feat: implement scheduled/recurring analysis with change alerting

Add APScheduler-based background task that periodically re-analyzes
tracked companies and alerts on significant patent count changes.

- Add tracked_companies and alerts tables to database schema
- Add SPARC/scheduler.py with configurable interval and threshold
- Add admin endpoints: GET/POST/DELETE /admin/tracked, GET /admin/alerts
- Scheduler starts at app startup; interval via SCHEDULE_INTERVAL_HOURS
- Change threshold configurable via CHANGE_THRESHOLD_PERCENT env var
- apscheduler is optional; graceful fallback if not installed

Closes leeworks-agents/SPARC#22

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

.env.example

@@ -1,21 +1,42 @@
 # SPARC Configuration
 
+# ---- Application Environment ----
+# Set to "production" or "staging" in deployed environments.
+# The API will refuse to start with the default JWT secret unless APP_ENV=development.
+APP_ENV=development
+
+# ---- API Keys ----
+
 # SerpAPI key for patent search
 API_KEY=your_serpapi_key_here
 
 # OpenRouter API key for LLM analysis
 OPENROUTER_API_KEY=your_openrouter_key_here
 
-# Database configuration
+# ---- Database ----
-# All messages are stored in the database for persistence and caching
-DATABASE_URL=postgresql://postgres:postgres@localhost:5432/sparc
 
-# Cache configuration
+# PostgreSQL credentials (used by docker-compose)
-# When USE_CACHE=true: check database for cached responses before making API calls
+POSTGRES_USER=postgres
-# When USE_CACHE=false: always make fresh API calls (still stores results in database)
+POSTGRES_PASSWORD=change-me-to-a-secure-password
-# Default: true
+POSTGRES_DB=sparc
-USE_CACHE=true
 
-# JWT Secret for authentication
+# Full database URL (must match the credentials above)
+DATABASE_URL=postgresql://postgres:change-me-to-a-secure-password@localhost:5432/sparc
+
+# ---- Authentication ----
+
+# JWT Secret for signing tokens
 # IMPORTANT: Change this to a secure random string in production
 JWT_SECRET=your-secure-jwt-secret-change-in-production
+
+# ---- CORS ----
+
+# Comma-separated list of allowed origins for CORS
+# Defaults to http://localhost:3000,http://localhost:5173 when unset
+# CORS_ORIGINS=https://sparc.example.com,https://app.example.com
+
+# ---- Cache ----
+
+# When USE_CACHE=true: check database for cached responses before making API calls
+# When USE_CACHE=false: always make fresh API calls (still stores results in database)
+USE_CACHE=true

.gitea/workflows/build.yaml

@@ -9,7 +9,43 @@ on:
   workflow_dispatch:
 
 jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Install system dependencies
+        shell: sh
+        run: |
+          apk add --no-cache git python3 py3-pip gcc musl-dev libpq-dev python3-dev
+
+      - name: Checkout code
+        shell: sh
+        run: |
+          git clone http://gitea.gitea.svc.cluster.local/${{ gitea.repository }}.git .
+          git checkout ${{ gitea.sha }}
+
+      - name: Install Python dependencies
+        shell: sh
+        run: |
+          pip3 install --break-system-packages -r requirements.txt ruff
+
+      - name: Run ruff linter
+        shell: sh
+        run: |
+          ruff check SPARC/ tests/
+
+      - name: Run pytest
+        shell: sh
+        env:
+          DATABASE_URL: "sqlite://"
+          API_KEY: "test-key"
+          OPENROUTER_API_KEY: "test-key"
+          JWT_SECRET: "test-secret-for-ci"
+          APP_ENV: "development"
+        run: |
+          python3 -m pytest tests/ -v --tb=short -x
+
   build-api:
+    needs: test
     runs-on: ubuntu-latest
     steps:
       - name: Install dependencies
@@ -81,6 +117,7 @@ jobs:
           echo "API image available at ${{ steps.tags.outputs.IMAGE_TAG }}"
 
   build-frontend:
+    needs: test
     runs-on: ubuntu-latest
     steps:
       - name: Install dependencies

.gitea/workflows/test.yaml

@@ -0,0 +1,46 @@
+name: Test and Lint
+
+on:
+  push:
+    branches:
+      - main
+  pull_request:
+    branches:
+      - main
+  workflow_dispatch:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Install system dependencies
+        shell: sh
+        run: |
+          apk add --no-cache git python3 py3-pip gcc musl-dev libpq-dev python3-dev
+
+      - name: Checkout code
+        shell: sh
+        run: |
+          git clone http://gitea.gitea.svc.cluster.local/${{ gitea.repository }}.git .
+          git checkout ${{ gitea.sha }}
+
+      - name: Install Python dependencies
+        shell: sh
+        run: |
+          pip3 install --break-system-packages -r requirements.txt ruff
+
+      - name: Run ruff linter
+        shell: sh
+        run: |
+          ruff check SPARC/ tests/
+
+      - name: Run pytest
+        shell: sh
+        env:
+          DATABASE_URL: "sqlite://"
+          API_KEY: "test-key"
+          OPENROUTER_API_KEY: "test-key"
+          JWT_SECRET: "test-secret-for-ci"
+          APP_ENV: "development"
+        run: |
+          python3 -m pytest tests/ -v --tb=short -x

README.md

@@ -54,6 +54,21 @@ docker-compose up -d
 # - API Docs: http://localhost:8000/docs
 ```
 
+#### Patent PDF Storage
+
+The API stores downloaded patent PDFs in a `patents/` directory. In Docker,
+this is mounted as a bind mount (`./patents:/app/patents`) so that PDFs persist
+across container restarts.
+
+If you deploy to a different environment, ensure the `patents/` directory is a
+persistent volume. Without it, PDFs will be re-downloaded on every analysis.
+
+```yaml
+# docker-compose.yml excerpt
+volumes:
+  - ./patents:/app/patents
+```
+
 ### NixOS
 
 ```bash

SPARC/__init__.py

@@ -1,3 +1,4 @@
-from .types import Patents, Patent
+from .types import Patent as Patent
+from .types import Patents as Patents
 
-all = ["Patents", "Patent"]
+__all__ = ["Patents", "Patent"]

SPARC/analyzer.py

@@ -5,14 +5,17 @@ to provide company performance estimation based on patent portfolios.
 """
 
 import hashlib
+import logging
 from concurrent.futures import ThreadPoolExecutor, as_completed
 from typing import Callable
 
 from SPARC import config
+
+logger = logging.getLogger(__name__)
 from SPARC.database import DatabaseClient
-from SPARC.serp_api import SERP
 from SPARC.llm import LLMAnalyzer
-from SPARC.types import Patent, Patents, CompanyAnalysisResult, BatchAnalysisResult
+from SPARC.serp_api import SERP
+from SPARC.types import BatchAnalysisResult, CompanyAnalysisResult, Patent, Patents
 
 
 class CompanyAnalyzer:
@@ -52,13 +55,13 @@ class CompanyAnalyzer:
             query_hash = hashlib.sha256(company_name.lower().encode()).hexdigest()
             cached_ids = self.db.get_cached_serp_query(query_hash)
             if cached_ids is not None:
-                print(f"Using cached SERP results for {company_name} ({len(cached_ids)} patents)")
+                logger.info("Using cached SERP results for %s (%d patents)", company_name, len(cached_ids))
                 patents = Patents(patents=[
                     Patent(patent_id=pid, pdf_link="")
                     for pid in cached_ids
                 ])
             else:
-                print(f"Retrieving patents for {company_name}...")
+                logger.info("Retrieving patents for %s...", company_name)
                 patents = SERP.query(company_name)
                 # Cache the SERP results
                 if patents.patents:
@@ -66,12 +69,13 @@ class CompanyAnalyzer:
                         company_name=company_name,
                         query_hash=query_hash,
                         patent_ids=[p.patent_id for p in patents.patents],
+                        ttl_hours=config.serp_cache_ttl_hours,
                     )
 
         if not patents.patents:
             return f"No patents found for {company_name}"
 
-        print(f"Found {len(patents.patents)} patents. Processing...")
+        logger.info("Found %d patents. Processing...", len(patents.patents))
 
         # Download, parse, and minimize patents in parallel
         processed_patents = []
@@ -87,12 +91,12 @@ class CompanyAnalyzer:
                     if result:
                         processed_patents.append(result)
                 except Exception as e:
-                    print(f"Warning: Failed to process {patent.patent_id}: {e}")
+                    logger.warning("Failed to process %s: %s", patent.patent_id, e)
 
         if not processed_patents:
             return f"Failed to process any patents for {company_name}"
 
-        print(f"Analyzing portfolio with LLM...")
+        logger.info("Analyzing portfolio with LLM...")
 
         # Analyze the full portfolio with LLM
         analysis = self.llm_analyzer.analyze_patent_portfolio(
@@ -104,21 +108,34 @@ class CompanyAnalyzer:
     def analyze_single_patent(self, patent_id: str, company_name: str) -> str:
         """Analyze a single patent by ID.
 
