feat(routers): split remaining 174 api.py routes into per-tag routers

- AST-extract all remaining routes from api.py into routers/*.py
- Group routes by OpenAPI tag (AI, Advanced, Agency, ..., x402)
- Keep existing auth/health/scraping/templates routers; add *_api.py for remaining routes of those tags
- Move shared helpers/models/variables with the routes that need them
- Update tests/test_api.py to import vision helpers from routers.vision
- Regenerate openapi.json (186 paths)
- All 497 tests pass; ruff clean

=1.14.0

@@ -1,15 +0,0 @@
-Collecting alembic
-  Using cached alembic-1.18.5-py3-none-any.whl.metadata (7.2 kB)
-Requirement already satisfied: SQLAlchemy>=1.4.23 in /tmp/test-pry/lib/python3.12/site-packages (from alembic) (2.0.51)
-Collecting Mako (from alembic)
-  Using cached mako-1.3.12-py3-none-any.whl.metadata (2.9 kB)
-Requirement already satisfied: typing-extensions>=4.12 in /tmp/test-pry/lib/python3.12/site-packages (from alembic) (4.16.0)
-Requirement already satisfied: greenlet>=1 in /tmp/test-pry/lib/python3.12/site-packages (from SQLAlchemy>=1.4.23->alembic) (3.5.3)
-Collecting MarkupSafe>=0.9.2 (from Mako->alembic)
-  Using cached markupsafe-3.0.3-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl.metadata (2.7 kB)
-Using cached alembic-1.18.5-py3-none-any.whl (264 kB)
-Using cached mako-1.3.12-py3-none-any.whl (78 kB)
-Using cached markupsafe-3.0.3-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (22 kB)
-Installing collected packages: MarkupSafe, Mako, alembic
-
-Successfully installed Mako-1.3.12 MarkupSafe-3.0.3 alembic-1.18.5

Dockerfile

@@ -41,7 +41,7 @@ COPY routers/ ./routers/
 COPY llm_providers/ ./llm_providers/
 COPY stealth_scripts/ ./stealth_scripts/
 COPY templates/ ./templates/
-RUN mkdir -p /app/sessions
+RUN mkdir -p /app/sessions /root/.pry
 
 EXPOSE 8002
 HEALTHCHECK --interval=15s --timeout=5s --start-period=10s --retries=3 \

STATUS.md

@@ -5,17 +5,17 @@
 > Where we are RIGHT NOW. Update before every commit.
 
 ## Last Updated
-2026-07-03 - Phase 0 router split + tests + Apify schema lib + PRY_API_KEY deployed
+2026-07-03 - Phase 0 infra hardening: metrics, redis/postgres, pinned deps, scrape metrics
 
 ## Current Status
-🟡 pre-production - core scraping/template routers split, e2e tests added, Talos PRY_API_KEY set, deploy pending.
+🟡 pre-production - router split deployed; infra hardening branch ready for review/deploy.
 
 ## Deployment Status
-- **Last deploy**: 2026-07-03
+- **Last deploy**: TBD
 - **Current version**: 3.0.0-phase0
-- **Health**: 🟢 healthy
+- **Health**: TBD
 - **Uptime (30d)**: TBD
-- **Active branch**: `feat/phase0-router-split` (PR #5 awaiting review)
+- **Active branch**: `feat/phase0-rebased` (reconciled on top of latest `main`)
 
