refactor: extract magic numbers to configuration constants

Created centralized configuration constants in settings/consts.py:

## New Constants

**PDF & Caching**
- PDF_CACHE_TTL_SECONDS = 300
- PDF_GENERATION_TIMEOUT_SECONDS = 60

**SWOT Validation**
- SWOT_MIN_ITEMS_PER_CATEGORY = 2
- SWOT_MAX_ITEMS_PER_CATEGORY = 10
- SWOT_MIN_ANALYSIS_LENGTH = 100
- SWOT_MAX_ANALYSIS_LENGTH = 5000

**Status Updates**
- STATUS_UPDATE_DELAY_MIN_SECONDS = 0
- STATUS_UPDATE_DELAY_MAX_SECONDS = 5
- STATUS_POLL_INTERVAL_SECONDS = 1

**HTTP Timeouts**
- HTTP_REQUEST_TIMEOUT_SECONDS = 30
- HTTP_CONNECT_TIMEOUT_SECONDS = 10

**Input Validation**
- MAX_PRIMARY_ENTITY_LENGTH = 500
- MAX_COMPARISON_ENTITIES_LENGTH = 2000
- MAX_COMPARISON_ENTITIES_COUNT = 10

**Reddit API**
- REDDIT_MAX_SUBREDDITS = 10
- REDDIT_CONCURRENT_THRESHOLD = 3

## Updated Files

- router.py: Use validation length constants
- utils.py: Use status update delay constants
- tools.py: Use HTTP timeout constants
- pdf_cache.py: Use PDF cache TTL constant

## Benefits

- Single source of truth for configuration
- Easy to adjust timeouts and limits
- Better maintainability
- Consistent behavior across modules

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

src/backend/core/pdf_cache.py

@@ -12,9 +12,10 @@ from typing import Any
 from backend.core.core import SwotAnalysis
 from backend.core.pdf_service import compute_content_hash
 from backend.logger import logger
+from backend.settings.consts import PDF_CACHE_TTL_SECONDS
 
 # Cache configuration
-CACHE_TTL_SECONDS = 300  # 5 minutes
+CACHE_TTL_SECONDS = PDF_CACHE_TTL_SECONDS  # Imported from settings
 CLEANUP_INTERVAL_SECONDS = 60  # Run cleanup every minute
 
 

src/backend/core/tools.py

@@ -9,6 +9,10 @@ from backend.core.consts import AI_MODEL
 from backend.core.core import SwotAgentDeps, SwotAnalysis, swot_agent
 from backend.core.utils import report_tool_usage
 from backend.logger import logger
+from backend.settings.consts import (
+    HTTP_CONNECT_TIMEOUT_SECONDS,
+    HTTP_REQUEST_TIMEOUT_SECONDS,
+)
 from backend.utils import get_val, set_event_loop, windows_sys_event_loop_check
 
 
@@ -25,8 +29,10 @@ async def fetch_website_content(
     :return: str
     """
     logger.info(f"Fetching website content for: {url}")
-    # Set reasonable timeouts: 10s connect, 30s total
-    timeout = httpx.Timeout(30.0, connect=10.0)
+    # Configure timeouts to prevent hanging on slow responses
+    timeout = httpx.Timeout(
+        HTTP_REQUEST_TIMEOUT_SECONDS, connect=HTTP_CONNECT_TIMEOUT_SECONDS
+    )
     async with httpx.AsyncClient(
         follow_redirects=True, timeout=timeout
     ) as http_client:

src/backend/settings/consts.py

@@ -54,3 +54,34 @@ all_dialects = enum.Enum(
 
 # TODO: include Oracle and MSSQL dialects
 SECRET_KEY: str = config("SECRET_KEY")
+
+# Application Configuration Constants
+# ===================================
+
+# PDF Generation and Caching
+PDF_CACHE_TTL_SECONDS = 300  # 5 minutes
+PDF_GENERATION_TIMEOUT_SECONDS = 60
+
+# SWOT Analysis Validation
+SWOT_MIN_ITEMS_PER_CATEGORY = 2
+SWOT_MAX_ITEMS_PER_CATEGORY = 10
+SWOT_MIN_ANALYSIS_LENGTH = 100
+SWOT_MAX_ANALYSIS_LENGTH = 5000
+
+# Status Updates and Polling
+STATUS_UPDATE_DELAY_MIN_SECONDS = 0
+STATUS_UPDATE_DELAY_MAX_SECONDS = 5
+STATUS_POLL_INTERVAL_SECONDS = 1
+
+# External API Configuration
+HTTP_REQUEST_TIMEOUT_SECONDS = 30
+HTTP_CONNECT_TIMEOUT_SECONDS = 10
+
+# Input Validation
+MAX_PRIMARY_ENTITY_LENGTH = 500
+MAX_COMPARISON_ENTITIES_LENGTH = 2000
+MAX_COMPARISON_ENTITIES_COUNT = 10
+
+# Reddit API Configuration
+REDDIT_MAX_SUBREDDITS = 10
+REDDIT_CONCURRENT_THRESHOLD = 3  # Switch to async when >3 subreddits

src/backend/site/router.py

@@ -11,6 +11,10 @@ from backend.core.pdf_cache import pdf_cache
 from backend.core.pdf_service import generate_swot_pdf
 from backend.logger import logger
 from backend.settings import app_settings
+from backend.settings.consts import (
+    MAX_COMPARISON_ENTITIES_LENGTH,
+    MAX_PRIMARY_ENTITY_LENGTH,
+)
 from backend.site.consts import (
     ANALYSIS_COMPLETE_MESSAGE,
     ANALYZING_MESSAGE,
@@ -45,12 +49,12 @@ class AnalysisInput(BaseModel):
     primary_entity: str = Field(
         ...,
         min_length=1,
-        max_length=500,
+        max_length=MAX_PRIMARY_ENTITY_LENGTH,
         description="Primary entity (company name or URL)",
     )
     comparison_entities: str = Field(
         default="",
-        max_length=2000,
+        max_length=MAX_COMPARISON_ENTITIES_LENGTH,
         description="Comma-separated comparison entities (optional)",
     )
 

src/backend/site/utils.py

@@ -7,6 +7,10 @@ from loguru import logger
 
 from backend.core.core import SwotAgentDeps, SwotAnalysis
 from backend.core.tools import run_agent
+from backend.settings.consts import (
+    STATUS_UPDATE_DELAY_MAX_SECONDS,
+    STATUS_UPDATE_DELAY_MIN_SECONDS,
+)
 from backend.site.consts import (
     ANALYSIS_COMPLETE_MESSAGE,
     result_store,
@@ -20,8 +24,11 @@ async def emulate_tool_completion(session_id: str, message: str) -> None:
 
     Uses asyncio.sleep to avoid blocking the event loop.
     """
-    # Sleep a random amount of time between 0 and 5 seconds (async)
-    await asyncio.sleep(random.randint(0, 5))
+    # Sleep a random amount of time (async)
+    delay = random.randint(
+        STATUS_UPDATE_DELAY_MIN_SECONDS, STATUS_UPDATE_DELAY_MAX_SECONDS
+    )
+    await asyncio.sleep(delay)
     status_store[session_id].append(message)