-        Useful for focused analysis of specific innovations.
+        Prerequisite:
+            The patent PDF must already exist at ``patents/{patent_id}.pdf``
+            before calling this method. PDFs are downloaded automatically when
+            using the batch analysis pipeline (``analyze_company`` or the
+            ``/analyze/batch`` API endpoint). For standalone usage, download
+            the PDF manually or call ``SERP.save_patents()`` first.
 
         Args:
-          patent_id: Publication ID of the patent
+          patent_id: Publication ID of the patent (e.g. "US-11234567-B2")
           company_name: Name of the company (for context)
 
         Returns:
           Analysis of the specific patent's innovation quality
+
+        Raises:
+          FileNotFoundError: If the patent PDF is not found at the expected path.
         """
-        # Note: This simplified version assumes the patent PDF is already downloaded
+        import os
-        # A more complete implementation would support direct patent ID lookup
+        logger.info("Analyzing patent %s for %s...", patent_id, company_name)
-        print(f"Analyzing patent {patent_id} for {company_name}...")
 
         patent_path = f"patents/{patent_id}.pdf"
 
+        if not os.path.exists(patent_path):
+            raise FileNotFoundError(
+                f"Patent PDF not found at '{patent_path}'. "
+                f"Download the PDF first using SERP.save_patents() or the batch analysis pipeline."
+            )
+
         try:
             sections = SERP.parse_patent_pdf(patent_path)
             minimized_content = SERP.minimize_patent_for_llm(sections)
@@ -129,6 +146,8 @@ class CompanyAnalyzer:
 
             return analysis
 
+        except FileNotFoundError:
+            raise
         except Exception as e:
             return f"Failed to analyze patent {patent_id}: {e}"
 
@@ -169,7 +188,7 @@ class CompanyAnalyzer:
 
             return {"patent_id": patent.patent_id, "content": minimized_content}
         except Exception as e:
-            print(f"Warning: Failed to process {patent.patent_id}: {e}")
+            logger.warning("Failed to process %s: %s", patent.patent_id, e)
             return None
 
     def _analyze_company_safe(self, company_name: str) -> CompanyAnalysisResult:
@@ -240,7 +259,7 @@ class CompanyAnalyzer:
         results: list[CompanyAnalysisResult] = []
         total = len(companies)
 
-        print(f"Starting batch analysis of {total} companies...")
+        logger.info("Starting batch analysis of %d companies...", total)
 
         with ThreadPoolExecutor(max_workers=max_workers) as executor:
             future_to_company = {
@@ -257,8 +276,8 @@ class CompanyAnalyzer:
                     result = future.result()
                     results.append(result)
 
-                    status = "✓" if result.success else "✗"
+                    status = "OK" if result.success else "FAIL"
-                    print(f"[{completed}/{total}] {status} {company}")
+                    logger.info("[%d/%d] %s %s", completed, total, status, company)
 
                     if progress_callback:
                         progress_callback(company, completed, total)
@@ -273,12 +292,12 @@ class CompanyAnalyzer:
                             error=str(e),
                         )
                     )
-                    print(f"[{completed}/{total}] ✗ {company}: {e}")
+                    logger.error("[%d/%d] FAIL %s: %s", completed, total, company, e)
 
         successful = sum(1 for r in results if r.success)
         failed = total - successful
 
-        print(f"\nBatch complete: {successful} succeeded, {failed} failed")
+        logger.info("Batch complete: %d succeeded, %d failed", successful, failed)
 
         return BatchAnalysisResult(
             results=results,
@@ -304,20 +323,20 @@ class CompanyAnalyzer:
         results: list[CompanyAnalysisResult] = []
         total = len(companies)
 
-        print(f"Starting sequential analysis of {total} companies...")
+        logger.info("Starting sequential analysis of %d companies...", total)
 
         for idx, company in enumerate(companies, 1):
-            print(f"\n[{idx}/{total}] Analyzing {company}...")
+            logger.info("[%d/%d] Analyzing %s...", idx, total, company)
             result = self._analyze_company_safe(company)
             results.append(result)
 
-            status = "✓" if result.success else "✗"
+            status = "OK" if result.success else "FAIL"
-            print(f"[{idx}/{total}] {status} {company}")
+            logger.info("[%d/%d] %s %s", idx, total, status, company)
 
         successful = sum(1 for r in results if r.success)
         failed = total - successful
 
-        print(f"\nBatch complete: {successful} succeeded, {failed} failed")
+        logger.info("Batch complete: %d succeeded, %d failed", successful, failed)
 
         return BatchAnalysisResult(
             results=results,

SPARC/api.py

@@ -20,11 +20,14 @@ from SPARC.analyzer import CompanyAnalyzer
 from SPARC.auth import (
     TokenResponse,
     UserResponse,
+    check_jwt_secret,
+    close_db_client,
     create_tokens,
     decode_token,
     get_current_admin,
     get_current_user,
     get_db_client,
+    init_db_client,
 )
 from SPARC.types import BatchAnalysisResult, CompanyAnalysisResult
 
@@ -118,8 +121,7 @@ class AnalyticsResponse(BaseModel):
     period_days: int
 
 
-# In-memory job storage (for demo; production would use Redis/DB)
+# Job counter for generating unique IDs (the actual state is in PostgreSQL)
-_jobs: dict[str, JobStatus] = {}
 _job_counter = 0
 
 
@@ -152,12 +154,28 @@ _analyzer: CompanyAnalyzer | None = None
 
 @asynccontextmanager
 async def lifespan(app: FastAPI):
-    """Initialize resources on startup."""
+    """Initialize resources on startup, clean up on shutdown."""
     global _analyzer
+    check_jwt_secret()
+    init_db_client()
     _analyzer = CompanyAnalyzer()
+    # Mark any jobs that were running/pending before the restart as failed
+    from SPARC.database import DatabaseClient
+    _db = DatabaseClient(config.database_url)
+    _db.connect()
+    _db.initialize_schema()
+    stale = _db.mark_stale_jobs_failed()
+    if stale:
+        import logging
+        logging.getLogger(__name__).warning("Marked %d stale jobs as failed on startup", stale)
+    _db.close()
+    # Start scheduled analysis if tracked companies are configured
+    from SPARC.scheduler import start_scheduler
+    start_scheduler()
     yield
-    # Cleanup if needed
+    # Cleanup
     _analyzer = None
+    close_db_client()
 
 
 app = FastAPI(
@@ -187,7 +205,7 @@ async def rate_limit_handler(request: Request, exc: RateLimitExceeded):
 # Add CORS middleware for React frontend
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["http://localhost:3000", "http://localhost:5173"],
+    allow_origins=config.cors_origins,
     allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
@@ -354,6 +372,60 @@ async def delete_user(
     return {"message": "User deleted"}
 
 
+# ============== Tracked Companies Endpoints ==============
+
+
+class TrackCompanyRequest(BaseModel):
+    """Request to add a company to tracking."""
+
+    company_name: str = Field(..., min_length=1, max_length=255)
+
+
+@app.get("/admin/tracked", tags=["Admin"])
+async def list_tracked_companies(
+    _: UserResponse = Depends(get_current_admin),
+):
+    """List all tracked companies (admin only)."""
+    db = get_db_client()
+    return db.list_tracked_companies()
+
+
+@app.post("/admin/tracked", tags=["Admin"])
+async def add_tracked_company(
+    request: TrackCompanyRequest,
+    _: UserResponse = Depends(get_current_admin),
+):
+    """Add a company to the tracked list (admin only)."""
+    db = get_db_client()
+    result = db.add_tracked_company(request.company_name)
+    if not result:
+        raise HTTPException(status_code=409, detail="Company already tracked")
+    return result
+
+
+@app.delete("/admin/tracked/{company_name}", tags=["Admin"])
+async def remove_tracked_company(
+    company_name: str,
+    _: UserResponse = Depends(get_current_admin),
+):
+    """Remove a company from the tracked list (admin only)."""
+    db = get_db_client()
+    removed = db.remove_tracked_company(company_name)
+    if not removed:
+        raise HTTPException(status_code=404, detail="Company not found in tracking list")
+    return {"message": f"Stopped tracking {company_name}"}
+
+
+@app.get("/admin/alerts", tags=["Admin"])
+async def list_alerts(
+    limit: int = Query(default=50, ge=1, le=200),
+    _: UserResponse = Depends(get_current_admin),
+):
+    """List recent alerts from scheduled analysis (admin only)."""
+    db = get_db_client()
+    return db.list_alerts(limit=limit)
+
+
 # ============== Analytics Endpoint ==============
 
 
@@ -444,20 +516,52 @@ async def analyze_companies_batch(
     return _convert_batch_result(result)
 