 ## Open Issues
 - _None — track in forgejo issues_
@@ -27,15 +27,24 @@
 - 2026-07-03: Added e2e tests for scraping and template routers (23 new/passing)
 - 2026-07-03: Added reusable Apify actor schema builder (`apify_schema.py`)
 - 2026-07-03: Set `PRY_API_KEY` in Talos `/srv/pry/.env`; auth now active on restart
-- 2026-07-03: Registered `/v1/templates/batch` as paid x402 operation
-- 2026-07-03: Deployed to Talos `/srv/pry/`; pry container healthy on 127.0.0.1:8005
-- 2026-07-03: Opened PR #5: https://git.rugmunch.io/RugMunchMedia/pryscraper/pulls/5
+- 2026-07-03: Added `/v1/templates/batch` to x402 paid endpoints + pricing
+- 2026-07-03: Added `/metrics` endpoint + Prometheus counters/histograms for requests and scrapes
+- 2026-07-03: Added Redis + Postgres services to `docker-compose.yml`
+- 2026-07-03: Fixed Ollama URL to Talos (`100.104.130.92:11434`)
+- 2026-07-03: Added `LLMRegistry.complete_or_empty()` graceful fallback
+- 2026-07-03: Regenerated `requirements.txt`; added pinned `requirements.lock`
 
 ## Known Issues / Tech Debt
-- PR #5 must be reviewed/merged to bring `main` in sync with deployed code
+- Deploy at `/srv/pry/` is out of sync with repo (needs pull + restart to activate PRY_API_KEY, router changes, and metrics)
+- `feat/phase0-rebased` branch reconciles infra-hardening commits on top of latest `main`
 - pry-flaresolverr host port moved from 8191 to 8192 to avoid conflict with `rmi-flaresolverr`
 - All 80+ site templates unverified - only ~30-40% known to work end-to-end. Run templates/validate_templates.py before claiming template coverage.
-- State storage in ~/.pry/*.json - no concurrency safety, no transactions. See plan to migrate to SQLite.
+- State storage in ~/.pry/*.json - no concurrency safety, no transactions. See plan to migrate to SQLite/Postgres.
+- PRY_X402_WALLET, PRY_X402_FACILITATOR env vars unset in deploy - x402 payments not active.
+- License collision resolved 2026-07-03: dual MIT (core) + BSL 1.1 (stealth/anti-detection). See ADR-0002.
+- mcp_production.py and x402.py exist in repo but are NOT deployed. Top priority for next deploy.
+- All 80+ site templates unverified - only ~30-40% known to work end-to-end. Run templates/validate_templates.py before claiming template coverage.
+- State storage in ~/.pry/*.json - no concurrency safety, no transactions. See plan to migrate to SQLite/Postgres.
 - PRY_X402_WALLET, PRY_X402_FACILITATOR env vars unset in deploy - x402 payments not active.
 - License collision resolved 2026-07-03: dual MIT (core) + BSL 1.1 (stealth/anti-detection). See ADR-0002.
 - mcp_production.py and x402.py exist in repo but are NOT deployed. Top priority for next deploy.

alembic/env.py

@@ -4,16 +4,15 @@ from logging.config import fileConfig
 from sqlalchemy import engine_from_config, pool
 
 from alembic import context
-from db import Base
+from db import Base, _resolve_database_url
 
 config = context.config
 
 if config.config_file_name is not None:
     fileConfig(config.config_file_name)
 
-db_url = os.getenv("PRY_DATABASE_URL")
-if db_url is not None:
-    config.set_main_option("sqlalchemy.url", db_url)
+# Always use the same URL resolution logic as the application.
+config.set_main_option("sqlalchemy.url", os.getenv("PRY_DATABASE_URL") or _resolve_database_url())
 
 target_metadata = Base.metadata
 exclude_tables = {"spatial_ref_sys"}

alembic/versions/0001_initial_schema.py

@@ -5,6 +5,7 @@ Revises:
 Create Date: 2026-07-03 00:00:00.000000
 
 """
+
 from __future__ import annotations
 
 from collections.abc import Sequence
@@ -58,7 +59,9 @@ def upgrade() -> None:
         sa.Column("ts", sa.DateTime(), index=True),
         sa.PrimaryKeyConstraint("id"),
     )
-    op.create_index("ix_intel_competitor_ts", "intel_snapshots", ["competitor_id", sa.text("ts DESC")])
+    op.create_index(
+        "ix_intel_competitor_ts", "intel_snapshots", ["competitor_id", sa.text("ts DESC")]
+    )
     op.create_table(
         "costing_entries",
         sa.Column("id", sa.Integer(), nullable=False),

