sites in v2_scan, v2_sample, and v2_schema (strict
block only) with try/except for google.api_core exceptions, translating to
HTTPException with a structured body shape: {error, message, details}.
Fixes Pavel's report (#134) where these endpoints returned bare HTTP 500
with no body when the SA on <your-dev-instance> hit cross-project Forbidden
on serviceusage.services.use.
Also fixes /sample's missing billing_project fallback (the bug 33a9964
fixed for /scan never landed here).
Body shape matches what the upcoming BqAccess refactor (next commit) will
produce, so client-side parsers see one consistent contract throughout
the staged rollout. _fetch_bq_table_options preserved exactly as-is —
its swallow-all-and-return-empty contract is intentional and survives
into the refactor."
```
This is the final Phase 1 commit (the one that goes into the PR).
---
# Phase 2 — Commit 2: `BqAccess` facade extraction + migration
Goal: extract the `BqAccess` facade and migrate all four call sites to use it. Delete the inline `try/except` blocks from Phase 1; route through `translate_bq_error` instead. Drop `_bq_client_factory` from `RemoteQueryEngine`.
---
### Task 2.1: Create `connectors/bigquery/access.py` skeleton with `BqProjects` + `BqAccessError`
**Files:**
- Create: `connectors/bigquery/access.py`.
- Create: `tests/test_bq_access.py`.
- [ ] **Step 1: Write failing tests for `BqProjects` and `BqAccessError`**
Create `tests/test_bq_access.py`:
```python
"""Tests for connectors/bigquery/access.py — the BqAccess facade."""
import pytest
class TestBqProjects:
def test_bq_projects_is_frozen_dataclass(self):
from connectors.bigquery.access import BqProjects
p = BqProjects(billing="b", data="d")
assert p.billing == "b"
assert p.data == "d"
with pytest.raises(Exception): # FrozenInstanceError or AttributeError
p.billing = "other"
class TestBqAccessError:
def test_carries_kind_message_details(self):
from connectors.bigquery.access import BqAccessError
e = BqAccessError("my_kind", "boom", {"foo": "bar"})
assert e.kind == "my_kind"
assert e.message == "boom"
assert e.details == {"foo": "bar"}
assert str(e) == "boom"
def test_default_details_is_empty_dict(self):
from connectors.bigquery.access import BqAccessError
e = BqAccessError("k", "m")
assert e.details == {}
def test_http_status_map_covers_all_kinds(self):
from connectors.bigquery.access import BqAccessError
expected = {
"not_configured": 500,
"bq_lib_missing": 500,
"auth_failed": 502,
"cross_project_forbidden": 502,
"bq_forbidden": 502,
"bq_bad_request": 400,
"bq_upstream_error": 502,
}
assert BqAccessError.HTTP_STATUS == expected
```
- [ ] **Step 2: Run — expect FAIL (module doesn't exist)**
Run: `pytest tests/test_bq_access.py -v`
Expected: collection error / ModuleNotFoundError on `connectors.bigquery.access`.
- [ ] **Step 3: Create the module skeleton**
Create `connectors/bigquery/access.py`:
```python
"""Single entry point for BigQuery access — config resolution, client construction,
DuckDB-extension session, and Google-API error translation.
See docs/superpowers/specs/2026-04-29-issue-134-bq-access-unify-design.md for the
full design rationale.
"""
from __future__ import annotations
import functools
import logging
from contextlib import contextmanager
from dataclasses import dataclass
from typing import Callable, Iterator, Literal
logger = logging.getLogger(__name__)
@dataclass(frozen=True)
class BqProjects:
"""Pair of GCP project IDs used by Agnes.
`billing` is the project the BQ client bills jobs to (also used as quota_project_id).
`data` is the default data project for FROM-clause construction. Today equal to
instance.yaml `data_source.bigquery.project`; locked to a single project per instance
until table_registry grows a per-table source_project column. See spec "Non-goals".
"""
billing: str
data: str
class BqAccessError(Exception):
"""Typed error for BQ access failures.
`kind` is one of HTTP_STATUS keys; endpoint translation maps it to status codes.
"""
HTTP_STATUS = {
"not_configured": 500, # admin/config bug — page on-call
"bq_lib_missing": 500, # deployment bug
"auth_failed": 502, # GCP metadata server unreachable
"cross_project_forbidden": 502, # SA lacks serviceusage.services.use on billing project
"bq_forbidden": 502, # other Forbidden from BQ
"bq_bad_request": 400, # 400 from BQ when caller flagged it as client-derived
"bq_upstream_error": 502, # all other upstream BQ failures
}
def __init__(self, kind: str, message: str, details: dict | None = None):
self.kind = kind
self.message = message
self.details = details or {}
super().__init__(message)
```
- [ ] **Step 4: Run — expect PASS**
Run: `pytest tests/test_bq_access.py::TestBqProjects tests/test_bq_access.py::TestBqAccessError -v`
Expected: all green.
- [ ] **Step 5: Commit**
```bash
git add connectors/bigquery/access.py tests/test_bq_access.py
git commit -m "feat(bq_access): skeleton — BqProjects + BqAccessError"
```
---
### Task 2.2: Add `translate_bq_error` to `connectors/bigquery/access.py`
**Files:**
- Modify: `connectors/bigquery/access.py`.
- Modify: `tests/test_bq_access.py`.
- [ ] **Step 1: Add failing tests for `translate_bq_error`**
Append to `tests/test_bq_access.py`:
```python
class TestTranslateBqError:
def setup_method(self):
from connectors.bigquery.access import BqProjects
self.projects = BqProjects(billing="bill", data="data")
def test_passes_through_BqAccessError(self):
"""CRITICAL: bq.client() / bq.duckdb_session() raise BqAccessError directly
for bq_lib_missing / auth_failed. translate_bq_error must pass them through,
not reclassify as 'unknown' and re-raise."""
from connectors.bigquery.access import BqAccessError, translate_bq_error
original = BqAccessError("bq_lib_missing", "no google lib")
result = translate_bq_error(original, self.projects, bad_request_status="client_error")
assert result is original
def test_forbidden_serviceusage_to_cross_project(self):
from google.api_core.exceptions import Forbidden
from connectors.bigquery.access import translate_bq_error
e = Forbidden("Permission denied: serviceusage.services.use on project foo")
result = translate_bq_error(e, self.projects, bad_request_status="client_error")
assert result.kind == "cross_project_forbidden"
assert "billing_project" in result.details
assert "hint" in result.details
def test_forbidden_no_serviceusage_to_bq_forbidden(self):
from google.api_core.exceptions import Forbidden
from connectors.bigquery.access import translate_bq_error
e = Forbidden("Permission denied on table-level ACL")
result = translate_bq_error(e, self.projects, bad_request_status="client_error")
assert result.kind == "bq_forbidden"
def test_forbidden_diff_projects_no_serviceusage_still_bq_forbidden(self):
"""billing != data is the NORMAL cross-project setup, not a signal of failure.