 
+def _get_job_db() -> "DatabaseClient":
+    """Get a DatabaseClient for job persistence."""
+    from SPARC.database import DatabaseClient
+    db = DatabaseClient(config.database_url)
+    return db
+
+
+def _job_row_to_status(row: dict) -> JobStatus:
+    """Convert a database job row to a JobStatus model."""
+    import json as _json
+    result = None
+    if row.get("result_json"):
+        result_data = row["result_json"]
+        if isinstance(result_data, str):
+            result_data = _json.loads(result_data)
+        result = BatchAnalysisResponse(**result_data)
+    return JobStatus(
+        job_id=row["job_id"],
+        status=row["status"],
+        progress=row["progress"],
+        total_companies=row["total_companies"],
+        completed_companies=row["completed_companies"],
+        result=result,
+        error=row.get("error"),
+    )
+
+
 def _run_batch_job(job_id: str, companies: list[str], max_workers: int):
     """Background task for batch analysis."""
-    global _jobs, _analyzer
+    import json as _json
+    global _analyzer
+
+    db = _get_job_db()
 
     if not _analyzer:
-        _jobs[job_id].status = "failed"
+        db.update_job(job_id, status="failed", error="Analyzer not initialized")
-        _jobs[job_id].error = "Analyzer not initialized"
         return
 
-    _jobs[job_id].status = "running"
+    db.update_job(job_id, status="running")
 
     def progress_callback(company: str, completed: int, total: int):
-        _jobs[job_id].completed_companies = completed
+        db.update_job(
-        _jobs[job_id].progress = int((completed / total) * 100)
+            job_id,
+            completed_companies=completed,
+            progress=int((completed / total) * 100),
+        )
 
     try:
         result = _analyzer.analyze_companies(
@@ -465,12 +569,15 @@ def _run_batch_job(job_id: str, companies: list[str], max_workers: int):
             max_workers=max_workers,
             progress_callback=progress_callback,
         )
-        _jobs[job_id].status = "completed"
+        batch_response = _convert_batch_result(result)
-        _jobs[job_id].progress = 100
+        db.update_job(
-        _jobs[job_id].result = _convert_batch_result(result)
+            job_id,
+            status="completed",
+            progress=100,
+            result_json=_json.dumps(batch_response.model_dump(), default=str),
+        )
     except Exception as e:
-        _jobs[job_id].status = "failed"
+        db.update_job(job_id, status="failed", error=str(e))
-        _jobs[job_id].error = str(e)
 
 
 @app.post("/analyze/batch/async", response_model=JobStatus, tags=["Analysis"])
@@ -495,19 +602,14 @@ async def analyze_companies_async(
     _job_counter += 1
     job_id = f"job_{_job_counter}_{datetime.now().strftime('%Y%m%d%H%M%S')}"
 
-    _jobs[job_id] = JobStatus(
+    db = _get_job_db()
-        job_id=job_id,
+    job_row = db.create_job(job_id=job_id, total_companies=len(request.companies))
-        status="pending",
-        progress=0,
-        total_companies=len(request.companies),
-        completed_companies=0,
-    )
 
     background_tasks.add_task(
         _run_batch_job, job_id, request.companies, request.max_workers
     )
 
-    return _jobs[job_id]
+    return _job_row_to_status(job_row)
 
 
 @app.get("/jobs/{job_id}", response_model=JobStatus, tags=["Jobs"])
@@ -523,10 +625,13 @@ async def get_job_status(
     Returns:
         Current job status including progress and results when complete
     """
-    if job_id not in _jobs:
+    db = _get_job_db()
+    job_row = db.get_job(job_id)
+
+    if not job_row:
         raise HTTPException(status_code=404, detail=f"Job {job_id} not found")
 
-    return _jobs[job_id]
+    return _job_row_to_status(job_row)
 
 
 @app.get("/jobs", response_model=list[JobStatus], tags=["Jobs"])
@@ -547,12 +652,6 @@ async def list_jobs(
     Returns:
         List of job statuses
     """
-    jobs = list(_jobs.values())
+    db = _get_job_db()
-
+    job_rows = db.list_jobs(status=status, limit=limit)
-    if status:
+    return [_job_row_to_status(row) for row in job_rows]
-        jobs = [j for j in jobs if j.status == status]
-
-    # Return most recent first
-    jobs.sort(key=lambda j: j.job_id, reverse=True)
-
-    return jobs[:limit]

SPARC/auth.py

@@ -13,11 +13,25 @@ from SPARC import config
 from SPARC.database import DatabaseClient
 
 # JWT Configuration
-JWT_SECRET = os.getenv("JWT_SECRET", "sparc-secret-key-change-in-production")
+_DEFAULT_JWT_SECRET = "sparc-secret-key-change-in-production"
+JWT_SECRET = os.getenv("JWT_SECRET", _DEFAULT_JWT_SECRET)
 JWT_ALGORITHM = "HS256"
 ACCESS_TOKEN_EXPIRE_MINUTES = 30
 REFRESH_TOKEN_EXPIRE_DAYS = 7
 
+
+def check_jwt_secret() -> None:
+    """Refuse to start with the default JWT secret in non-development environments.
+
+    Raises:
+        RuntimeError: If JWT_SECRET is the default value and APP_ENV is not 'development'.
+    """
+    if JWT_SECRET == _DEFAULT_JWT_SECRET and config.app_env != "development":
+        raise RuntimeError(
+            f"FATAL: JWT_SECRET is set to the default value and APP_ENV={config.app_env!r}. "
+            "Set a secure JWT_SECRET environment variable before running in non-development environments."
+        )
+
 security = HTTPBearer()
 
 
@@ -132,11 +146,36 @@ def decode_token(token: str) -> Optional[TokenPayload]:
         return None
 
 
+# Shared database client singleton, initialized at startup via init_db_client()
+_db_client: DatabaseClient | None = None
+
+
+def init_db_client() -> None:
+    """Initialize the shared database client. Call once at app startup."""
+    global _db_client
+    _db_client = DatabaseClient(config.database_url)
+    _db_client.connect()
+
+
+def close_db_client() -> None:
+    """Close the shared database client. Call at app shutdown."""
+    global _db_client
+    if _db_client:
+        _db_client.close()
+        _db_client = None
+
+
 def get_db_client() -> DatabaseClient:
-    """Get database client for auth operations."""
+    """Get the shared pooled database client for auth operations.
-    client = DatabaseClient(config.database_url)
+
-    client.connect()
+    Returns the module-level singleton DatabaseClient. If not yet initialized
-    return client
+    (e.g., during tests), creates a new instance as a fallback.
+    """
+    global _db_client
+    if _db_client is None:
+        _db_client = DatabaseClient(config.database_url)
+        _db_client.connect()
+    return _db_client
 
 
 async def get_current_user(

SPARC/config.py

@@ -2,11 +2,20 @@
 
 Loads environment variables from .env file for API keys and other secrets.
 """
-from dotenv import load_dotenv
+import logging
 import os
 
+from dotenv import load_dotenv
+
 load_dotenv()
 
+# Logging configuration
+log_level = os.getenv("LOG_LEVEL", "INFO").upper()
+logging.basicConfig(
+    level=getattr(logging, log_level, logging.INFO),
+    format="%(asctime)s %(levelname)s %(name)s %(message)s",
+)
+
 # SerpAPI key for patent search
 api_key = os.getenv("API_KEY")
 
@@ -30,6 +39,25 @@ use_database = os.getenv("USE_DATABASE", "false").lower() in ("true", "1", "yes"
 patent_search_days = int(os.getenv("PATENT_SEARCH_DAYS", "90"))
 patent_thread_workers = int(os.getenv("PATENT_THREAD_WORKERS", "5"))
 
+# LLM model to use via OpenRouter (e.g. "anthropic/claude-3.5-sonnet", "openai/gpt-4o")
+model = os.getenv("MODEL", "anthropic/claude-3.5-sonnet")
+
+# SERP cache TTL in hours (how long cached search results are considered fresh)
+serp_cache_ttl_hours = int(os.getenv("SERP_CACHE_TTL_HOURS", "24"))
+
 # Root path for running behind a reverse proxy (e.g., "/api" when served at /api/)
 # This ensures OpenAPI docs work correctly when accessed via the proxy
 root_path = os.getenv("ROOT_PATH", "")
+
+# Application environment: "development", "staging", or "production"
+# Used for safety checks (e.g., refusing default JWT secret in production)
+app_env = os.getenv("APP_ENV", "development")
+
+# CORS allowed origins (comma-separated)
+# Defaults to localhost dev origins when unset
+_cors_origins_raw = os.getenv("CORS_ORIGINS", "")
+cors_origins: list[str] = (
+    [o.strip() for o in _cors_origins_raw.split(",") if o.strip()]
+    if _cors_origins_raw
+    else ["http://localhost:3000", "http://localhost:5173"]
+)

SPARC/database.py

@@ -1,14 +1,15 @@
 """Database client for storing and retrieving LLM messages and user authentication."""
 
 import contextlib
-import psycopg2
-from psycopg2.pool import ThreadedConnectionPool
-from psycopg2.extras import RealDictCursor
-from typing import Dict, List, Optional
-from datetime import datetime, timedelta
-import json
 import hashlib
+import json
+from datetime import datetime, timedelta
+from typing import Dict, List, Optional
+
 import bcrypt
+import psycopg2
+from psycopg2.extras import RealDictCursor
+from psycopg2.pool import ThreadedConnectionPool
 
 
 class DatabaseClient:
@@ -171,6 +172,55 @@ class DatabaseClient:
                 ON serp_queries(query_hash)
             """)
 