api.py

@@ -27,6 +27,7 @@ from errors import (
 )
 from mcp_production import make_fallback_server, register_all
 from mcp_sse import mcp_post_message, mcp_sse_endpoint
+from observability import REQUEST_COUNT, REQUEST_LATENCY, setup_tracing
 from pipeline import get_pipeline
 from routers.advanced import router as advanced_router
 from routers.agency import router as agency_router
@@ -105,8 +106,12 @@ except ImportError:
 async def lifespan(app: FastAPI) -> AsyncIterator[None]:
     """Startup: validate deps. Shutdown: cleanup clients."""
     logger.info("pry_startup", version="3.0.0")
+    setup_tracing()
     if not settings.api_key:
-        logger.warning("pry_api_key_unset", message="PRY_API_KEY is not set; all protected endpoints are publicly accessible")
+        logger.warning(
+            "pry_api_key_unset",
+            message="PRY_API_KEY is not set; all protected endpoints are publicly accessible",
+        )
     get_pipeline()  # Initialize pipeline
     yield
     logger.info("pry_shutdown")
@@ -274,8 +279,10 @@ class PryHttpMiddleware:
         start = time.time()
         request_id = self._get_header(scope, b"x-request-id") or uuid.uuid4().hex[:12]
         path = scope.get("path", "")
+        method = scope.get("method", "GET")
         client = scope.get("client")
         ip = client[0] if client else "unknown"
+        response_status = 200
 
         # Authentication check
         api_key = settings.api_key
@@ -349,7 +356,9 @@ class PryHttpMiddleware:
         request_id_bytes = request_id.encode()
 
         async def wrapped_send(message: dict[str, Any]) -> None:
+            nonlocal response_status
             if message["type"] == "http.response.start":
+                response_status = message.get("status", 200)
                 headers = list(message.get("headers", []))
                 headers.append((b"x-ratelimit-limit", ratelimit_limit))
                 headers.append((b"x-ratelimit-remaining", ratelimit_remaining))
@@ -361,12 +370,16 @@ class PryHttpMiddleware:
         await self.app(scope, receive, wrapped_send)
 
         elapsed = time.time() - start
+        if REQUEST_COUNT and REQUEST_LATENCY:
+            REQUEST_COUNT.labels(method=method, endpoint=path, status=str(response_status)).inc()
+            REQUEST_LATENCY.labels(endpoint=path).observe(elapsed)
         logger.info(
             "request_end",
             extra={
                 "request_id": request_id,
-                "method": scope.get("method", "GET"),
+                "method": method,
                 "path": path,
+                "status": response_status,
                 "duration": f"{elapsed:.3f}s",
             },
         )
@@ -719,6 +732,8 @@ app.include_router(vision_router)
 app.include_router(webhooks_router)
 app.include_router(x402_router)
 
+
+
 # ── Review ──
 
 

apify_schema.py

@@ -231,7 +231,9 @@ class ObjectField(FieldSpec):
 class ApifySchemaBuilder:
     """Fluent builder for Apify actor input/output schemas."""
 
-    def __init__(self, title: str = "Actor Input", description: str = "", schema_version: int = 1) -> None:
+    def __init__(
+        self, title: str = "Actor Input", description: str = "", schema_version: int = 1
+    ) -> None:
         self.title = title
         self.description = description
         self.schema_version = schema_version
@@ -284,7 +286,13 @@ class ApifySchemaBuilder:
         maximum: int | None = None,
     ) -> ApifySchemaBuilder:
         field = IntegerField(
-            name, title, description=description, default=default, nullable=nullable, minimum=minimum, maximum=maximum
+            name,
+            title,
+            description=description,
+            default=default,
+            nullable=nullable,
+            minimum=minimum,
+            maximum=maximum,
         )
         return self.add_field(field, required=required)
 