Heuristic must rely on 'serviceusage' substring only."""
from google.api_core.exceptions import Forbidden
from connectors.bigquery.access import translate_bq_error, BqProjects
e = Forbidden("Permission denied on table-level ACL")
result = translate_bq_error(e, BqProjects(billing="b", data="d"),
bad_request_status="client_error")
assert result.kind == "bq_forbidden" # NOT cross_project_forbidden
def test_bad_request_client_error_to_bq_bad_request_400(self):
from google.api_core.exceptions import BadRequest
from connectors.bigquery.access import translate_bq_error, BqAccessError
e = BadRequest("Syntax error at line 1")
result = translate_bq_error(e, self.projects, bad_request_status="client_error")
assert result.kind == "bq_bad_request"
assert BqAccessError.HTTP_STATUS[result.kind] == 400
def test_bad_request_upstream_error_to_bq_upstream_error_502(self):
from google.api_core.exceptions import BadRequest
from connectors.bigquery.access import translate_bq_error, BqAccessError
e = BadRequest("malformed identifier")
result = translate_bq_error(e, self.projects, bad_request_status="upstream_error")
assert result.kind == "bq_upstream_error"
assert BqAccessError.HTTP_STATUS[result.kind] == 502
def test_other_google_api_error_to_bq_upstream_error(self):
from google.api_core.exceptions import InternalServerError
from connectors.bigquery.access import translate_bq_error
e = InternalServerError("BQ borked")
result = translate_bq_error(e, self.projects, bad_request_status="client_error")
assert result.kind == "bq_upstream_error"
def test_unknown_exception_reraises(self):
from connectors.bigquery.access import translate_bq_error
with pytest.raises(RuntimeError, match="oops"):
translate_bq_error(RuntimeError("oops"), self.projects,
bad_request_status="client_error")
```
- [ ] **Step 2: Run — expect FAIL (function doesn't exist)**
Run: `pytest tests/test_bq_access.py::TestTranslateBqError -v`
Expected: ImportError on `translate_bq_error`.
- [ ] **Step 3: Implement `translate_bq_error`**
Append to `connectors/bigquery/access.py`:
```python
def translate_bq_error(
e: Exception,
projects: BqProjects,
*,
bad_request_status: Literal["client_error", "upstream_error"],
) -> BqAccessError:
"""Convert Google API exceptions into a typed BqAccessError.
Mapping (FIRST match wins):
1. BqAccessError -> pass through unchanged (CRITICAL: bq.client()
and bq.duckdb_session() can raise BqAccessError
directly for bq_lib_missing / auth_failed; those
must round-trip without reclassification)
2. Forbidden + 'serviceusage' in str(e).lower()
-> cross_project_forbidden (with hint)
3. Forbidden -> bq_forbidden
4. BadRequest, bad_request_status='client_error'
-> bq_bad_request (HTTP 400)
5. BadRequest, bad_request_status='upstream_error'
-> bq_upstream_error (HTTP 502)
6. GoogleAPICallError (other) -> bq_upstream_error
7. Anything else -> RE-RAISED unchanged (don't swallow programmer errors)
"""
if isinstance(e, BqAccessError):
return e
try:
from google.api_core import exceptions as gax # type: ignore
except ImportError:
# No google lib installed → can't classify Google errors. Re-raise.
raise e
msg = str(e)
if isinstance(e, gax.Forbidden):
if "serviceusage" in msg.lower():
return BqAccessError(
"cross_project_forbidden",
msg,
details={
"billing_project": projects.billing,
"data_project": projects.data,
"hint": (
"Set data_source.bigquery.billing_project in instance.yaml to a project "
"where the SA has serviceusage.services.use, or grant the SA that role "
"on the data project."
),
},
)
return BqAccessError(
"bq_forbidden",
msg,
details={"billing_project": projects.billing, "data_project": projects.data},
)
if isinstance(e, gax.BadRequest):
if bad_request_status == "client_error":
return BqAccessError("bq_bad_request", msg)
return BqAccessError("bq_upstream_error", msg)
if isinstance(e, gax.GoogleAPICallError):
return BqAccessError("bq_upstream_error", msg)
# Don't swallow programmer errors / unknown exceptions
raise e
```
- [ ] **Step 4: Run — expect PASS**
Run: `pytest tests/test_bq_access.py::TestTranslateBqError -v`
Expected: all green.
- [ ] **Step 5: Commit**
```bash
git add connectors/bigquery/access.py tests/test_bq_access.py
git commit -m "feat(bq_access): translate_bq_error — typed BqAccessError mapping"
```
---
### Task 2.3: Add `_default_client_factory` and `_default_duckdb_session_factory`
**Files:**
- Modify: `connectors/bigquery/access.py`.
- Modify: `tests/test_bq_access.py`.
- [ ] **Step 1: Write failing tests**
Append to `tests/test_bq_access.py`:
```python
class TestDefaultClientFactory:
def test_constructs_client_with_billing_project_as_quota(self, monkeypatch):
"""quota_project_id must be projects.billing, NOT projects.data."""
from connectors.bigquery.access import _default_client_factory, BqProjects
captured = {}
class FakeClientOptions:
def __init__(self, **kwargs):
captured["client_options_kwargs"] = kwargs
class FakeClient:
def __init__(self, project, client_options):
captured["project"] = project
captured["client_options"] = client_options
import google.cloud.bigquery as bq_mod
import google.api_core.client_options as co_mod
monkeypatch.setattr(bq_mod, "Client", FakeClient)
monkeypatch.setattr(co_mod, "ClientOptions", FakeClientOptions)
_default_client_factory(BqProjects(billing="bill", data="data"))
assert captured["project"] == "bill"
assert captured["client_options_kwargs"]["quota_project_id"] == "bill"
def test_raises_bq_lib_missing_on_importerror(self, monkeypatch):
"""If google-cloud-bigquery is not installed, raise BqAccessError, not ImportError."""
from connectors.bigquery.access import _default_client_factory, BqProjects, BqAccessError
import builtins
real_import = builtins.__import__
def fake_import(name, *args, **kwargs):
if name == "google.cloud" or name.startswith("google.cloud.bigquery"):
raise ImportError("no google-cloud-bigquery")
return real_import(name, *args, **kwargs)
monkeypatch.setattr(builtins, "__import__", fake_import)
with pytest.raises(BqAccessError) as exc_info:
_default_client_factory(BqProjects(billing="b", data="d"))
assert exc_info.value.kind == "bq_lib_missing"
class TestDefaultDuckdbSessionFactory:
def test_yields_duckdb_conn_with_secret_then_closes(self, monkeypatch):
from connectors.bigquery.access import _default_duckdb_session_factory, BqProjects
executed_sql = []
class FakeConn:
def __init__(self):
self.closed = False
def execute(self, sql, params=None):
executed_sql.append((sql, params))
return self
def close(self):
self.closed = True
fake_conn = FakeConn()
monkeypatch.setattr("duckdb.connect", lambda _: fake_conn)
monkeypatch.setattr("connectors.bigquery.auth.get_metadata_token", lambda: "tok123")
with _default_duckdb_session_factory(BqProjects(billing="b", data="d")) as conn:
assert conn is fake_conn
assert fake_conn.closed is True
# Verify INSTALL/LOAD/SECRET sequence ran
assert any("INSTALL bigquery" in sql for sql, _ in executed_sql)
assert any("LOAD bigquery" in sql for sql, _ in executed_sql)
assert any("CREATE OR REPLACE SECRET" in sql and "tok123" in sql for sql, _ in executed_sql)
def test_closes_on_exception_inside_with_block(self, monkeypatch):
from connectors.bigquery.access import _default_duckdb_session_factory, BqProjects
class FakeConn:
closed = False
def execute(self, *a, **kw): return self
def close(self): self.closed = True
fake_conn = FakeConn()
monkeypatch.setattr("duckdb.connect", lambda _: fake_conn)
monkeypatch.setattr("connectors.bigquery.auth.get_metadata_token", lambda: "t")
with pytest.raises(RuntimeError, match="boom"):
with _default_duckdb_session_factory(BqProjects(billing="b", data="d")) as conn:
raise RuntimeError("boom")
assert fake_conn.closed is True
def test_translates_metadata_auth_error_to_auth_failed(self, monkeypatch):
from connectors.bigquery.access import _default_duckdb_session_factory, BqProjects, BqAccessError
from connectors.bigquery.auth import BQMetadataAuthError
def fail():
raise BQMetadataAuthError("metadata server unreachable")
monkeypatch.setattr("connectors.bigquery.auth.get_metadata_token", fail)
with pytest.raises(BqAccessError) as exc_info:
with _default_duckdb_session_factory(BqProjects(billing="b", data="d")):
pass
assert exc_info.value.kind == "auth_failed"
```
- [ ] **Step 2: Run — expect FAIL**
Run: `pytest tests/test_bq_access.py::TestDefaultClientFactory tests/test_bq_access.py::TestDefaultDuckdbSessionFactory -v`
Expected: FAIL — factories don't exist.