+            # Create jobs table for persisting async batch job state
+            cursor.execute("""
+                CREATE TABLE IF NOT EXISTS jobs (
+                    job_id VARCHAR(128) PRIMARY KEY,
+                    status VARCHAR(20) NOT NULL DEFAULT 'pending',
+                    progress INTEGER NOT NULL DEFAULT 0,
+                    total_companies INTEGER NOT NULL DEFAULT 0,
+                    completed_companies INTEGER NOT NULL DEFAULT 0,
+                    result_json JSONB,
+                    error TEXT,
+                    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
+                    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+                )
+            """)
+
+            cursor.execute("""
+                CREATE INDEX IF NOT EXISTS idx_jobs_status
+                ON jobs(status)
+            """)
+
+            # Create tracked companies table for scheduled analysis
+            cursor.execute("""
+                CREATE TABLE IF NOT EXISTS tracked_companies (
+                    id SERIAL PRIMARY KEY,
+                    company_name VARCHAR(255) UNIQUE NOT NULL,
+                    last_patent_count INTEGER DEFAULT 0,
+                    last_analysis_at TIMESTAMP,
+                    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+                )
+            """)
+
+            # Create alerts table for significant changes
+            cursor.execute("""
+                CREATE TABLE IF NOT EXISTS alerts (
+                    id SERIAL PRIMARY KEY,
+                    company_name VARCHAR(255) NOT NULL,
+                    alert_type VARCHAR(50) NOT NULL,
+                    message TEXT NOT NULL,
+                    old_value NUMERIC,
+                    new_value NUMERIC,
+                    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+                )
+            """)
+
+            cursor.execute("""
+                CREATE INDEX IF NOT EXISTS idx_alerts_company
+                ON alerts(company_name)
+            """)
+
             self.conn.commit()
 
     @staticmethod
@@ -201,8 +251,6 @@ class DatabaseClient:
         Returns:
             Cached message dict if found, None otherwise
         """
-        self.connect()
-
         prompt_hash = self.hash_prompt(prompt)
 
         query = """
@@ -225,10 +273,11 @@ class DatabaseClient:
 
         query += " ORDER BY timestamp DESC LIMIT 1"
 
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+        with self.get_conn() as conn:
-            cursor.execute(query, params)
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
-            result = cursor.fetchone()
+                cursor.execute(query, params)
-            return dict(result) if result else None
+                result = cursor.fetchone()
+                return dict(result) if result else None
 
     def store_message(
         self,
@@ -256,33 +305,32 @@ class DatabaseClient:
         Returns:
             The ID of the inserted record
         """
-        self.connect()
-
         prompt_hash = self.hash_prompt(prompt)
 
-        with self.conn.cursor() as cursor:
+        with self.get_conn() as conn:
-            cursor.execute(
+            with conn.cursor() as cursor:
-                """
+                cursor.execute(
-                INSERT INTO llm_messages
+                    """
-                (prompt, prompt_hash, response, company_name, analysis_type, model, metadata, token_usage, is_cached)
+                    INSERT INTO llm_messages
-                VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s)
+                    (prompt, prompt_hash, response, company_name, analysis_type, model, metadata, token_usage, is_cached)
-                RETURNING id
+                    VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s)
-                """,
+                    RETURNING id
-                (
+                    """,
-                    prompt,
+                    (
-                    prompt_hash,
+                        prompt,
-                    response,
+                        prompt_hash,
-                    company_name,
+                        response,
-                    analysis_type,
+                        company_name,
-                    model,
+                        analysis_type,
-                    json.dumps(metadata) if metadata else None,
+                        model,
-                    json.dumps(token_usage) if token_usage else None,
+                        json.dumps(metadata) if metadata else None,
-                    is_cached,
+                        json.dumps(token_usage) if token_usage else None,
-                ),
+                        is_cached,
-            )
+                    ),
+                )
 
-            message_id = cursor.fetchone()[0]
+                message_id = cursor.fetchone()[0]
-            self.conn.commit()
+            conn.commit()
 
             return message_id
 
@@ -304,8 +352,6 @@ class DatabaseClient:
         Returns:
             List of message dictionaries
         """
-        self.connect()
-
         query = "SELECT * FROM llm_messages WHERE 1=1"
         params = []
 
@@ -320,9 +366,10 @@ class DatabaseClient:
         query += " ORDER BY timestamp DESC LIMIT %s OFFSET %s"
         params.extend([limit, offset])
 
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+        with self.get_conn() as conn:
-            cursor.execute(query, params)
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
-            return [dict(row) for row in cursor.fetchall()]
+                cursor.execute(query, params)
+                return [dict(row) for row in cursor.fetchall()]
 
     def get_analytics(self, days: int = 30) -> Dict:
         """Get analytics on message usage.
@@ -333,53 +380,52 @@ class DatabaseClient:
         Returns:
             Dictionary with analytics data
         """
-        self.connect()
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                # Total messages
+                cursor.execute(
+                    """
+                    SELECT COUNT(*) as total_messages
+                    FROM llm_messages
+                    WHERE timestamp >= NOW() - INTERVAL '%s days'
+                    """,
+                    (days,),
+                )
+                total = cursor.fetchone()["total_messages"]
 
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                # Messages by company
-            # Total messages
+                cursor.execute(
-            cursor.execute(
+                    """
-                """
+                    SELECT company_name, COUNT(*) as count
-                SELECT COUNT(*) as total_messages
+                    FROM llm_messages
-                FROM llm_messages
+                    WHERE timestamp >= NOW() - INTERVAL '%s days'
-                WHERE timestamp >= NOW() - INTERVAL '%s days'
+                    GROUP BY company_name
-                """,
+                    ORDER BY count DESC
-                (days,),
+                    LIMIT 10
-            )
+                    """,
-            total = cursor.fetchone()["total_messages"]
+                    (days,),
+                )
+                by_company = cursor.fetchall()
 
-            # Messages by company
+                # Messages by type
-            cursor.execute(
+                cursor.execute(
-                """
+                    """
-                SELECT company_name, COUNT(*) as count
+                    SELECT analysis_type, COUNT(*) as count
-                FROM llm_messages
+                    FROM llm_messages
-                WHERE timestamp >= NOW() - INTERVAL '%s days'
+                    WHERE timestamp >= NOW() - INTERVAL '%s days'
-                GROUP BY company_name
+                    GROUP BY analysis_type
-                ORDER BY count DESC
+                    ORDER BY count DESC
-                LIMIT 10
+                    """,
-                """,
+                    (days,),
-                (days,),
+                )
-            )
+                by_type = cursor.fetchall()
-            by_company = cursor.fetchall()
 
-            # Messages by type
+                return {
-            cursor.execute(
+                    "total_messages": total,
-                """
+                    "by_company": [dict(row) for row in by_company],
-                SELECT analysis_type, COUNT(*) as count
+                    "by_type": [dict(row) for row in by_type],
-                FROM llm_messages
+                    "period_days": days,
-                WHERE timestamp >= NOW() - INTERVAL '%s days'
+                }
-                GROUP BY analysis_type
-                ORDER BY count DESC
-                """,
-                (days,),
-            )
-            by_type = cursor.fetchall()
-
-            return {
-                "total_messages": total,
-                "by_company": [dict(row) for row in by_company],
-                "by_type": [dict(row) for row in by_type],
-                "period_days": days,
-            }
 
     # Patent Cache Methods
 
@@ -462,6 +508,131 @@ class DatabaseClient:
                 )
             conn.commit()
 