@@ -300,7 +308,13 @@ class ApifySchemaBuilder:
         maximum: float | None = None,
     ) -> ApifySchemaBuilder:
         field = NumberField(
-            name, title, description=description, default=default, nullable=nullable, minimum=minimum, maximum=maximum
+            name,
+            title,
+            description=description,
+            default=default,
+            nullable=nullable,
+            minimum=minimum,
+            maximum=maximum,
         )
         return self.add_field(field, required=required)
 
@@ -313,7 +327,9 @@ class ApifySchemaBuilder:
         default: bool | None = None,
         nullable: bool = False,
     ) -> ApifySchemaBuilder:
-        field = BooleanField(name, title, description=description, default=default, nullable=nullable)
+        field = BooleanField(
+            name, title, description=description, default=default, nullable=nullable
+        )
         return self.add_field(field, required=required)
 
     def add_enum(
@@ -326,7 +342,9 @@ class ApifySchemaBuilder:
         default: str | None = None,
         nullable: bool = False,
     ) -> ApifySchemaBuilder:
-        field = EnumField(name, title, enum, description=description, default=default, nullable=nullable)
+        field = EnumField(
+            name, title, enum, description=description, default=default, nullable=nullable
+        )
         return self.add_field(field, required=required)
 
     def add_array(
@@ -365,7 +383,13 @@ class ApifySchemaBuilder:
         nested_required: list[str] | None = None,
     ) -> ApifySchemaBuilder:
         field = ObjectField(
-            name, title, properties, description=description, default=default, nullable=nullable, required=nested_required
+            name,
+            title,
+            properties,
+            description=description,
+            default=default,
+            nullable=nullable,
+            required=nested_required,
         )
         return self.add_field(field, required=required)
 

cli.py

@@ -83,6 +83,7 @@ def _spinner(label: str):
             if stop:
                 break
             print(f"  {CYAN}{c}{NC} {label}...  ", end="\r", flush=True)
+            # CLI spinner runs in a daemon thread; sync sleep is acceptable here.
             time.sleep(0.1)
 
     t = threading.Thread(target=_spin, daemon=True)
@@ -196,6 +197,7 @@ def cmd_screenshot(url, output=None, timeout=30):
 def cmd_migrate():
     """Run pending database migrations (alembic upgrade head)."""
     from db import run_migrations
+
     run_migrations()
     print(f"{GREEN}✓{NC} Migrations up to date")
 

docker-compose.yml

@@ -10,6 +10,7 @@ services:
     ports:
       - "127.0.0.1:8005:8002"
     volumes:
+      - pry-data:/root/.pry
       - pry-sessions:/app/sessions
     env_file: .env
     environment:
@@ -77,4 +78,42 @@ services:
       - tor
 
 volumes:
+  pry-data:
   pry-sessions:
+
+  redis:
+    image: redis:7-alpine
+    container_name: pry-redis
+    restart: unless-stopped
+    ports:
+      - "127.0.0.1:6379:6379"
+    volumes:
+      - pry-redis:/data
+    deploy:
+      resources:
+        limits:
+          memory: 512M
+          cpus: "0.5"
+
+  postgres:
+    image: postgres:15-alpine
+    container_name: pry-postgres
+    restart: unless-stopped
+    ports:
+      - "127.0.0.1:5432:5432"
+    volumes:
+      - pry-postgres:/var/lib/postgresql/data
+    environment:
+      - POSTGRES_USER=pry
+      - POSTGRES_PASSWORD=pry
+      - POSTGRES_DB=pry
+    deploy:
+      resources:
+        limits:
+          memory: 512M
+          cpus: "0.5"
+
+volumes:
+  pry-sessions:
+  pry-redis:
+  pry-postgres:

llm_providers/registry.py

@@ -94,6 +94,35 @@ class LLMRegistry:
                     break
         raise Exception(f"All LLM providers failed. Last: {last_error}")
 
