[Python] Playwright Python: PDF renders in new tab (200 application/pdf) but I can’t save it —...

I’m automating the Pima County inspections site with Playwright (Python). The flow is:

• 
  Open listings: https://publicworks.pima.gov/inspections
  
  
• 
  Click a record link which goes to an Accela page like: https://aca-prod.accela.com/PIMA/Cap/CapDetail.aspx?Module=Building&...
  
  
• 
  Click View inspection summary. This opens a new tab to the Accela report parameter page: https://aca-prod.accela.com/PIMA/Re...T_SUMMARY_REPORT&reportID=762&agencyCode=PIMA
  

That page then loads the actual PDF from:
https://aca-prod.accela.com/PIMA/Re...T_SUMMARY_REPORT&reportID=762&agencyCode=PIMA

In the browser I can see the PDF rendered in the new tab. My network log also shows the PDF response clearly:

RESP 200 https://aca-prod.accela.com/PIMA/Report/ShowReport.aspx?Module=Building&reportType=PRINT_PERMIT_SUMMARY_REPORT&reportID=762&agencyCode=PIMA | ct=application/pdf cl=160319


…but my code never saves the bytes, it is very frustrating. acceptDownloads=True doesn’t help (server is Content-Type: application/pdf inline, not an attachment). I tried several approaches.

Environment:

• Windows 11
• Python 3.13.7
• Playwright Python 1.47.0
• Browsers: channel="chrome" (headed) and chromium (headed). Same result.

Minimal code that reproduces

from pathlib import Path
from playwright.sync_api import sync_playwright

BASE = "https://publicworks.pima.gov/inspections"
DOWNLOADS = Path("downloads"); DOWNLOADS.mkdir(exist_ok=True)

def is_pdf_response(resp):
return ("Report/ShowReport.aspx" in resp.url
and "application/pdf" in (resp.headers.get("content-type","").lower()))

with sync_playwright() as p:
browser = p.chromium.launch(channel="chrome", headless=False)
ctx = browser.new_context(accept_downloads=True) # inline PDF, so this never fires
page = ctx.new_page()

# (1) Go to listings and click Prev until rows exist (omitted for brevity)
page.goto(BASE)

# (2) Click a record link -> record details page (omitted: selector resolution)
# page.click("a[href*='CapDetail.aspx']")

# (3) Click "View inspection summary" which opens a new tab
with ctx.expect_page() as new_page_info:
page.click("a#ctl00_PlaceHolderMain_lnkPrintSummary") # link text/ID varies, but this opens the tab
pdf_tab = new_page_info.value

# (4) The parameter page loads, then the PDF URL is fetched by the site.
# Try to catch the actual PDF response:
caught = {"saved": False}

def on_response(resp):
if is_pdf_response(resp):
# Attempt 1: save body() from the response we just saw
content = resp.body() # <- returns bytes?
if content and content[:4] == b"%PDF":
out = DOWNLOADS / "report_from_response.pdf"
out.write_bytes(content)
print("Saved via context.on('response'):", out)
caught["saved"] = True

ctx.on("response", on_response)

# Also try a tab-scoped expect_response around the time it should load:
try:
with pdf_tab.expect_response(lambda r: is_pdf_response(r), timeout=15000) as resp_info:
# Sometimes there is a submit/run-report button on ReportParameter.aspx;
# sometimes it loads automatically. So just wait here.
pass
r = resp_info.value
body = r.body()
if body and body[:4] == b"%PDF":
out = DOWNLOADS / "report_from_expect.pdf"
out.write_bytes(body)
print("Saved via expect_response:", out)
caught["saved"] = True
except Exception as e:
print("expect_response failed:", e)

# Fallback Attempt 3: if we saw the parameter page URL, re-GET the PDF directly
# with cookies + Referer (same origin).
if not caught["saved"]:
param_url = pdf_tab.url
pdf_url = "https://aca-prod.accela.com/PIMA/Report/ShowReport.aspx?Module=Building&reportType=PRINT_PERMIT_SUMMARY_REPORT&reportID=762&agencyCode=PIMA"
r = ctx.request.get(pdf_url, headers={"Referer": param_url})
print("Re-GET status:", r.status, r.headers.get("content-type"))
body = r.body()
if body and body[:4] == b"%PDF":
out = DOWNLOADS / "report_from_reget.pdf"
out.write_bytes(body)
print("Saved via re-GET:", out)
caught["saved"] = True

print("Done; saved?", caught["saved"])


Questions:

What’s the reliable Playwright pattern to capture inline PDFs that render in a new tab (navigation to ShowReport.aspx) and save the bytes?

Is there a timing/targeting gotcha with expect_response on the newly opened tab (navigation responses vs subresource responses)?

If the PDF stream is delivered as a navigation response, is there a recommended way to read that response body (CDP hook? another Playwright API?) from Python?

This Accela flow is public (no login). Any pointers or a minimal snippet that consistently grabs the ShowReport.aspx bytes would be hugely appreciated.

What I’ve tried:

BrowserContext.on("response") and page.expect_response(...) with a predicate for ShowReport.aspx + content-type: application/pdf. I can see the response in my own network log, but the save code never triggers, or body() isn’t giving me PDF bytes.

Fallback ctx.request.get(pdf_url, headers={"Referer": param_url}) using the same cookies/session. Status is 200 and content-type: application/pdf in the network log, still no saved file in my script.

Clicking the Chrome PDF viewer toolbar “download” button isn’t accessible via Playwright (browser UI).

accept_downloads=True + page.on("download") doesn’t fire because the server returns inline PDF (no Content-Disposition: attachment).

Continue reading...