+    # Job Persistence Methods
+
+    def create_job(
+        self,
+        job_id: str,
+        total_companies: int,
+    ) -> Dict:
+        """Create a new job record.
+
+        Args:
+            job_id: Unique job identifier
+            total_companies: Number of companies in the batch
+
+        Returns:
+            Job dict
+        """
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
+                    """
+                    INSERT INTO jobs (job_id, status, progress, total_companies, completed_companies)
+                    VALUES (%s, 'pending', 0, %s, 0)
+                    RETURNING *
+                    """,
+                    (job_id, total_companies),
+                )
+                job = cursor.fetchone()
+            conn.commit()
+            return dict(job)
+
+    def update_job(
+        self,
+        job_id: str,
+        status: Optional[str] = None,
+        progress: Optional[int] = None,
+        completed_companies: Optional[int] = None,
+        result_json: Optional[str] = None,
+        error: Optional[str] = None,
+    ) -> Optional[Dict]:
+        """Update a job's state.
+
+        Only non-None fields are updated.
+        """
+        updates = []
+        params = []
+        if status is not None:
+            updates.append("status = %s")
+            params.append(status)
+        if progress is not None:
+            updates.append("progress = %s")
+            params.append(progress)
+        if completed_companies is not None:
+            updates.append("completed_companies = %s")
+            params.append(completed_companies)
+        if result_json is not None:
+            updates.append("result_json = %s")
+            params.append(result_json)
+        if error is not None:
+            updates.append("error = %s")
+            params.append(error)
+
+        if not updates:
+            return self.get_job(job_id)
+
+        updates.append("updated_at = CURRENT_TIMESTAMP")
+        params.append(job_id)
+
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
+                    f"UPDATE jobs SET {', '.join(updates)} WHERE job_id = %s RETURNING *",
+                    params,
+                )
+                job = cursor.fetchone()
+            conn.commit()
+            return dict(job) if job else None
+
+    def get_job(self, job_id: str) -> Optional[Dict]:
+        """Get a job by ID."""
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute("SELECT * FROM jobs WHERE job_id = %s", (job_id,))
+                job = cursor.fetchone()
+                return dict(job) if job else None
+
+    def list_jobs(
+        self,
+        status: Optional[str] = None,
+        limit: int = 10,
+    ) -> List[Dict]:
+        """List jobs, optionally filtered by status."""
+        query = "SELECT * FROM jobs"
+        params: list = []
+        if status:
+            query += " WHERE status = %s"
+            params.append(status)
+        query += " ORDER BY created_at DESC LIMIT %s"
+        params.append(limit)
+
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(query, params)
+                return [dict(row) for row in cursor.fetchall()]
+
+    def mark_stale_jobs_failed(self) -> int:
+        """Mark any jobs in 'running' or 'pending' state as 'failed'.
+
+        Called at startup to clean up jobs that were interrupted by a restart.
+
+        Returns:
+            Number of jobs marked as failed.
+        """
+        with self.get_conn() as conn:
+            with conn.cursor() as cursor:
+                cursor.execute(
+                    """
+                    UPDATE jobs SET status = 'failed', error = 'Interrupted by server restart',
+                    updated_at = CURRENT_TIMESTAMP
+                    WHERE status IN ('running', 'pending')
+                    """
+                )
+                count = cursor.rowcount
+            conn.commit()
+            return count
+
     # User Authentication Methods
 
     @staticmethod
@@ -505,25 +676,23 @@ class DatabaseClient:
         Returns:
             Created user dict or None if email exists
         """
-        self.connect()
-
         password_hash = self.hash_password(password)
 
         try:
-            with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+            with self.get_conn() as conn:
-                cursor.execute(
+                with conn.cursor(cursor_factory=RealDictCursor) as cursor:
-                    """
+                    cursor.execute(
-                    INSERT INTO users (email, password_hash, role)
+                        """
-                    VALUES (%s, %s, %s)
+                        INSERT INTO users (email, password_hash, role)
-                    RETURNING id, email, role, created_at
+                        VALUES (%s, %s, %s)
-                    """,
+                        RETURNING id, email, role, created_at
-                    (email, password_hash, role),
+                        """,
-                )
+                        (email, password_hash, role),
-                user = cursor.fetchone()
+                    )
-                self.conn.commit()
+                    user = cursor.fetchone()
+                conn.commit()
                 return dict(user) if user else None
         except psycopg2.errors.UniqueViolation:
-            self.conn.rollback()
             return None
 
     def authenticate_user(self, email: str, password: str) -> Optional[Dict]:
@@ -536,23 +705,22 @@ class DatabaseClient:
         Returns:
             User dict if authenticated, None otherwise
         """
-        self.connect()
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
+                    "SELECT * FROM users WHERE email = %s",
+                    (email,),
+                )
+                user = cursor.fetchone()
 
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                if user and self.verify_password(password, user["password_hash"]):
-            cursor.execute(
+                    return {
-                "SELECT * FROM users WHERE email = %s",
+                        "id": user["id"],
-                (email,),
+                        "email": user["email"],
-            )
+                        "role": user["role"],
-            user = cursor.fetchone()
+                        "created_at": user["created_at"],
-
+                    }
-            if user and self.verify_password(password, user["password_hash"]):
+                return None
-                return {
-                    "id": user["id"],
-                    "email": user["email"],
-                    "role": user["role"],
-                    "created_at": user["created_at"],
-                }
-            return None
 
     def get_user_by_id(self, user_id: int) -> Optional[Dict]:
         """Get a user by ID.
@@ -563,15 +731,14 @@ class DatabaseClient:
         Returns:
             User dict or None
         """
-        self.connect()
+        with self.get_conn() as conn:
-
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
-            cursor.execute(
+                    "SELECT id, email, role, created_at FROM users WHERE id = %s",
-                "SELECT id, email, role, created_at FROM users WHERE id = %s",
+                    (user_id,),
-                (user_id,),
+                )
-            )
+                user = cursor.fetchone()
-            user = cursor.fetchone()
+                return dict(user) if user else None
-            return dict(user) if user else None
 
     def get_user_by_email(self, email: str) -> Optional[Dict]:
         """Get a user by email.
@@ -582,15 +749,14 @@ class DatabaseClient:
         Returns:
             User dict or None
         """
-        self.connect()
+        with self.get_conn() as conn:
-
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
-            cursor.execute(
+                    "SELECT id, email, role, created_at FROM users WHERE email = %s",
-                "SELECT id, email, role, created_at FROM users WHERE email = %s",
+                    (email,),
-                (email,),
+                )
-            )
+                user = cursor.fetchone()
-            user = cursor.fetchone()
+                return dict(user) if user else None
-            return dict(user) if user else None
 
     def get_all_users(self, limit: int = 100, offset: int = 0) -> List[Dict]:
         """Get all users (admin only).
@@ -602,19 +768,18 @@ class DatabaseClient:
         Returns:
             List of user dicts
         """
-        self.connect()
+        with self.get_conn() as conn:
-
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
-            cursor.execute(
+                    """
-                """
+                    SELECT id, email, role, created_at
-                SELECT id, email, role, created_at
+                    FROM users
-                FROM users
+                    ORDER BY created_at DESC
-                ORDER BY created_at DESC
+                    LIMIT %s OFFSET %s
-                LIMIT %s OFFSET %s
+                    """,
-                """,
+                    (limit, offset),
-                (limit, offset),
+                )
-            )
+                return [dict(row) for row in cursor.fetchall()]
-            return [dict(row) for row in cursor.fetchall()]
 
     def update_user_role(self, user_id: int, role: str) -> Optional[Dict]:
         """Update a user's role (admin only).
@@ -626,20 +791,19 @@ class DatabaseClient:
         Returns:
             Updated user dict or None
         """
-        self.connect()
+        with self.get_conn() as conn:
-
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
-        with self.conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
-            cursor.execute(
+                    """
-                """
+                    UPDATE users
-                UPDATE users
+                    SET role = %s, updated_at = CURRENT_TIMESTAMP
-                SET role = %s, updated_at = CURRENT_TIMESTAMP
+                    WHERE id = %s
-                WHERE id = %s
+                    RETURNING id, email, role, created_at
-                RETURNING id, email, role, created_at
+                    """,
-                """,
+                    (role, user_id),
-                (role, user_id),
+                )
-            )
+                user = cursor.fetchone()
-            user = cursor.fetchone()
+            conn.commit()
-            self.conn.commit()
             return dict(user) if user else None
 
     def delete_user(self, user_id: int) -> bool:
@@ -651,12 +815,11 @@ class DatabaseClient:
         Returns:
             True if deleted
         """
-        self.connect()
+        with self.get_conn() as conn:
-
+            with conn.cursor() as cursor:
-        with self.conn.cursor() as cursor:
+                cursor.execute("DELETE FROM users WHERE id = %s", (user_id,))
-            cursor.execute("DELETE FROM users WHERE id = %s", (user_id,))
+                deleted = cursor.rowcount > 0
-            deleted = cursor.rowcount > 0
+            conn.commit()
-            self.conn.commit()
             return deleted
 
     def get_user_count(self) -> int:
@@ -665,8 +828,85 @@ class DatabaseClient:
         Returns:
             Number of users
         """
-        self.connect()
+        with self.get_conn() as conn:
+            with conn.cursor() as cursor:
+                cursor.execute("SELECT COUNT(*) FROM users")
+                return cursor.fetchone()[0]
 