+    async def complete_or_empty(
+        self,
+        prompt: str,
+        system: str = "",
+        provider_name: str = "",
+        max_tokens: int = 1024,
+        temperature: float = 0.7,
+        model: str = "",
+        fallback: bool = True,
+    ) -> LLMResponse:
+        """Complete via LLM, returning an empty response on total failure.
+
+        Use this at API boundaries where a 500 is worse than degraded output.
+        """
+        try:
+            return await self.complete(
+                prompt, system, provider_name, max_tokens, temperature, model, fallback
+            )
+        except Exception as e:  # noqa: BLE001
+            logger.warning("llm_complete_failed_all_providers", extra={"error": str(e)[:120]})
+            return LLMResponse(
+                text="",
+                provider="none",
+                model="",
+                input_tokens=0,
+                output_tokens=0,
+                cost_usd=0.0,
+            )
+
     async def embed(self, text: str, provider_name: str = "", model: str = "") -> list[float]:
         names = [provider_name] if provider_name else list(self.fallback_chain)
         for name in names:

pyproject.toml

@@ -50,6 +50,9 @@ dependencies = [
     "structlog>=24.0.0",
     "sqlalchemy>=2.0.0",
     "aiosqlite>=0.19.0",
+    "prometheus-client>=0.21.0",
+    "opentelemetry-api>=1.29.0",
+    "opentelemetry-sdk>=1.29.0",
 ]
 
 [project.optional-dependencies]

requirements.lock

@@ -0,0 +1,31 @@
+# SPDX-License-Identifier: MIT
+# Copyright (c) 2026 Rug Munch Media LLC
+# Part of Pry — https://git.rugmunch.io/RugMunchMedia/pryscraper
+# Generated from pyproject.toml; pinned to currently installed versions.
+
+fastapi==0.128.8
+uvicorn==0.49.0
+trafilatura==2.1.0
+readability-lxml==0.8.4.1
+lxml==6.1.1
+httpx==0.28.1
+markdownify==1.2.3
+pydantic==2.12.5
+pydantic-settings==2.14.2
+playwright==1.60.0
+redis==8.0.0
+pypdf==6.14.2
+python-docx==1.2.0
+tiktoken==0.12.0
+pillow==12.2.0
+click==8.1.8
+pyyaml==6.0.3
+pandas==3.0.3
+anyio==4.14.1
+croniter==6.2.3
+structlog==26.1.0
+sqlalchemy==2.0.50
+aiosqlite==0.22.1
+prometheus-client==0.25.0
+opentelemetry-api==1.37.0
+opentelemetry-sdk==1.37.0

requirements.txt

@@ -27,3 +27,6 @@ structlog>=24.0.0
 sqlalchemy>=2.0.0
 aiosqlite>=0.19.0
 alembic>=1.14.0
+prometheus-client>=0.21.0
+opentelemetry-api>=1.29.0
+opentelemetry-sdk>=1.29.0

routers/scraping.py

@@ -32,6 +32,7 @@ from pydantic import BaseModel
 from client import get_client
 from deps import cache, extractor, queue, scraper
 from errors import ExternalServiceError, InvalidRequestError, PryError, ScrapeError
+from observability import track_scrape
 from scraper import BlockDetector
 from settings import settings
 
@@ -107,15 +108,16 @@ async def scrape(request: ScrapeRequest) -> dict[str, Any]:
         return cached
 
     try:
-        result = await scraper.scrape(
-            request.url,
-            {
-                "timeout": request.timeout,
-                "bypass_cloudflare": request.bypassCloudflare,
-                "js_render": request.jsRender,
-                "formats": request.formats,
-            },
-        )
+        with track_scrape(method="direct"):
+            result = await scraper.scrape(
+                request.url,
+                {
+                    "timeout": request.timeout,
+                    "bypass_cloudflare": request.bypassCloudflare,
+                    "js_render": request.jsRender,
+                    "formats": request.formats,
+                },
+            )
         if result.get("status") != "ok":
             raise ScrapeError(result.get("error", "Scrape failed"))
 