- [ ] **Step 3: Implement the two default factories**
Append to `connectors/bigquery/access.py`:
```python
def _default_client_factory(projects: BqProjects):
"""Real BigQuery client construction. Raises BqAccessError on import / config issues."""
try:
from google.cloud import bigquery # type: ignore
from google.api_core.client_options import ClientOptions # type: ignore
except ImportError as e:
raise BqAccessError(
"bq_lib_missing",
"google-cloud-bigquery is not installed",
details={"original": str(e)},
)
return bigquery.Client(
project=projects.billing,
client_options=ClientOptions(quota_project_id=projects.billing),
)
@contextmanager
def _default_duckdb_session_factory(projects: BqProjects):
"""Yield an in-memory DuckDB conn with bigquery extension loaded + SECRET set
from get_metadata_token(). Auto-cleanup. Translates auth/install failures
to BqAccessError(kind='auth_failed' or 'bq_lib_missing').
Note: `projects.billing` is not used by this factory directly — bigquery_query()
callers pass it themselves as the first positional arg to identify the billing
project. The factory keeps the parameter for symmetry with _default_client_factory.
"""
import duckdb # type: ignore
from connectors.bigquery.auth import get_metadata_token, BQMetadataAuthError
try:
token = get_metadata_token()
except BQMetadataAuthError as e:
raise BqAccessError(
"auth_failed",
f"could not fetch GCP metadata token: {e}",
details={"original": str(e)},
)
conn = duckdb.connect(":memory:")
try:
try:
conn.execute("INSTALL bigquery FROM community; LOAD bigquery;")
escaped = token.replace("'", "''")
conn.execute(
f"CREATE OR REPLACE SECRET bq_s (TYPE bigquery, ACCESS_TOKEN '{escaped}')"
)
except Exception as e:
raise BqAccessError(
"bq_lib_missing",
f"failed to install/load BigQuery DuckDB extension: {e}",
details={"original": str(e)},
)
yield conn
finally:
conn.close()
```
- [ ] **Step 4: Run — expect PASS**
Run: `pytest tests/test_bq_access.py::TestDefaultClientFactory tests/test_bq_access.py::TestDefaultDuckdbSessionFactory -v`
Expected: all green.
- [ ] **Step 5: Commit**
```bash
git add connectors/bigquery/access.py tests/test_bq_access.py
git commit -m "feat(bq_access): default factories for client + duckdb_session"
```
---
### Task 2.4: Add `BqAccess` class
**Files:**
- Modify: `connectors/bigquery/access.py`.
- Modify: `tests/test_bq_access.py`.
- [ ] **Step 1: Write failing tests**
Append to `tests/test_bq_access.py`:
```python
class TestBqAccess:
def test_uses_default_factories_when_none_passed(self, monkeypatch):
from connectors.bigquery.access import BqAccess, BqProjects
captured = []
monkeypatch.setattr(
"connectors.bigquery.access._default_client_factory",
lambda projects: captured.append(("client", projects)) or "FAKE_CLIENT",
)
bq = BqAccess(BqProjects(billing="b", data="d"))
assert bq.client() == "FAKE_CLIENT"
assert captured == [("client", BqProjects(billing="b", data="d"))]
def test_injected_client_factory_overrides_default(self):
from connectors.bigquery.access import BqAccess, BqProjects
bq = BqAccess(
BqProjects(billing="b", data="d"),
client_factory=lambda projects: "MOCK_CLIENT",
)
assert bq.client() == "MOCK_CLIENT"
def test_injected_duckdb_session_factory_overrides_default(self):
from connectors.bigquery.access import BqAccess, BqProjects
from contextlib import contextmanager
@contextmanager
def fake_session(projects):
yield "FAKE_CONN"
bq = BqAccess(
BqProjects(billing="b", data="d"),
duckdb_session_factory=fake_session,
)
with bq.duckdb_session() as conn:
assert conn == "FAKE_CONN"
def test_projects_property(self):
from connectors.bigquery.access import BqAccess, BqProjects
p = BqProjects(billing="b", data="d")
bq = BqAccess(p)
assert bq.projects is p
```
- [ ] **Step 2: Run — expect FAIL (BqAccess class doesn't exist)**
Run: `pytest tests/test_bq_access.py::TestBqAccess -v`
Expected: AttributeError on `BqAccess`.
- [ ] **Step 3: Implement `BqAccess`**
Append to `connectors/bigquery/access.py`:
```python
class BqAccess:
"""Single entry point for BigQuery access. Stateless after construction.
Factories are injectable for tests:
bq = BqAccess(
BqProjects(billing="test-billing", data="test-data"),
client_factory=lambda projects: mock_client,
)
"""
def __init__(
self,
projects: BqProjects,
*,
client_factory: Callable[[BqProjects], object] | None = None,
duckdb_session_factory: Callable[[BqProjects], object] | None = None,
):
self._projects = projects
self._client_factory = client_factory or _default_client_factory
self._duckdb_session_factory = duckdb_session_factory or _default_duckdb_session_factory
@property
def projects(self) -> BqProjects:
return self._projects
def client(self):
"""Construct (or retrieve from injected factory) a BigQuery client."""
return self._client_factory(self._projects)
@contextmanager
def duckdb_session(self) -> Iterator[object]:
"""Yield in-memory DuckDB conn with bigquery extension loaded + SECRET set."""
with self._duckdb_session_factory(self._projects) as conn:
yield conn
```
- [ ] **Step 4: Run — expect PASS**
Run: `pytest tests/test_bq_access.py::TestBqAccess -v`
Expected: all green.
- [ ] **Step 5: Commit**
```bash
git add connectors/bigquery/access.py tests/test_bq_access.py
git commit -m "feat(bq_access): BqAccess facade with injectable factories"
```
---
### Task 2.5: Add `get_bq_access` (module-level cached entry point)
**Files:**
- Modify: `connectors/bigquery/access.py`.
- Modify: `tests/test_bq_access.py`.