-        with self.conn.cursor() as cursor:
+    # Tracked Companies Methods
-            cursor.execute("SELECT COUNT(*) FROM users")
+
-            return cursor.fetchone()[0]
+    def add_tracked_company(self, company_name: str) -> Optional[Dict]:
+        """Add a company to the tracking list."""
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                try:
+                    cursor.execute(
+                        "INSERT INTO tracked_companies (company_name) VALUES (%s) RETURNING *",
+                        (company_name,),
+                    )
+                    row = cursor.fetchone()
+                    conn.commit()
+                    return dict(row) if row else None
+                except Exception:
+                    conn.rollback()
+                    return None
+
+    def remove_tracked_company(self, company_name: str) -> bool:
+        """Remove a company from the tracking list."""
+        with self.get_conn() as conn:
+            with conn.cursor() as cursor:
+                cursor.execute(
+                    "DELETE FROM tracked_companies WHERE LOWER(company_name) = LOWER(%s)",
+                    (company_name,),
+                )
+                conn.commit()
+                return cursor.rowcount > 0
+
+    def list_tracked_companies(self) -> List[Dict]:
+        """List all tracked companies."""
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute("SELECT * FROM tracked_companies ORDER BY company_name")
+                return [dict(row) for row in cursor.fetchall()]
+
+    def update_tracked_company(
+        self, company_name: str, patent_count: int
+    ) -> None:
+        """Update the last analysis stats for a tracked company."""
+        with self.get_conn() as conn:
+            with conn.cursor() as cursor:
+                cursor.execute(
+                    """UPDATE tracked_companies
+                       SET last_patent_count = %s, last_analysis_at = CURRENT_TIMESTAMP
+                       WHERE LOWER(company_name) = LOWER(%s)""",
+                    (patent_count, company_name),
+                )
+                conn.commit()
+
+    def store_alert(
+        self,
+        company_name: str,
+        alert_type: str,
+        message: str,
+        old_value: float | None = None,
+        new_value: float | None = None,
+    ) -> None:
+        """Record an alert for a significant change."""
+        with self.get_conn() as conn:
+            with conn.cursor() as cursor:
+                cursor.execute(
+                    """INSERT INTO alerts (company_name, alert_type, message, old_value, new_value)
+                       VALUES (%s, %s, %s, %s, %s)""",
+                    (company_name, alert_type, message, old_value, new_value),
+                )
+                conn.commit()
+
+    def list_alerts(self, limit: int = 50) -> List[Dict]:
+        """List recent alerts."""
+        with self.get_conn() as conn:
+            with conn.cursor(cursor_factory=RealDictCursor) as cursor:
+                cursor.execute(
+                    "SELECT * FROM alerts ORDER BY created_at DESC LIMIT %s",
+                    (limit,),
+                )
+                return [dict(row) for row in cursor.fetchall()]

SPARC/llm.py

@@ -1,9 +1,14 @@
 """LLM integration for patent analysis using OpenRouter."""
 
+import logging
+from typing import Dict
+
 from openai import OpenAI
+
 from SPARC import config
 from SPARC.database import DatabaseClient
-from typing import Dict
+
+logger = logging.getLogger(__name__)
 
 
 class LLMAnalyzer:
@@ -20,7 +25,7 @@ class LLMAnalyzer:
         """
         self.test_mode = test_mode
         self.use_cache = use_cache if use_cache is not None else config.use_cache
-        self.model = "anthropic/claude-3.5-sonnet"
+        self.model = config.model
 
         # Always initialize database client for storage and caching
         self.db_client = DatabaseClient(config.database_url)
@@ -59,11 +64,7 @@ Patent Content:
 Provide a concise analysis (2-3 paragraphs) focusing on what this patent reveals about the company's technical direction and competitive advantage."""
 
         if self.test_mode:
-            print("=" * 80)
+            logger.debug("TEST MODE - Prompt that would be sent to LLM:\n%s", prompt)
-            print("TEST MODE - Prompt that would be sent to LLM:")
-            print("=" * 80)
-            print(prompt)
-            print("=" * 80)
             return "[TEST MODE - No API call made]"
 
         # Check cache first
@@ -165,7 +166,7 @@ Patent Portfolio:
 Provide a comprehensive analysis (4-5 paragraphs) with a final verdict on the company's innovation strength and performance outlook."""
 
         if self.test_mode:
-            print(prompt)
+            logger.debug("TEST MODE - Portfolio prompt:\n%s", prompt)
             return "[TEST MODE]"
 
         metadata = {

SPARC/scheduler.py

@@ -0,0 +1,109 @@
+"""Scheduled patent analysis for tracked companies.
+
+Uses APScheduler to periodically re-analyze tracked companies and
+detect significant changes in patent counts.
+"""
+
+import logging
+import os
+
+from SPARC import config
+from SPARC.analyzer import CompanyAnalyzer
+from SPARC.database import DatabaseClient
+
+logger = logging.getLogger(__name__)
+
+# Configurable via environment variable (in hours, default 24)
+SCHEDULE_INTERVAL_HOURS = int(os.getenv("SCHEDULE_INTERVAL_HOURS", "24"))
+
+# Patent count change threshold (percentage) to trigger an alert
+CHANGE_THRESHOLD_PERCENT = int(os.getenv("CHANGE_THRESHOLD_PERCENT", "20"))
+
+
+def run_scheduled_analysis() -> None:
+    """Re-analyze all tracked companies and check for significant changes."""
+    db = DatabaseClient(config.database_url)
+    db.connect()
+    db.initialize_schema()
+
+    tracked = db.list_tracked_companies()
+    if not tracked:
+        logger.info("No tracked companies configured; skipping scheduled analysis")
+        return
+
+    logger.info("Running scheduled analysis for %d tracked companies", len(tracked))
+
+    analyzer = CompanyAnalyzer(db_client=db)
+
+    for company_row in tracked:
+        name = company_row["company_name"]
+        old_count = company_row.get("last_patent_count", 0) or 0
+
+        try:
+            result = analyzer._analyze_company_safe(name)
+
+            if result.success:
+                new_count = result.patent_count
+
+                # Update tracking record
+                db.update_tracked_company(name, new_count)
+
+                # Check for significant change
+                if old_count > 0:
+                    delta_pct = abs(new_count - old_count) / old_count * 100
+                    if delta_pct >= CHANGE_THRESHOLD_PERCENT:
+                        direction = "increased" if new_count > old_count else "decreased"
+                        message = (
+                            f"Patent count for {name} {direction} by {delta_pct:.0f}% "
+                            f"({old_count} -> {new_count})"
+                        )
+                        logger.warning("ALERT: %s", message)
+                        db.store_alert(
+                            company_name=name,
+                            alert_type="patent_count_change",
+                            message=message,
+                            old_value=old_count,
+                            new_value=new_count,
+                        )
+                elif new_count > 0:
+                    # First analysis -- record baseline
+                    logger.info("Baseline for %s: %d patents", name, new_count)
+            else:
+                logger.warning("Scheduled analysis failed for %s: %s", name, result.error)
+
+        except Exception as e:
+            logger.error("Error analyzing tracked company %s: %s", name, e)
+
+    db.close()
+    logger.info("Scheduled analysis complete")
+
+
+def start_scheduler() -> None:
+    """Start the APScheduler background scheduler.
+
+    Safe to call at application startup. If apscheduler is not installed,
+    the function logs a warning and returns without starting anything.
+    """
+    try:
+        from apscheduler.schedulers.background import BackgroundScheduler
+    except ImportError:
+        logger.warning(
+            "apscheduler not installed; scheduled analysis disabled. "
+            "Install with: pip install apscheduler"
+        )
+        return
+
+    scheduler = BackgroundScheduler()
+    scheduler.add_job(
+        run_scheduled_analysis,
+        "interval",
+        hours=SCHEDULE_INTERVAL_HOURS,
+        id="scheduled_patent_analysis",
+        replace_existing=True,
+    )
+    scheduler.start()
+    logger.info(
+        "Scheduled patent analysis started (every %d hours, threshold %d%%)",
+        SCHEDULE_INTERVAL_HOURS,
+        CHANGE_THRESHOLD_PERCENT,
+    )

SPARC/serp_api.py

@@ -1,12 +1,15 @@
 import os
-import serpapi
-from SPARC import config
 import re
-import pdfplumber  # pip install pdfplumber
-import requests
 from datetime import datetime, timedelta
 from typing import Dict
-from SPARC.types import Patents, Patent
+
+import pdfplumber  # pip install pdfplumber
+import requests
+import serpapi
+
+from SPARC import config
+from SPARC.types import Patent, Patents
+
 
 class SERP:
   def query(company: str, days_back: int = None) -> Patents:

SPARC/types.py

@@ -4,7 +4,7 @@ from datetime import datetime
 
 @dataclass
 class Patent:
-    patent_id: int
+    patent_id: str
     pdf_link: str
     pdf_path: str | None = None
     summary: dict | None = None

docker-compose.yml

@@ -3,15 +3,15 @@ services:
     image: postgres:16-alpine
     container_name: sparc-postgres
     environment:
-      POSTGRES_USER: postgres
+      POSTGRES_USER: ${POSTGRES_USER}
-      POSTGRES_PASSWORD: postgres
+      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
-      POSTGRES_DB: sparc
+      POSTGRES_DB: ${POSTGRES_DB}
     ports:
       - "5432:5432"
     volumes:
       - postgres_data:/var/lib/postgresql/data
     healthcheck:
-      test: ["CMD-SHELL", "pg_isready -U postgres"]
+      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER}"]
       interval: 5s
       timeout: 5s
       retries: 5
@@ -22,7 +22,7 @@ services:
     container_name: sparc-init-db
     command: python scripts/init_database.py
     environment:
-      DATABASE_URL: postgresql://postgres:postgres@postgres:5432/sparc
+      DATABASE_URL: postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@postgres:5432/${POSTGRES_DB}
     depends_on:
       postgres:
         condition: service_healthy
@@ -35,9 +35,11 @@ services:
     environment:
       API_KEY: ${API_KEY}
       OPENROUTER_API_KEY: ${OPENROUTER_API_KEY}
-      DATABASE_URL: postgresql://postgres:postgres@postgres:5432/sparc
+      DATABASE_URL: postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@postgres:5432/${POSTGRES_DB}
       USE_CACHE: "true"
       JWT_SECRET: ${JWT_SECRET:-sparc-secret-key-change-in-production}
+      CORS_ORIGINS: ${CORS_ORIGINS:-}
+      APP_ENV: ${APP_ENV:-development}
       ROOT_PATH: /api
     ports:
       - "8000:8000"