@@ -212,7 +214,8 @@ async def detect_lazy_content(
         generate_scroll_script,
     )
 
-    result = await scraper.scrape(url, {"bypass_cloudflare": True})
+    with track_scrape(method="lazy"):
+        result = await scraper.scrape(url, {"bypass_cloudflare": True})
     if result.get("status") != "ok":
         raise ScrapeError(result.get("error") or "Scrape failed")
 
@@ -255,14 +258,15 @@ async def crawl(request: CrawlRequest) -> dict[str, Any]:
         task.add_done_callback(_log_crawl_job_failure)
         return {"success": True, "data": {"id": job_id, "status": "pending"}}
 
-    pages = await scraper.crawl(
-        request.url,
-        {
-            "max_pages": request.maxPages,
-            "max_depth": request.maxDepth,
-            "timeout": request.scrapeOptions.get("timeout", 60) if request.scrapeOptions else 60,
-        },
-    )
+    with track_scrape(method="crawl"):
+        pages = await scraper.crawl(
+            request.url,
+            {
+                "max_pages": request.maxPages,
+                "max_depth": request.maxDepth,
+                "timeout": request.scrapeOptions.get("timeout", 60) if request.scrapeOptions else 60,
+            },
+        )
     return {"success": True, "data": {"id": "sync", "url": request.url, "pages": pages}}
 
 
@@ -272,7 +276,8 @@ async def crawl(request: CrawlRequest) -> dict[str, Any]:
 @router.post("/v1/map", summary="Discover URLs on a site")
 async def map_pages(request: MapRequest) -> dict[str, Any]:
     """Discover URLs on a site."""
-    urls = await scraper.map_urls(request.url, {"limit": request.limit})
+    with track_scrape(method="map"):
+        urls = await scraper.map_urls(request.url, {"limit": request.limit})
     return {"success": True, "data": {"links": urls}}
 
 
@@ -284,8 +289,9 @@ async def batch_scrape(urls: list[str] = Body(...), timeout: int = 30) -> dict[s
     """Scrape multiple URLs in parallel. Firecrawl charges extra for batch."""
     if len(urls) > 50:
         raise InvalidRequestError("Max 50 URLs per batch")
-    tasks = [scraper.scrape(u, {"timeout": timeout, "bypass_cloudflare": True}) for u in urls]
-    results = await asyncio.gather(*tasks, return_exceptions=True)
+    with track_scrape(method="batch"):
+        tasks = [scraper.scrape(u, {"timeout": timeout, "bypass_cloudflare": True}) for u in urls]
+        results = await asyncio.gather(*tasks, return_exceptions=True)
     pages = []
     for i, r in enumerate(results):
         if isinstance(r, BaseException):

routers/system.py

@@ -10,13 +10,22 @@ from __future__ import annotations
 
 import logging
 
-from fastapi import APIRouter
+from fastapi import APIRouter, Response
 from fastapi.responses import JSONResponse
 
+from observability import get_metrics_output
+
 logger = logging.getLogger(__name__)
 
 router = APIRouter(tags=["System"])
 
+@router.get("/metrics", tags=["System"], summary="Prometheus metrics", include_in_schema=False)
+async def metrics() -> Response:
+    """Expose Prometheus metrics for scraping."""
+    data, content_type = get_metrics_output()
+    return Response(content=data, media_type=content_type)
+
+
 @router.get(
     "/openapi.json",
     tags=["System"],

settings.py

@@ -25,7 +25,7 @@ class PrySettings(BaseSettings):
     port: int = 8002
 
     # Ollama LLM endpoint
-    ollama_url: str = "http://100.100.18.18:11434"
+    ollama_url: str = "http://100.104.130.92:11434"
 