- [ ] **Step 1: Write failing tests**
Append to `tests/test_bq_access.py`:
```python
class TestGetBqAccess:
def setup_method(self):
# Clear the cache between tests
from connectors.bigquery.access import get_bq_access
get_bq_access.cache_clear()
def test_env_var_wins(self, monkeypatch):
from connectors.bigquery.access import get_bq_access
monkeypatch.setenv("BIGQUERY_PROJECT", "env-proj")
bq = get_bq_access()
assert bq.projects.billing == "env-proj"
assert bq.projects.data == "env-proj"
def test_billing_project_from_yaml_when_no_env(self, monkeypatch):
from connectors.bigquery.access import get_bq_access
monkeypatch.delenv("BIGQUERY_PROJECT", raising=False)
def fake_get_value(*keys, default=""):
return {
("data_source", "bigquery", "billing_project"): "yaml-bill",
("data_source", "bigquery", "project"): "yaml-data",
}.get(keys, default)
monkeypatch.setattr("app.instance_config.get_value", fake_get_value)
bq = get_bq_access()
assert bq.projects.billing == "yaml-bill"
assert bq.projects.data == "yaml-data"
def test_billing_falls_back_to_project_when_no_billing(self, monkeypatch):
from connectors.bigquery.access import get_bq_access
monkeypatch.delenv("BIGQUERY_PROJECT", raising=False)
def fake_get_value(*keys, default=""):
return {
("data_source", "bigquery", "project"): "yaml-data",
}.get(keys, default)
monkeypatch.setattr("app.instance_config.get_value", fake_get_value)
bq = get_bq_access()
assert bq.projects.billing == "yaml-data"
assert bq.projects.data == "yaml-data"
def test_raises_not_configured_when_neither_set(self, monkeypatch):
from connectors.bigquery.access import get_bq_access, BqAccessError
monkeypatch.delenv("BIGQUERY_PROJECT", raising=False)
monkeypatch.setattr("app.instance_config.get_value", lambda *k, default="": default)
with pytest.raises(BqAccessError) as exc_info:
get_bq_access()
assert exc_info.value.kind == "not_configured"
assert "billing_project" in exc_info.value.details["hint"].lower() or \
"project" in exc_info.value.details["hint"].lower()
def test_is_cached(self, monkeypatch):
from connectors.bigquery.access import get_bq_access
monkeypatch.setenv("BIGQUERY_PROJECT", "p")
a = get_bq_access()
b = get_bq_access()
assert a is b
def test_does_not_cache_exceptions(self, monkeypatch):
"""functools.cache does not cache exceptions — config can be fixed and retried."""
from connectors.bigquery.access import get_bq_access, BqAccessError
monkeypatch.delenv("BIGQUERY_PROJECT", raising=False)
monkeypatch.setattr("app.instance_config.get_value", lambda *k, default="": default)
with pytest.raises(BqAccessError):
get_bq_access()
# Now "fix" the config
monkeypatch.setenv("BIGQUERY_PROJECT", "p")
bq = get_bq_access()
assert bq.projects.billing == "p"
```
- [ ] **Step 2: Run — expect FAIL (function doesn't exist)**
Run: `pytest tests/test_bq_access.py::TestGetBqAccess -v`
Expected: ImportError on `get_bq_access`.
- [ ] **Step 3: Implement `get_bq_access`**
Append to `connectors/bigquery/access.py`:
```python
@functools.cache
def get_bq_access() -> BqAccess:
"""Module-level FastAPI Depends target. Resolves projects from config and returns
a BqAccess instance with default factories.
Resolution order:
1. BIGQUERY_PROJECT env var → both billing + data (legacy override)
2. instance.yaml data_source.bigquery.billing_project → billing
3. instance.yaml data_source.bigquery.project → data, and billing if (2) is unset
Process-cached. Hot-reload of instance.yaml is out of scope; restart the container
on config change. functools.cache does NOT cache exceptions, so a failed call is
retried on the next invocation.
Tests inject via `app.dependency_overrides[get_bq_access] = lambda: bq` for
endpoints, or construct `BqAccess(...)` directly for non-endpoint code.
Module-level (not a classmethod) to avoid the @classmethod + @functools.cache
stacking footgun and to give FastAPI's dependency introspection a clean signature.
"""
import os
env_project = os.environ.get("BIGQUERY_PROJECT", "").strip()
if env_project:
return BqAccess(BqProjects(billing=env_project, data=env_project))
from app.instance_config import get_value
billing = (get_value("data_source", "bigquery", "billing_project", default="") or "").strip()
data = (get_value("data_source", "bigquery", "project", default="") or "").strip()
if not data:
raise BqAccessError(
"not_configured",
"BigQuery project not configured",
details={
"hint": (
"Set data_source.bigquery.project in instance.yaml "
"(and optionally data_source.bigquery.billing_project for cross-project "
"deployments). BIGQUERY_PROJECT env var also accepted as legacy override."
),
},
)
if not billing:
billing = data
return BqAccess(BqProjects(billing=billing, data=data))
```
- [ ] **Step 4: Run — expect PASS**
Run: `pytest tests/test_bq_access.py::TestGetBqAccess -v`
Expected: all green.
- [ ] **Step 5: Run the WHOLE bq_access test file as a sanity check**
Run: `pytest tests/test_bq_access.py -v`
Expected: every test green.
- [ ] **Step 6: Commit**
```bash
git add connectors/bigquery/access.py tests/test_bq_access.py
git commit -m "feat(bq_access): get_bq_access — cached module-level entry point"
```
---
### Task 2.6: Add `bq_access` fixture to `tests/conftest.py`
**Files:**
- Modify: `tests/conftest.py`.
- [ ] **Step 1: Add fixture without test (it's infrastructure for later tasks)**
Append to `tests/conftest.py`:
```python
import contextlib as _contextlib
@pytest.fixture
def bq_access():
"""Build a BqAccess with pluggable factories and override the FastAPI Depends.
Usage:
def test_x(bq_access):
mock_client = MagicMock()
bq = bq_access(client=mock_client)
# endpoint test code
Override is auto-cleared on fixture teardown.
NOTE: `contextlib.nullcontext(duckdb_conn)` does NOT close the conn on exit.
The production path closes via _default_duckdb_session_factory. Tests that
care about close behavior should use that factory directly (see
tests/test_bq_access.py::TestDefaultDuckdbSessionFactory).
"""
from connectors.bigquery.access import BqAccess, BqProjects, get_bq_access
from app.main import app
def _build(*, client=None, duckdb_conn=None,
billing="test-billing", data="test-data"):
bq = BqAccess(
BqProjects(billing=billing, data=data),
client_factory=(lambda projects: client) if client is not None else None,
duckdb_session_factory=(
lambda projects: _contextlib.nullcontext(duckdb_conn)
) if duckdb_conn is not None else None,
)
app.dependency_overrides[get_bq_access] = lambda: bq
return bq
yield _build
from app.main import app as _app
_app.dependency_overrides.pop(get_bq_access, None)
```
(Note: import path for `app` may need adjustment depending on how tests import it; match the existing pattern.)
- [ ] **Step 2: Verify fixture loads (no test, just collection)**
Run: `pytest --collect-only tests/conftest.py 2>&1 | tail -5`
Expected: no collection errors.
- [ ] **Step 3: Commit**
```bash
git add tests/conftest.py
git commit -m "test: add bq_access fixture for FastAPI dep override"
```
---
### Task 2.7: Migrate `app/api/v2_scan.py` to `BqAccess`
**Files:**
- Modify: `app/api/v2_scan.py`.
- Modify: `tests/test_v2_scan.py`, `tests/test_v2_scan_estimate.py` (update lambda signatures + use new fixture).
