Move history crawl interval to env

This commit is contained in:
zetaphor committed 2025-04-12 14:45:59 -05:00
1 parent 5629a92acf
commit 6307d962d8
4 files changed
+10 -1

No files matched your search

+1
View File
@@ -0,0 +1 @@
CRAWL_INTERVAL=10
+1
View File
@@ -10,4 +10,5 @@ dependencies = [
"uvicorn[standard]",
"crawl4ai",
"browser-history",
"python-dotenv"
]
+6 -1
View File
@@ -10,9 +10,14 @@ from base_crawler import BaseCrawler
import asyncio
from contextlib import asynccontextmanager
from browser_history import get_history
from dotenv import load_dotenv
# Create logs directory if it doesn't exist
os.makedirs('logs', exist_ok=True)
# Load environment variables
load_dotenv()
CRAWL_INTERVAL = int(os.getenv('CRAWL_INTERVAL', 30)) # Default to 30 seconds if not set
@asynccontextmanager
async def lifespan(app: FastAPI):
# Startup
@@ -62,7 +67,7 @@ class HistoryCrawler(BaseCrawler):
if success:
self.logger.info(f"Processed historical URL: {url}")
await asyncio.sleep(30) # Wait 30 seconds before next crawl
await asyncio.sleep(CRAWL_INTERVAL) # Use environment variable for interval
@app.websocket("/ws")
async def websocket_endpoint(websocket: WebSocket):
Generated
+2
View File
@@ -197,6 +197,7 @@ dependencies = [
{ name = "browser-history" },
{ name = "crawl4ai" },
{ name = "fastapi" },
{ name = "python-dotenv" },
{ name = "uvicorn", extra = ["standard"] },
]
@@ -205,6 +206,7 @@ requires-dist = [
{ name = "browser-history" },
{ name = "crawl4ai" },
{ name = "fastapi" },
{ name = "python-dotenv" },
{ name = "uvicorn", extras = ["standard"] },
]