     # FlareSolverr Cloudflare bypass endpoint
     flaresolverr_url: str = "http://flaresolverr:8191/v1"

tests/test_api.py

@@ -47,11 +47,17 @@ async def test_api_key_required_when_set(client: TestClient, monkeypatch) -> Non
     """Protected endpoints require a valid Bearer PRY_API_KEY."""
     monkeypatch.setattr("api.settings.api_key", "secret-pry-key")
     # Disable x402 payment gating so we can test auth in isolation.
-    with patch("x402_middleware.X402Middleware.__call__", lambda self, scope, receive, send: self.app(scope, receive, send)):
+    with patch(
+        "x402_middleware.X402Middleware.__call__",
+        lambda self, scope, receive, send: self.app(scope, receive, send),
+    ):
         resp = client.post("/v1/scrape", json={"url": "https://example.com"})
     assert resp.status_code == 401
 
-    with patch("x402_middleware.X402Middleware.__call__", lambda self, scope, receive, send: self.app(scope, receive, send)):
+    with patch(
+        "x402_middleware.X402Middleware.__call__",
+        lambda self, scope, receive, send: self.app(scope, receive, send),
+    ):
         resp = client.post(
             "/v1/scrape",
             json={"url": "https://example.com"},
@@ -59,7 +65,10 @@ async def test_api_key_required_when_set(client: TestClient, monkeypatch) -> Non
         )
     assert resp.status_code == 401
 
-    with patch("x402_middleware.X402Middleware.__call__", lambda self, scope, receive, send: self.app(scope, receive, send)):
+    with patch(
+        "x402_middleware.X402Middleware.__call__",
+        lambda self, scope, receive, send: self.app(scope, receive, send),
+    ):
         resp = client.post(
             "/v1/scrape",
             json={"url": "https://example.com"},
@@ -67,4 +76,3 @@ async def test_api_key_required_when_set(client: TestClient, monkeypatch) -> Non
         )
     # auth passes; endpoint returns 402 because we did not mock scraper, which is fine
     assert resp.status_code in (200, 402)
-

tests/test_api_scraping.py

@@ -11,6 +11,7 @@ TestClient. External I/O (HTTP requests, browser automation) is mocked at the
 deterministic.
 """
 
+from collections.abc import Iterator
 from unittest.mock import AsyncMock, patch
 
 import pytest
@@ -20,14 +21,17 @@ from api import app
 
 
 @pytest.fixture(autouse=True)
-def bypass_x402_middleware() -> None:
+def bypass_x402_middleware() -> Iterator[None]:
     """Disable x402 payment gating so scraping endpoints can be tested."""
-    with patch("x402_middleware.X402Middleware.__call__", lambda self, scope, receive, send: self.app(scope, receive, send)):
+    with patch(
+        "x402_middleware.X402Middleware.__call__",
+        lambda self, scope, receive, send: self.app(scope, receive, send),
+    ):
         yield
 
 
 @pytest.fixture(autouse=True)
-def clear_response_cache() -> None:
+def clear_response_cache() -> Iterator[None]:
     """Clear the shared response cache so tests don't see each other's data."""
     from deps import cache as response_cache
 
@@ -71,7 +75,10 @@ async def test_v1_scrape_success(client: TestClient, mock_scrape_result: dict) -
 async def test_v1_scrape_with_json_schema(client: TestClient, mock_scrape_result: dict) -> None:
     schema = {"title": "page title"}
     extracted = {"title": "Example"}
-    with patch("routers.scraping.scraper") as mock_scraper, patch("routers.scraping.extractor") as mock_extractor:
+    with (
+        patch("routers.scraping.scraper") as mock_scraper,
+        patch("routers.scraping.extractor") as mock_extractor,
+    ):
         mock_scraper.scrape = AsyncMock(return_value=mock_scrape_result)
         mock_extractor.extract = AsyncMock(return_value=extracted)
         resp = client.post("/v1/scrape", json={"url": "https://example.com", "jsonSchema": schema})
@@ -94,7 +101,11 @@ async def test_v1_detect_block_success(client: TestClient) -> None:
             )
         )
         mock_get_client.return_value = mock_client
-        resp = client.post("/v1/detect-block", content="\"https://example.com\"", headers={"content-type": "application/json"})
+        resp = client.post(
+            "/v1/detect-block",
+            content='"https://example.com"',
+            headers={"content-type": "application/json"},
+        )
 