requirements.txt

@@ -15,3 +15,4 @@ pandas
 bcrypt
 PyJWT
 slowapi
+apscheduler

ruff.toml

@@ -0,0 +1,8 @@
+[lint]
+select = ["E", "F", "I"]
+ignore = [
+    "E501",  # line too long (handled by formatter)
+]
+
+[lint.per-file-ignores]
+"tests/*" = ["E402", "F841"]  # allow import not at top of file, unused vars (mocks) in tests

scripts/init_database.py

@@ -40,6 +40,9 @@ def main():
         print("\nTables created:")
         print("  - llm_messages: Stores all LLM prompts and responses")
         print("  - users: Stores user accounts")
+        print("  - jobs: Stores async batch job state")
+        print("  - patents: Patent PDF cache")
+        print("  - serp_queries: SERP query result cache")
         print("\nIndexes created:")
         print("  - idx_messages_timestamp: For time-based queries")
         print("  - idx_messages_company: For company-specific queries")

tests/test_analyzer.py

@@ -1,9 +1,11 @@
 """Tests for the high-level company analyzer orchestration."""
 
+from unittest.mock import MagicMock, Mock
+
 import pytest
-from unittest.mock import Mock, patch, call, MagicMock
+
 from SPARC.analyzer import CompanyAnalyzer
-from SPARC.types import Patent, Patents, CompanyAnalysisResult, BatchAnalysisResult
+from SPARC.types import BatchAnalysisResult, Patent, Patents
 
 
 @pytest.fixture(autouse=True)
@@ -24,7 +26,7 @@ class TestCompanyAnalyzer:
         """Test analyzer initialization with API key."""
         mock_llm = mocker.patch("SPARC.analyzer.LLMAnalyzer")
 
-        analyzer = CompanyAnalyzer(openrouter_api_key="test-key")
+        _analyzer = CompanyAnalyzer(openrouter_api_key="test-key")  # noqa: F841
 
         mock_llm.assert_called_once_with(api_key="test-key")
 

tests/test_api.py

@@ -1,12 +1,13 @@
 """Tests for FastAPI web service endpoints."""
 
-import pytest
 from datetime import datetime
-from unittest.mock import Mock, patch
+from unittest.mock import Mock
+
+import pytest
 from fastapi.testclient import TestClient
 
-from SPARC.api import app, _analyzer, _jobs
+from SPARC.api import app
-from SPARC.types import CompanyAnalysisResult, BatchAnalysisResult
+from SPARC.types import BatchAnalysisResult, CompanyAnalysisResult
 
 
 @pytest.fixture

tests/test_auth.py

@@ -0,0 +1,302 @@
+"""Tests for JWT authentication flow: register, login, protected routes, refresh, admin access."""
+
+from datetime import datetime, timezone
+from unittest.mock import MagicMock, patch
+
+import pytest
+from fastapi.testclient import TestClient
+
+from SPARC.api import app
+from SPARC.auth import create_access_token, create_refresh_token
+
+
+@pytest.fixture
+def client():
+    """Create test client."""
+    return TestClient(app)
+
+
+@pytest.fixture(autouse=True)
+def mock_db(monkeypatch):
+    """Mock the database client used by auth endpoints.
+
+    Returns a MagicMock with all DB methods pre-configured.
+    """
+    db = MagicMock()
+
+    # Default: no users exist
+    db.get_user_count.return_value = 0
+    db.get_user_by_id.return_value = None
+    db.get_user_by_email.return_value = None
+    db.authenticate_user.return_value = None
+    db.create_user.return_value = None
+    db.get_all_users.return_value = []
+    db.update_user_role.return_value = None
+    db.delete_user.return_value = False
+
+    with patch("SPARC.api.get_db_client", return_value=db), \
+         patch("SPARC.auth.get_db_client", return_value=db):
+        yield db
+
+
+def _make_admin_user():
+    return {
+        "id": 1,
+        "email": "admin@test.com",
+        "role": "admin",
+        "created_at": datetime(2025, 1, 1, tzinfo=timezone.utc),
+    }
+
+
+def _make_regular_user():
+    return {
+        "id": 2,
+        "email": "user@test.com",
+        "role": "user",
+        "created_at": datetime(2025, 1, 1, tzinfo=timezone.utc),
+    }
+
+
+def _auth_header(user_dict):
+    """Create an Authorization header with a valid access token for the given user."""
+    token = create_access_token(user_dict["id"], user_dict["email"], user_dict["role"])
+    return {"Authorization": f"Bearer {token}"}
+
+
+class TestRegister:
+    """POST /auth/register"""
+
+    def test_register_first_user_becomes_admin(self, client, mock_db):
+        """First registered user should get admin role."""
+        mock_db.get_user_count.return_value = 0
+        mock_db.create_user.return_value = {
+            "id": 1,
+            "email": "admin@test.com",
+            "role": "admin",
+            "created_at": datetime(2025, 1, 1, tzinfo=timezone.utc),
+        }
+
+        response = client.post(
+            "/auth/register",
+            json={"email": "admin@test.com", "password": "securepass123"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["email"] == "admin@test.com"
+        assert data["role"] == "admin"
+        mock_db.create_user.assert_called_once_with(
+            email="admin@test.com", password="securepass123", role="admin"
+        )
+
+    def test_register_subsequent_user_gets_user_role(self, client, mock_db):
+        """Non-first user should get regular user role."""
+        mock_db.get_user_count.return_value = 1
+        mock_db.create_user.return_value = _make_regular_user()
+
+        response = client.post(
+            "/auth/register",
+            json={"email": "user@test.com", "password": "securepass123"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["role"] == "user"
+
+    def test_register_duplicate_email_returns_400(self, client, mock_db):
+        """Registering with an existing email should return 400."""
+        mock_db.get_user_count.return_value = 1
+        mock_db.create_user.return_value = None  # indicates duplicate
+
+        response = client.post(
+            "/auth/register",
+            json={"email": "existing@test.com", "password": "securepass123"},
+        )
+
+        assert response.status_code == 400
+        assert "already registered" in response.json()["detail"].lower()
+
+
+class TestLogin:
+    """POST /auth/login"""
+
+    def test_login_valid_credentials_returns_tokens(self, client, mock_db):
+        """Valid credentials should return access and refresh tokens."""
+        user = _make_regular_user()
+        mock_db.authenticate_user.return_value = user
+
+        response = client.post(
+            "/auth/login",
+            json={"email": "user@test.com", "password": "correctpassword"},
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert "access_token" in data
+        assert "refresh_token" in data
+        assert data["token_type"] == "bearer"
+
+    def test_login_invalid_credentials_returns_401(self, client, mock_db):
+        """Invalid credentials should return 401."""
+        mock_db.authenticate_user.return_value = None
+
+        response = client.post(
+            "/auth/login",
+            json={"email": "user@test.com", "password": "wrongpassword"},
+        )
+
+        assert response.status_code == 401
+        assert "invalid" in response.json()["detail"].lower()
+
+
+class TestGetMe:
+    """GET /auth/me"""
+
+    def test_valid_access_token_returns_user(self, client, mock_db):
+        """A valid access token should return the user's data."""
+        user = _make_regular_user()
+        mock_db.get_user_by_id.return_value = user
+
+        response = client.get("/auth/me", headers=_auth_header(user))
+
+        assert response.status_code == 200
+        data = response.json()
+        assert data["email"] == "user@test.com"
+        assert data["id"] == 2
+
+    def test_missing_token_returns_401(self, client):
+        """No token should return 401 (403 from HTTPBearer)."""
+        response = client.get("/auth/me")
+        assert response.status_code in (401, 403)
+
+    def test_expired_token_returns_401(self, client, mock_db):
+        """An expired token should return 401."""
+        # Create a token that has already expired
+        from datetime import timedelta
+
+        import jwt as pyjwt
+        from SPARC.auth import JWT_ALGORITHM, JWT_SECRET
+
+        payload = {
+            "sub": "1",
+            "email": "user@test.com",
+            "role": "user",
+            "exp": datetime.now(timezone.utc) - timedelta(hours=1),
+            "type": "access",
+        }
+        expired_token = pyjwt.encode(payload, JWT_SECRET, algorithm=JWT_ALGORITHM)
+
+        response = client.get(
+            "/auth/me", headers={"Authorization": f"Bearer {expired_token}"}
+        )
+        assert response.status_code == 401
+
+    def test_refresh_token_as_access_returns_401(self, client, mock_db):
+        """Using a refresh token as an access token should return 401."""
+        user = _make_regular_user()
+        refresh_token = create_refresh_token(user["id"], user["email"], user["role"])
+
+        response = client.get(
+            "/auth/me", headers={"Authorization": f"Bearer {refresh_token}"}
+        )
+        assert response.status_code == 401
+
+
+class TestRefreshToken:
+    """POST /auth/refresh"""
+
+    def test_valid_refresh_token_returns_new_tokens(self, client, mock_db):
+        """A valid refresh token should issue new access and refresh tokens."""
+        user = _make_regular_user()
+        mock_db.get_user_by_id.return_value = user
+        refresh = create_refresh_token(user["id"], user["email"], user["role"])
+
+        response = client.post(
+            "/auth/refresh", json={"refresh_token": refresh}
+        )
+
+        assert response.status_code == 200
+        data = response.json()
+        assert "access_token" in data
+        assert "refresh_token" in data
+
+    def test_invalid_refresh_token_returns_401(self, client, mock_db):
+        """An invalid refresh token should return 401."""
+        response = client.post(
+            "/auth/refresh", json={"refresh_token": "invalid-token-string"}
+        )
+        assert response.status_code == 401
+
+    def test_access_token_as_refresh_returns_401(self, client, mock_db):
+        """Using an access token as a refresh token should return 401."""
+        user = _make_regular_user()
+        access = create_access_token(user["id"], user["email"], user["role"])
+
+        response = client.post(
+            "/auth/refresh", json={"refresh_token": access}
+        )
+        assert response.status_code == 401
+
+
+class TestAdminUsers:
+    """GET /admin/users and PATCH /admin/users/{id}/role"""
+
+    def test_admin_can_list_users(self, client, mock_db):
+        """Admin token should allow listing users."""
+        admin = _make_admin_user()
+        mock_db.get_user_by_id.return_value = admin
+        mock_db.get_all_users.return_value = [admin, _make_regular_user()]
+
+        response = client.get("/admin/users", headers=_auth_header(admin))
+
+        assert response.status_code == 200
+        data = response.json()
+        assert len(data) == 2
+
+    def test_regular_user_cannot_list_users(self, client, mock_db):
+        """Regular user token should be rejected with 403."""
+        user = _make_regular_user()
+        mock_db.get_user_by_id.return_value = user
+
+        response = client.get("/admin/users", headers=_auth_header(user))
+
+        assert response.status_code == 403
+
+    def test_no_token_cannot_list_users(self, client):
+        """No token should be rejected."""
+        response = client.get("/admin/users")
+        assert response.status_code in (401, 403)
+
+    def test_admin_can_change_user_role(self, client, mock_db):
+        """Admin should be able to change another user's role."""
+        admin = _make_admin_user()
+        mock_db.get_user_by_id.return_value = admin
+        mock_db.update_user_role.return_value = {
+            "id": 2,
+            "email": "user@test.com",
+            "role": "admin",
+            "created_at": datetime(2025, 1, 1, tzinfo=timezone.utc),
+        }
+
+        response = client.patch(
+            "/admin/users/2/role",
+            json={"role": "admin"},
+            headers=_auth_header(admin),
+        )
+
+        assert response.status_code == 200
+        assert response.json()["role"] == "admin"
+
+    def test_admin_cannot_change_own_role(self, client, mock_db):
+        """Admin should not be able to change their own role."""
+        admin = _make_admin_user()
+        mock_db.get_user_by_id.return_value = admin
+
+        response = client.patch(
+            "/admin/users/1/role",
+            json={"role": "user"},
+            headers=_auth_header(admin),
+        )
+
+        assert response.status_code == 400
+        assert "own role" in response.json()["detail"].lower()