- [ ] **Step 1: Read current state to confirm Phase-1 try/except blocks are still in place**
Run: `sed -n '43,90p' app/api/v2_scan.py`
Expected: `_bq_dry_run_bytes` has the inline try/except added in Task 1.2.
- [ ] **Step 2: Update `_bq_dry_run_bytes` signature and body**
Replace `app/api/v2_scan.py:43-95` (the entire `_bq_dry_run_bytes` from Task 1.2):
```python
def _bq_dry_run_bytes(bq, sql: str) -> int:
"""Run a BQ dry-run via the google-cloud-bigquery client and return totalBytesProcessed."""
from google.cloud import bigquery
from connectors.bigquery.access import translate_bq_error
client = bq.client() # raises BqAccessError(bq_lib_missing/auth_failed) — propagates as-is
try:
job = client.query(
sql, job_config=bigquery.QueryJobConfig(dry_run=True, use_query_cache=False),
)
return int(job.total_bytes_processed or 0)
except Exception as e:
raise translate_bq_error(e, bq.projects, bad_request_status="client_error")
```
- [ ] **Step 3: Update `_run_bq_scan` similarly**
Replace `app/api/v2_scan.py:266-...` (the entire `_run_bq_scan` from Task 1.3):
```python
def _run_bq_scan(bq, sql: str):
"""Run a BQ query via DuckDB BQ extension. Returns Arrow table."""
from connectors.bigquery.access import translate_bq_error
with bq.duckdb_session() as conn:
try:
return conn.execute(
"SELECT * FROM bigquery_query(?, ?)",
[bq.projects.billing, sql],
).arrow()
except Exception as e:
raise translate_bq_error(e, bq.projects, bad_request_status="client_error")
```
- [ ] **Step 4: Update `estimate` function signature and call site**
In `app/api/v2_scan.py:135-...` (`def estimate(...)`):
- Remove `project_id: str, billing_project: str | None = None` parameters.
- Add `bq` parameter.
- Inside: replace `_bq_dry_run_bytes(billing_project or project_id, bq_sql)` with `_bq_dry_run_bytes(bq, bq_sql)`.
- Replace `_build_bq_sql(row, project_id, req)` with `_build_bq_sql(row, bq.projects.data, req)`.
- Replace `_resolve_schema(conn, user, req.table_id, project_id)` with `_resolve_schema(conn, user, req.table_id, bq.projects.data)`.
- [ ] **Step 5: Update `run_scan` function similarly**
Same pattern: drop `project_id`, `billing_project` params; add `bq`. Replace internal references.
- [ ] **Step 6: Update `scan_endpoint` and `scan_estimate_endpoint`**
Replace lines `app/api/v2_scan.py:217-...` (`scan_estimate_endpoint`) and `:378-...` (`scan_endpoint`):
```python
@router.post("/scan/estimate")
async def scan_estimate_endpoint(
raw: dict,
user: dict = Depends(get_current_user),
conn: duckdb.DuckDBPyConnection = Depends(_get_db),
bq: BqAccess = Depends(get_bq_access),
):
try:
return estimate(conn, user, raw, bq=bq)
except WhereValidationError as e:
raise HTTPException(status_code=400, detail={"error": "validator_rejected", "kind": e.kind, "details": e.detail or {}})
except PermissionError:
raise HTTPException(status_code=403, detail="not authorized for this table")
except FileNotFoundError as e:
raise HTTPException(status_code=404, detail=f"table {e!s} not found")
except ValueError as e:
raise HTTPException(status_code=400, detail=str(e))
except BqAccessError as e:
raise HTTPException(
status_code=BqAccessError.HTTP_STATUS.get(e.kind, 500),
detail={"error": e.kind, "message": e.message, "details": e.details},
)
@router.post("/scan")
async def scan_endpoint(
raw: dict,
user: dict = Depends(get_current_user),
conn: duckdb.DuckDBPyConnection = Depends(_get_db),
bq: BqAccess = Depends(get_bq_access),
):
quota = _build_quota_tracker()
try:
ipc = run_scan(conn, user, raw, bq=bq, quota=quota)
return Response(content=ipc, media_type=CONTENT_TYPE)
except WhereValidationError as e:
raise HTTPException(status_code=400, detail={"error": "validator_rejected", "kind": e.kind, "details": e.detail or {}})
except QuotaExceededError as e:
raise HTTPException(status_code=429, detail={"error": "quota_exceeded", **(e.detail or {})})
except PermissionError:
raise HTTPException(status_code=403, detail="not authorized for this table")
except FileNotFoundError as e:
raise HTTPException(status_code=404, detail=f"table {e!s} not found")
except ValueError as e:
raise HTTPException(status_code=400, detail=str(e))
except BqAccessError as e:
raise HTTPException(
status_code=BqAccessError.HTTP_STATUS.get(e.kind, 500),
detail={"error": e.kind, "message": e.message, "details": e.details},
)
```
Add at top of `app/api/v2_scan.py`:
```python
from connectors.bigquery.access import BqAccess, BqAccessError, get_bq_access
```
- [ ] **Step 7: Update existing tests in `tests/test_v2_scan.py` and `tests/test_v2_scan_estimate.py` for the new signature**
Tests that previously did `monkeypatch.setattr(v2_scan, "_bq_dry_run_bytes", lambda project, sql: ...)` now need `lambda bq, sql: ...`. Same for `_run_bq_scan`. Trace through all such monkeypatches and update.
The Forbidden / BadRequest tests from Task 1.2 / 1.3 also need updating: instead of patching `_bq_dry_run_bytes` to raise, you can either (a) keep patching that function (now with new signature), or (b) use the new `bq_access` fixture to inject a client that raises. Pattern (b) is cleaner long-term:
```python
def test_scan_estimate_returns_502_on_bq_forbidden_serviceusage(reload_db, bq_access):
from google.api_core.exceptions import Forbidden
from unittest.mock import MagicMock
mock_client = MagicMock()
mock_client.query.side_effect = Forbidden("Permission denied: serviceusage.services.use on project foo")
bq_access(client=mock_client)
# ... call POST /api/v2/scan/estimate
# Assert response.status_code == 502
# Assert response.json()["detail"]["error"] == "cross_project_forbidden"
```
- [ ] **Step 8: Run v2_scan tests — expect PASS**
Run: `pytest tests/test_v2_scan.py tests/test_v2_scan_estimate.py -v`
Expected: all green.
- [ ] **Step 9: Commit**
```bash
git add app/api/v2_scan.py tests/test_v2_scan.py tests/test_v2_scan_estimate.py
git commit -m "refactor(v2_scan): #134 migrate to BqAccess facade"
```
---
### Task 2.8: Migrate `app/api/v2_sample.py` to `BqAccess`
**Files:**
- Modify: `app/api/v2_sample.py`.
- Modify: `tests/test_v2_sample.py`.