     assert resp.status_code == 200
     data = resp.json()
@@ -121,7 +132,9 @@ async def test_v1_crawl_success(client: TestClient, mock_scrape_result: dict) ->
     pages = [mock_scrape_result]
     with patch("routers.scraping.scraper") as mock_scraper:
         mock_scraper.crawl = AsyncMock(return_value=pages)
-        resp = client.post("/v1/crawl", json={"url": "https://example.com", "maxPages": 2, "maxDepth": 1})
+        resp = client.post(
+            "/v1/crawl", json={"url": "https://example.com", "maxPages": 2, "maxDepth": 1}
+        )
 
     assert resp.status_code == 200
     data = resp.json()

tests/test_api_templates.py

@@ -5,6 +5,7 @@
 # Licensed under MIT.
 """End-to-end tests for the templates router endpoints."""
 
+from collections.abc import Iterator
 from unittest.mock import AsyncMock, patch
 
 import pytest
@@ -14,9 +15,12 @@ from api import app
 
 
 @pytest.fixture(autouse=True)
-def bypass_x402_middleware() -> None:
+def bypass_x402_middleware() -> Iterator[None]:
     """Disable x402 payment gating for template execution tests."""
-    with patch("x402_middleware.X402Middleware.__call__", lambda self, scope, receive, send: self.app(scope, receive, send)):
+    with patch(
+        "x402_middleware.X402Middleware.__call__",
+        lambda self, scope, receive, send: self.app(scope, receive, send),
+    ):
         yield
 
 
@@ -65,7 +69,10 @@ async def test_get_template_not_found(client: TestClient) -> None:
 async def test_execute_template(client: TestClient) -> None:
     result = {"success": True, "data": {"title": "Widget"}}
     with patch("routers.templates.execute_template", new=AsyncMock(return_value=result)):
-        resp = client.post("/v1/templates/execute", json={"template_id": "amazon_product", "url": "https://example.com"})
+        resp = client.post(
+            "/v1/templates/execute",
+            json={"template_id": "amazon_product", "url": "https://example.com"},
+        )
 
     assert resp.status_code == 200
     data = resp.json()

tests/test_apify_schema.py

@@ -63,10 +63,17 @@ def test_array_and_object_fields() -> None:
 
 
 def test_nullable_field() -> None:
-    schema = ApifySchemaBuilder().add_string("optional", "Optional", nullable=True).build_input_schema()
+    schema = (
+        ApifySchemaBuilder().add_string("optional", "Optional", nullable=True).build_input_schema()
+    )
     assert schema["properties"]["optional"]["type"] == ["string", "null"]
 
 
 def test_prefab() -> None:
-    schema = ApifySchemaBuilder().set_prefab("START_URLS").add_string("url", "URL", required=True).build_input_schema()
+    schema = (
+        ApifySchemaBuilder()
+        .set_prefab("START_URLS")
+        .add_string("url", "URL", required=True)
+        .build_input_schema()
+    )
     assert schema["prefab"] == "START_URLS"

ultimate_scraper.py

@@ -314,6 +314,11 @@ class UltimateScraper:
 
                 driver = uc.Chrome(headless=True, use_subprocess=True)
                 driver.get(url)
+                # Note: time.sleep is used inside a sync thread-pool task because
+                # undetected_chromedriver's blocking API runs in asyncio.to_thread.
+                # Replacing with asyncio.sleep would require restructuring the tier
+                # to be fully async; this sync sleep is bounded and does not block
+                # the event loop directly.
                 time.sleep(random.uniform(2, 4))
                 html = driver.page_source
                 driver.quit()