tests/test_llm.py

@@ -1,7 +1,9 @@
 """Tests for LLM analysis functionality."""
 
+from unittest.mock import Mock
+
 import pytest
-from unittest.mock import Mock, MagicMock, patch
+
 from SPARC.llm import LLMAnalyzer
 
 

tests/test_security.py

@@ -0,0 +1,116 @@
+"""Tests for security hardening: JWT secret startup check, CORS config, credential handling."""
+
+import os
+from unittest.mock import patch
+
+import pytest
+
+
+class TestJWTSecretStartupCheck:
+    """Test the startup guard that refuses default JWT secret in non-dev environments."""
+
+    def test_default_secret_in_production_raises(self):
+        """Starting with default secret and APP_ENV=production must raise RuntimeError."""
+        with patch.dict(os.environ, {"APP_ENV": "production"}):
+            # Reload config to pick up the new APP_ENV
+            import importlib
+            import SPARC.config
+            importlib.reload(SPARC.config)
+
+            from SPARC.auth import _DEFAULT_JWT_SECRET, check_jwt_secret
+            # Patch JWT_SECRET to the default
+            with patch("SPARC.auth.JWT_SECRET", _DEFAULT_JWT_SECRET):
+                with pytest.raises(RuntimeError, match="FATAL.*JWT_SECRET"):
+                    check_jwt_secret()
+
+            # Restore config
+            with patch.dict(os.environ, {"APP_ENV": "development"}):
+                importlib.reload(SPARC.config)
+
+    def test_default_secret_in_development_succeeds(self):
+        """Starting with default secret and APP_ENV=development must not raise."""
+        with patch.dict(os.environ, {"APP_ENV": "development"}):
+            import importlib
+            import SPARC.config
+            importlib.reload(SPARC.config)
+
+            from SPARC.auth import _DEFAULT_JWT_SECRET, check_jwt_secret
+            with patch("SPARC.auth.JWT_SECRET", _DEFAULT_JWT_SECRET):
+                # Should not raise
+                check_jwt_secret()
+
+            # Restore
+            importlib.reload(SPARC.config)
+
+    def test_custom_secret_in_production_succeeds(self):
+        """Starting with a custom secret in production must not raise."""
+        with patch.dict(os.environ, {"APP_ENV": "production"}):
+            import importlib
+            import SPARC.config
+            importlib.reload(SPARC.config)
+
+            from SPARC.auth import check_jwt_secret
+            with patch("SPARC.auth.JWT_SECRET", "my-secure-random-secret-abc123"):
+                # Should not raise
+                check_jwt_secret()
+
+            with patch.dict(os.environ, {"APP_ENV": "development"}):
+                importlib.reload(SPARC.config)
+
+    def test_default_secret_unset_env_succeeds(self):
+        """When APP_ENV is unset (defaults to development), default secret is allowed."""
+        with patch.dict(os.environ, {}, clear=False):
+            # Remove APP_ENV if present
+            env = os.environ.copy()
+            env.pop("APP_ENV", None)
+            with patch.dict(os.environ, env, clear=True):
+                import importlib
+                import SPARC.config
+                importlib.reload(SPARC.config)
+
+                from SPARC.auth import _DEFAULT_JWT_SECRET, check_jwt_secret
+                with patch("SPARC.auth.JWT_SECRET", _DEFAULT_JWT_SECRET):
+                    # Should not raise (defaults to development)
+                    check_jwt_secret()
+
+                with patch.dict(os.environ, {"APP_ENV": "development"}):
+                    importlib.reload(SPARC.config)
+
+
+class TestCORSConfig:
+    """Test that CORS origins are configurable via environment variable."""
+
+    def test_default_cors_origins(self):
+        """When CORS_ORIGINS is unset, defaults to localhost origins."""
+        with patch.dict(os.environ, {"CORS_ORIGINS": ""}):
+            import importlib
+            import SPARC.config
+            importlib.reload(SPARC.config)
+            assert SPARC.config.cors_origins == [
+                "http://localhost:3000",
+                "http://localhost:5173",
+            ]
+
+    def test_custom_cors_origins(self):
+        """Setting CORS_ORIGINS configures allowed origins."""
+        with patch.dict(os.environ, {"CORS_ORIGINS": "https://sparc.example.com,https://app.example.com"}):
+            import importlib
+            import SPARC.config
+            importlib.reload(SPARC.config)
+            assert SPARC.config.cors_origins == [
+                "https://sparc.example.com",
+                "https://app.example.com",
+            ]
+            # Restore
+            with patch.dict(os.environ, {"CORS_ORIGINS": ""}):
+                importlib.reload(SPARC.config)
+
+    def test_single_cors_origin(self):
+        """A single origin without comma works correctly."""
+        with patch.dict(os.environ, {"CORS_ORIGINS": "https://sparc.example.com"}):
+            import importlib
+            import SPARC.config
+            importlib.reload(SPARC.config)
+            assert SPARC.config.cors_origins == ["https://sparc.example.com"]
+            with patch.dict(os.environ, {"CORS_ORIGINS": ""}):
+                importlib.reload(SPARC.config)

tests/test_serp_api.py

@@ -1,9 +1,8 @@
 """Tests for SERP API patent retrieval and parsing functionality."""
 
-import os
-import pytest
-from unittest.mock import patch, Mock
 from datetime import datetime, timedelta
+from unittest.mock import Mock
+
 from SPARC.serp_api import SERP
 from SPARC.types import Patent