- [ ] **Step 1: Update `_fetch_bq_sample` signature and body**
Replace the entire function (added in Task 1.1):
```python
def _fetch_bq_sample(bq, dataset: str, table: str, n: int) -> list[dict]:
from connectors.bigquery.access import translate_bq_error
from src.identifier_validation import validate_quoted_identifier
if not (validate_quoted_identifier(bq.projects.data, "BQ project")
and validate_quoted_identifier(dataset, "BQ dataset")
and validate_quoted_identifier(table, "BQ source_table")):
raise ValueError("unsafe BQ identifier in registry — refusing to query")
bq_sql = f"SELECT * FROM `{bq.projects.data}.{dataset}.{table}` LIMIT {int(n)}"
with bq.duckdb_session() as conn:
try:
df = conn.execute(
"SELECT * FROM bigquery_query(?, ?)",
[bq.projects.billing, bq_sql],
).fetchdf()
return df.to_dict(orient="records")
except Exception as e:
raise translate_bq_error(e, bq.projects, bad_request_status="upstream_error")
```
- [ ] **Step 2: Update `build_sample` signature**
Replace `def build_sample(conn, user, table_id, *, n, project_id):` with:
```python
def build_sample(conn, user, table_id, *, n, bq):
```
Inside, replace `_fetch_bq_sample(project_id, ...)` with `_fetch_bq_sample(bq, ...)`. Local-source (parquet) path unchanged.
- [ ] **Step 3: Update `sample` endpoint**
Replace `app/api/v2_sample.py:97-110`:
```python
@router.get("/sample/{table_id}")
async def sample(
table_id: str,
n: int = Query(default=5, ge=1, le=_MAX_N),
user: dict = Depends(get_current_user),
conn: duckdb.DuckDBPyConnection = Depends(_get_db),
bq: BqAccess = Depends(get_bq_access),
):
try:
return build_sample(conn, user, table_id, n=n, bq=bq)
except FileNotFoundError:
raise HTTPException(status_code=404, detail=f"table {table_id!r} not found")
except PermissionError:
raise HTTPException(status_code=403, detail="not authorized for this table")
except ValueError as e:
raise HTTPException(
status_code=400,
detail={"error": "unsafe_identifier", "message": str(e), "details": {}},
)
except BqAccessError as e:
raise HTTPException(
status_code=BqAccessError.HTTP_STATUS.get(e.kind, 500),
detail={"error": e.kind, "message": e.message, "details": e.details},
)
```
Add at top:
```python
from connectors.bigquery.access import BqAccess, BqAccessError, get_bq_access
```
- [ ] **Step 4: Update tests in `tests/test_v2_sample.py`**
The Phase-1 tests patched `_fetch_bq_sample` with a 4-arg lambda `(project, dataset, table, n)`. Update all such patches to `(bq, dataset, table, n)`. Convert the Forbidden test to use `bq_access` fixture (matches pattern from Task 2.7 step 7).
- [ ] **Step 5: Run — expect PASS**
Run: `pytest tests/test_v2_sample.py -v`
Expected: all green.
- [ ] **Step 6: Commit**
```bash
git add app/api/v2_sample.py tests/test_v2_sample.py
git commit -m "refactor(v2_sample): #134 migrate to BqAccess facade"
```
---
### Task 2.9: Migrate `app/api/v2_schema.py` (both blocks, preserve best-effort semantics)
**Files:**
- Modify: `app/api/v2_schema.py`.
- Modify: `tests/test_v2_schema.py`.
- [ ] **Step 1: Update `_fetch_bq_schema` (strict block) — replace Phase-1 inline try/except with translate_bq_error**
Replace `app/api/v2_schema.py:36-...` (the `_fetch_bq_schema` function). Body becomes:
```python
def _fetch_bq_schema(bq, dataset: str, table: str) -> list[dict]:
from connectors.bigquery.access import translate_bq_error
from src.identifier_validation import validate_quoted_identifier
if not (validate_quoted_identifier(bq.projects.data, "BQ project")
and validate_quoted_identifier(dataset, "BQ dataset")
and validate_quoted_identifier(table, "BQ source_table")):
raise ValueError("unsafe BQ identifier in registry")
bq_sql = (
f"SELECT column_name, data_type FROM `{bq.projects.data}.{dataset}.INFORMATION_SCHEMA.COLUMNS` "
f"WHERE table_name = ? "
f"ORDER BY ordinal_position"
)
with bq.duckdb_session() as conn:
try:
rows = conn.execute(
"SELECT * FROM bigquery_query(?, ?, ?)",
[bq.projects.billing, bq_sql, table],
).fetchall()
except Exception as e:
raise translate_bq_error(e, bq.projects, bad_request_status="upstream_error")
return [{"name": r[0], "type": r[1]} for r in rows]
```
(Adjust the SELECT shape to match what the function returns today; this sketch preserves the {name, type} shape of the existing function.)
- [ ] **Step 2: Update `_fetch_bq_table_options` (best-effort block) — preserve swallow-all**
Replace `app/api/v2_schema.py:90-129`:
```python
def _fetch_bq_table_options(bq, dataset: str, table: str) -> dict:
"""Best-effort partition/cluster info. Returns {} on ANY failure (preserved
from pre-refactor; /schema endpoint must keep returning successfully even
when partition queries fail)."""
from src.identifier_validation import validate_quoted_identifier
if not (validate_quoted_identifier(bq.projects.data, "BQ project")
and validate_quoted_identifier(dataset, "BQ dataset")
and validate_quoted_identifier(table, "BQ source_table")):
return {}
try:
with bq.duckdb_session() as conn:
bq_sql = (
f"SELECT column_name, is_partitioning_column, clustering_ordinal_position "
f"FROM `{bq.projects.data}.{dataset}.INFORMATION_SCHEMA.COLUMNS` "
f"WHERE table_name = ? "
f"ORDER BY clustering_ordinal_position NULLS LAST"
)
rows = conn.execute(
"SELECT * FROM bigquery_query(?, ?, ?)",
[bq.projects.billing, bq_sql, table],
).fetchall()
if not rows:
return {}
partition_by = next(
(r[0] for r in rows if (r[1] or "").upper() == "YES"),
None,
)
clustered_by = [r[0] for r in rows if r[2] is not None]
return {"partition_by": partition_by, "clustered_by": clustered_by}
except Exception as e:
logger.warning(
"BQ table options fetch failed for %s.%s.%s: %s",
bq.projects.data, dataset, table, e,
)
return {}
```
**Note**: the outer `try/except Exception → return {}` is the load-bearing contract preserved from today.
- [ ] **Step 3: Update `build_schema` and `schema` endpoint signatures**
`build_schema(conn, user, table_id, *, project_id)``build_schema(conn, user, table_id, *, bq)`.
Inside: `_fetch_bq_schema(project_id, dataset, table)``_fetch_bq_schema(bq, dataset, table)`. Same for `_fetch_bq_table_options`.
`schema` endpoint (`app/api/v2_schema.py:...`):
```python
@router.get("/schema/{table_id}")
async def schema(
table_id: str,
user: dict = Depends(get_current_user),
conn: duckdb.DuckDBPyConnection = Depends(_get_db),
bq: BqAccess = Depends(get_bq_access),
):
try:
return build_schema(conn, user, table_id, bq=bq)
except FileNotFoundError:
raise HTTPException(status_code=404, detail=f"table {table_id!r} not found")
except PermissionError:
raise HTTPException(status_code=403, detail="not authorized for this table")
except ValueError as e:
raise HTTPException(
status_code=400,
detail={"error": "unsafe_identifier", "message": str(e), "details": {}},
)
except BqAccessError as e:
raise HTTPException(
status_code=BqAccessError.HTTP_STATUS.get(e.kind, 500),
detail={"error": e.kind, "message": e.message, "details": e.details},
)
```
Add at top:
```python
from connectors.bigquery.access import BqAccess, BqAccessError, get_bq_access
```
- [ ] **Step 4: Update tests in `tests/test_v2_schema.py`**
Update `_fetch_bq_schema` and `_fetch_bq_table_options` monkeypatches to new signatures `(bq, dataset, table)`. The Phase-1 regression-guard test for partition-info failure should still pass (the contract is preserved).
