Working /experimenting

screenshot.py

@@ -1,11 +1,10 @@
-# main.py
-from fastapi import FastAPI, HTTPException, Query
+# scrape.py
+from fastapi import FastAPI, HTTPException
 from pydantic import BaseModel
-from typing import List, Optional
+from typing import Optional
 import base64
 import json
 from playwright.async_api import async_playwright, TimeoutError as PlaywrightTimeoutError
-import asyncio
 
 app = FastAPI(title="Web Analyzer API")
 
@@ -37,23 +36,42 @@ async def get_metadata(url: str):
     page, browser, pw = await get_page(url)
     try:
         title = await page.title()
-        desc = await page.get_attribute("meta[name='description']", "content")
-        og = {}
-        twitter = {}
+
+        # Get description meta tag
+        try:
+            desc = await page.get_attribute("meta[name='description']", "content")
+        except Exception:
+            desc = None
+
+        # Extract Open Graph metadata
         og = {}
         for prop in ["title", "description", "image"]:
-            selector = f"meta[property='og:{prop}']"
             try:
+                selector = f"meta[property='og:{prop}']"
                 if await page.query_selector(selector):
                     og[f"og:{prop}"] = await page.get_attribute(selector, "content")
                 else:
                     og[f"og:{prop}"] = None
-            except Exception as e:
+            except Exception:
                 og[f"og:{prop}"] = None
 
-            og[f"og:{prop}"] = await page.get_attribute(f"meta[property='og:{prop}']", "content")
-            twitter[f"twitter:{prop}"] = await page.get_attribute(f"meta[name='twitter:{prop}']", "content")
-        canonical = await page.get_attribute("link[rel='canonical']", "href")
+        # Extract Twitter metadata
+        twitter = {}
+        for prop in ["title", "description", "image"]:
+            try:
+                selector = f"meta[name='twitter:{prop}']"
+                if await page.query_selector(selector):
+                    twitter[f"twitter:{prop}"] = await page.get_attribute(selector, "content")
+                else:
+                    twitter[f"twitter:{prop}"] = None
+            except Exception:
+                twitter[f"twitter:{prop}"] = None
+
+        # Get canonical URL
+        try:
+            canonical = await page.get_attribute("link[rel='canonical']", "href")
+        except Exception:
+            canonical = None
         return {
             "title": title,
             "description": desc,
@@ -94,8 +112,15 @@ async def seo_audit(url: str):
                     internal += 1
                 else:
                     external += 1
-        robots = await page.get_attribute("meta[name='robots']", "content")
-        canonical = await page.get_attribute("link[rel='canonical']", "href")
+        try:
+            robots = await page.get_attribute("meta[name='robots']", "content")
+        except Exception:
+            robots = None
+
+        try:
+            canonical = await page.get_attribute("link[rel='canonical']", "href")
+        except Exception:
+            canonical = None
         return {
             "h1_count": h1_count,
             "missing_image_alts": missing_alts,
@@ -108,19 +133,39 @@ async def seo_audit(url: str):
         await browser.close()
         await pw.stop()
 
-
 @app.get("/performance")
 async def performance_metrics(url: str):
     page, browser, pw = await get_page(url)
     try:
-        nav_timing = await page.evaluate("JSON.stringify(performance.getEntriesByType('navigation'))")
-        timing = json.loads(nav_timing)[0]
-        fcp = await page.evaluate("performance.getEntriesByName('first-contentful-paint')[0]?.startTime")
-        lcp = await page.evaluate("performance.getEntriesByType('largest-contentful-paint')[0]?.renderTime")
-        cls_entries = await page.evaluate("JSON.stringify(performance.getEntriesByType('layout-shift'))")
-        cls = sum(e['value'] for e in json.loads(cls_entries))
+        # Get navigation timing
+        try:
+            nav_timing = await page.evaluate("JSON.stringify(performance.getEntriesByType('navigation'))")
+            timing = json.loads(nav_timing)[0] if nav_timing else {}
+            page_load_time = timing.get('duration', None)
+        except Exception:
+            page_load_time = None
+
+        # Get First Contentful Paint
+        try:
+            fcp = await page.evaluate("performance.getEntriesByName('first-contentful-paint')[0]?.startTime")
+        except Exception:
+            fcp = None
+
+        # Get Largest Contentful Paint
+        try:
+            lcp = await page.evaluate("performance.getEntriesByType('largest-contentful-paint')[0]?.renderTime")
+        except Exception:
+            lcp = None
+
+        # Get Cumulative Layout Shift
+        try:
+            cls_entries = await page.evaluate("JSON.stringify(performance.getEntriesByType('layout-shift'))")
+            cls = sum(e.get('value', 0) for e in json.loads(cls_entries) if isinstance(e, dict))
+        except Exception:
+            cls = None
+
         return {
-            "page_load_time_ms": timing['duration'],
+            "page_load_time_ms": page_load_time,
             "first_contentful_paint": fcp,
             "largest_contentful_paint": lcp,
             "cumulative_layout_shift": cls
@@ -176,4 +221,4 @@ async def accessibility_check(url: str):
         }
     finally:
         await browser.close()
-        await pw.stop()
+        await pw.stop()