- [ ] **Step 5: Run — expect PASS**
Run: `pytest tests/test_v2_schema.py -v`
Expected: all green, including the regression guard.
- [ ] **Step 6: Commit**
```bash
git add app/api/v2_schema.py tests/test_v2_schema.py
git commit -m "refactor(v2_schema): #134 migrate to BqAccess (strict + best-effort blocks)"
```
---
### Task 2.10: Migrate `src/remote_query.py` — lazy `bq_access`, drop `_bq_client_factory`
**Files:**
- Modify: `src/remote_query.py`.
- [ ] **Step 1: Read current state of `__init__` and `_get_bq_client`**
Run: `sed -n '195,230p' src/remote_query.py && echo '---' && sed -n '407,455p' src/remote_query.py`
Expected: see today's `_bq_client_factory` parameter and the fallback chain in `_get_bq_client`.
- [ ] **Step 2: Update `__init__` signature**
In `src/remote_query.py` `RemoteQueryEngine.__init__`:
- Remove `_bq_client_factory=None` parameter.
- Add `bq_access: "BqAccess | None" = None` parameter.
- Add `self._bq = bq_access` line in body.
- Remove `self._bq_client_factory = _bq_client_factory` line.
- Update the docstring to remove the old factory reference (the stale line 204 about `scripts.duckdb_manager._create_bq_client`).
- [ ] **Step 3: Replace `_get_bq_client`**
Delete the existing `_get_bq_client` (lines ~407-450) and replace with:
```python
def _get_bq_client(self):
"""Lazy-resolve BqAccess on first use. Many tests construct RemoteQueryEngine
for DuckDB-only paths and never touch BQ — those must not fail with not_configured."""
if self._bq is None:
from connectors.bigquery.access import get_bq_access
self._bq = get_bq_access() # may raise BqAccessError; that's fine
return self._bq.client()
```
- [ ] **Step 4: Run existing remote_query tests — expect MOSTLY PASS, with failures in tests that injected `_bq_client_factory`**
Run: `pytest tests/test_remote_query.py -v 2>&1 | tail -40`
Expected: ~10 failures from tests that pass `_bq_client_factory=...` (now an unknown kwarg) or that mocked the old factory factory chain.
- [ ] **Step 5: DO NOT commit yet** — Task 2.11 fixes the tests. Both must land together.
---
### Task 2.11: Migrate `tests/test_remote_query.py` from `_bq_client_factory` to `bq_access` injection
**Files:**
- Modify: `tests/test_remote_query.py`.
- [ ] **Step 1: Find all `_bq_client_factory` usages**
Run: `grep -n "_bq_client_factory" tests/test_remote_query.py`
Note: the spec mentions ~12+ call sites; expect roughly that count.
- [ ] **Step 2: For each occurrence, migrate to `bq_access=BqAccess(..., client_factory=...)`**
Pattern:
```python
# Before
engine = RemoteQueryEngine(conn, _bq_client_factory=lambda project: mock_client)
# After
from connectors.bigquery.access import BqAccess, BqProjects
bq = BqAccess(
BqProjects(billing="test-billing", data="test-data"),
client_factory=lambda projects: mock_client,
)
engine = RemoteQueryEngine(conn, bq_access=bq)
```
For tests where the factory was `lambda project: ...` and used `project` argument: now it's `lambda projects: ...` (a `BqProjects` instance). If the test asserted on the project argument, change to `assert call.projects.billing == "expected-project"`.
- [ ] **Step 3: DuckDB-only tests need NO change**
Tests like `RemoteQueryEngine(analytics_conn)` (no factory) at lines 106, 148, 188, 196, 417, 520, 529, 538: these now have `bq_access=None` by default and don't trigger `get_bq_access()` because they never call `_get_bq_client`. Confirm these pass without modification.
- [ ] **Step 4: Run — expect PASS**
Run: `pytest tests/test_remote_query.py -v`
Expected: all green.
- [ ] **Step 5: Commit Tasks 2.10 + 2.11 together**
```bash
git add src/remote_query.py tests/test_remote_query.py
git commit -m "refactor(remote_query): #134 lazy BqAccess injection, drop _bq_client_factory
RemoteQueryEngine.__init__ no longer accepts _bq_client_factory. Tests
migrate to bq_access=BqAccess(projects, client_factory=...) injection.
DuckDB-only tests that never touch BQ continue to work without changes
because bq_access defaults to None and get_bq_access() is only invoked
on first BQ call (lazy resolution).
Drops the stale docstring at the previous src/remote_query.py:204
that referenced scripts.duckdb_manager._create_bq_client as the default
factory — RemoteQueryEngine never actually used that function.
CLI (cli/commands/query.py) is unaffected: it never injected the factory,
and the new bq_access kwarg's None default routes through get_bq_access()
on first BQ call — matching today's eager behavior for the CLI path."
```
---
### Task 2.12: CHANGELOG entries for Phase 2
**Files:** `CHANGELOG.md`.
- [ ] **Step 1: Add `### Changed` BREAKING entry under `## [Unreleased]`**
Add immediately below the `### Fixed` block from Phase 1:
```markdown
### Changed
- **BREAKING for deployments using `BIGQUERY_PROJECT` env var alongside `data_source.bigquery.project` in `instance.yaml`.** The env var now sets BOTH billing and data project, overriding `data_source.bigquery.project` for FROM-clause construction in `v2_scan` / `v2_sample` / `v2_schema`. Migrate by clearing `BIGQUERY_PROJECT` and using `data_source.bigquery.billing_project` + `data_source.bigquery.project` in `instance.yaml`. (Previously `BIGQUERY_PROJECT` only affected `RemoteQueryEngine` billing.)
```
- [ ] **Step 2: Add `### Internal` entries**
```markdown
### Internal
- New shared module `connectors/bigquery/access.py``BqAccess` facade unifies BQ project resolution, client construction, DuckDB-extension session, and Google-API error translation across `v2_scan`, `v2_sample`, `v2_schema`, and `RemoteQueryEngine`. Replaces four duplicate code paths with one.
- **Internal API change:** `RemoteQueryEngine.__init__` no longer accepts `_bq_client_factory`. Callers that injected it migrate to `RemoteQueryEngine(..., bq_access=BqAccess(projects, client_factory=...))`. The CLI (`cli/commands/query.py`) is unaffected — it never injected the factory and the new `bq_access` kwarg defaults to `None` (lazy `get_bq_access()` on first BQ call).
- Removed stale docstring in `src/remote_query.py` referencing `scripts.duckdb_manager._create_bq_client` as the default BQ client factory (the engine never actually used that function).
- Two known-duplicate BQ-access sites (`connectors/bigquery/extractor.py`, `scripts/duckdb_manager.register_bq_table`) explicitly out of scope for this PR; tracked as follow-up.
```
- [ ] **Step 3: Commit**
```bash
git add CHANGELOG.md
git commit -m "docs(changelog): #134 Phase 2 BqAccess refactor — Changed BREAKING + Internal"
```
---
# Phase 3 — Verification
### Task 3.1: Run full test suite
- [ ] **Step 1: Run all tests**
Run:
```bash
pytest tests/ -v --tb=short 2>&1 | tee /tmp/test-output-issue-134.log
```
Expected: every test passes. If failures appear that aren't in BQ/v2 paths, investigate — they may be flaky tests unrelated to this PR; document any pre-existing failures.
- [ ] **Step 2: Run linters / type checks if the repo enforces them**
Run: `ruff check connectors/bigquery/access.py app/api/v2_*.py src/remote_query.py tests/test_bq_access.py 2>&1 | head -30`
Expected: no new lint errors introduced.
Run: `mypy connectors/bigquery/access.py 2>&1 | head -30`
Expected: no new type errors. (CI may run with `continue-on-error` per repo convention — match that.)
- [ ] **Step 3: If any failures, fix them before proceeding**
Do NOT proceed to E2E with red CI.
---
### Task 3.2: Squash dev commits into final PR shape (two commits)
The intermediate commits during development can be squashed into the two PR-shape commits that the spec calls for.
- [ ] **Step 1: Identify the boundary**
Phase 1 ended with the squash commit `fix(v2): #134 structured 502/400 on BQ errors across /scan, /scan/estimate, /sample, /schema`. Find its SHA:
```bash
git log --oneline | grep "structured 502/400 on BQ errors across"
```
Note this SHA — call it `$PHASE1`.
- [ ] **Step 2: Soft-reset everything after `$PHASE1` to staging**
```bash
git reset --soft $PHASE1
git status # expect a large pile of staged changes representing Phase 2
```
- [ ] **Step 3: Make the single Phase 2 commit**
```bash
git commit -m "refactor(bq): #134 BqAccess facade — unify v2_scan, v2_sample, v2_schema, RemoteQueryEngine
Extracts the duplicated BigQuery-access pattern (project resolution +
client construction + DuckDB-extension session + Google-API error
translation) into connectors/bigquery/access.py. Migrates four
call sites to use it:
- app/api/v2_scan.py — _bq_dry_run_bytes, _run_bq_scan
- app/api/v2_sample.py — _fetch_bq_sample
- app/api/v2_schema.py — _fetch_bq_schema (strict translation),
_fetch_bq_table_options (preserves swallow-all best-effort contract)
- src/remote_query.py — RemoteQueryEngine, lazy bq_access kwarg
Removes _bq_client_factory parameter from RemoteQueryEngine.__init__
and the stale docstring referencing scripts.duckdb_manager._create_bq_client.
Tests migrate from _bq_client_factory injection to
bq_access=BqAccess(projects, client_factory=...) injection. DuckDB-only
RemoteQueryEngine tests need no changes (lazy resolution skips
get_bq_access() when bq_access is None and BQ is never touched).
BREAKING for deployments combining BIGQUERY_PROJECT env var with
data_source.bigquery.project in instance.yaml — the env var now
overrides data project too. See CHANGELOG.
Two known-duplicate BQ-access sites (connectors/bigquery/extractor.py,
scripts/duckdb_manager.register_bq_table) explicitly out of scope;
tracked as follow-up."
```
- [ ] **Step 4: Verify final shape**
Run: `git log --oneline | head -10`
Expected: top two commits are exactly the two intended for the PR (one fix(v2), one refactor(bq)).
---
### Task 3.3: E2E verification on `<your-dev-instance>`
This task happens AFTER the PR is merged and deployed to `<your-dev-instance>`. Per the spec, this is the success criterion for closing #134 — without it, "fixed" is unverifiable.
**Prerequisite:** `<your-dev-instance>` is running the PR's image; reproduce the PAT used in Pavel's report.
- [ ] **Step 1: Pre-config baseline (BEFORE setting `billing_project`)**
```bash
PAT=...
curl -k -i -H "Authorization: Bearer $PAT" \
https://<your-agnes-host>/api/v2/sample/<bq_table_id>?n=2
curl -k -i -X POST -H "Authorization: Bearer $PAT" -H "Content-Type: application/json" \
https://<your-agnes-host>/api/v2/scan/estimate \
-d '{"table_id":"<bq_table_id>","select":["event_date"],"where":"event_date = DATE \"2026-04-21\""}'
curl -k -i -X POST -H "Authorization: Bearer $PAT" -H "Content-Type: application/json" \
https://<your-agnes-host>/api/v2/scan \
-d '{"table_id":"<bq_table_id>","select":["event_date"],"where":"event_date = DATE \"2026-04-21\"","limit":50}'
curl -k -i -H "Authorization: Bearer $PAT" \
https://<your-agnes-host>/api/v2/schema/<bq_table_id>
```
Expected (per spec § E2E manual verification):
- `/sample`, `/scan`, `/scan/estimate`: HTTP 502 with JSON body containing `error=cross_project_forbidden` + `details.hint` mentioning `billing_project`.
- `/schema`: HTTP 200 (INFORMATION_SCHEMA queries don't trip serviceusage; only fails if something else breaks the strict block).
- [ ] **Step 2: Operator action — set `billing_project` and restart**
In `instance.yaml` on the VM:
```yaml
data_source:
bigquery:
project: <your-data-project> # data project, unchanged
billing_project: <project-where-SA-can-bill> # NEW
```
Restart the container.
- [ ] **Step 3: Post-config verification**
Repeat the four curls from Step 1.
Expected: HTTP 200 on all four, with valid response bodies.
- [ ] **Step 4: Document and close #134**
Add a comment to issue #134 with:
- The four curl invocations + their pre/post status codes.
- Confirmation that the structured body shape is what was specced.
- Note any unexpected behavior.
Close the issue.
---
## Self-review (against the spec)
| Spec section | Plan task |
|---|---|
| Bug A — `v2_scan` missing try/except | Tasks 1.2, 1.3 + 2.7 |
| Bug B — `v2_sample` missing billing_project + try/except | Tasks 1.1 + 2.8 |
| Bug C — `v2_schema` (two blocks, two semantics) | Tasks 1.4 + 2.9 |
| `BqProjects` + `BqAccessError` | Task 2.1 |
| `translate_bq_error` (incl. BqAccessError pass-through) | Task 2.2 |
| `_default_client_factory`, `_default_duckdb_session_factory` | Task 2.3 |
| `BqAccess` class with injectable factories | Task 2.4 |
| `get_bq_access` module-level cached | Task 2.5 |
| `bq_access` fixture in conftest | Task 2.6 |
| RemoteQueryEngine lazy bq_access; drop `_bq_client_factory` | Tasks 2.10 + 2.11 |
| CHANGELOG (Fixed + Changed BREAKING + Internal) | Tasks 1.5 + 2.12 |
| Two-commit PR shape | Task 3.2 |
| E2E verification protocol | Task 3.3 |
| Cross-project Forbidden heuristic ('serviceusage' only) | Task 2.2 (`test_forbidden_diff_projects_no_serviceusage_still_bq_forbidden`) |
| Status code split (400/502/500) | Task 2.1 (`test_http_status_map_covers_all_kinds`) |
| `_fetch_bq_table_options` swallow-all preserved | Task 2.9 + Task 1.4 regression guard |
All spec sections have at least one task implementing them. No placeholders in the plan body.
## Final notes
- Each task's tests come BEFORE the implementation per TDD. Run the failing test first, then implement, then verify pass.
- Intermediate commits during dev are fine — Task 3.2 squashes them into the two PR-shape commits.
- The plan stays in the worktree (`<worktree-root>`, branch `fix/134-bq-access-unify`). All commits land there. Push to GitHub when ready for PR.
- If a task's TDD step finds an issue with the spec's design, STOP and update the spec before proceeding. Don't paper over design holes in implementation.