Finviz - Stock fundamentals and preset screeners

import json, os, re, tempfile, time, urllib.parse
 
def js_retry(expr, tries=6):
    # Finviz pages are ad-heavy; CDP calls sometimes miss the 5 s daemon timeout while the page is busy.
    for _ in range(tries):
        try:
            return js(expr)
        except TimeoutError:
            pass
        except RuntimeError as e:
            if "timed out" not in str(e):
                raise
        time.sleep(2)
    raise RuntimeError(f"page stayed busy for {tries} tries: {expr[:60]}")
 
def fv_open(url, timeout=45, settle=2.5):
    """Open url; return the landed URL. Raises on a Cloudflare challenge or if the page never loads."""
    prev = js_retry("location.href")
    try:
        goto_url(url)
    except TimeoutError:
        pass  # Page.navigate can exceed 5 s while the previous page unloads; the navigation still runs
    end = time.time() + timeout
    while True:
        if time.time() > end:
            raise RuntimeError(f"{url} did not finish loading in {timeout}s")
        time.sleep(1)
        href, state = js_retry("[location.href, document.readyState]")
        if state == "complete" and (href != prev or prev == url):
            break
    time.sleep(settle)
    title = js_retry("document.title") or ""
    challenge = js_retry("!!document.querySelector('#challenge-form, iframe[src*=\"challenges.cloudflare.com\"]')")
    if "Just a moment" in title or challenge:
        raise RuntimeError(f"Cloudflare challenge on {url}: stop and tell the user; do not retry around it")
    return js_retry("location.href")
 
SNAPSHOT_JS = """(() => {
  const rows = [];
  for (const t of document.querySelectorAll('table.js-snapshot-table'))
    for (const tr of t.rows) {
      const c = [...tr.cells];
      for (let i = 0; i + 1 < c.length; i += 2)
        rows.push([c[i].innerText.trim(), c[i + 1].innerText.trim(), c[i].getAttribute('data-boxover-html') || '']);
    }
  const txt = s => { const e = document.querySelector(s); return e ? e.innerText.trim() : null; };
  const link = p => { const a = document.querySelector(`a[href*="f=${p}_"]`); return a ? a.innerText.trim() : null; };
  const site = [...document.querySelectorAll('.quote-header a[href^="http"]')].map(a => a.href)[0] || null;
  const inc = document.getElementById('income-breakdown-init-data-0');
  return {rows, company: txt('.quote-header h2'), price: txt('.quote-price_price'), price_date: txt('.quote-price_date'),
          sector: link('sec'), industry: link('ind'), country: link('geo'), exchange: link('exch'), website: site,
          income: inc ? JSON.parse(inc.textContent).statement : null};
})()"""
 
def finviz_stock(ticker):
    """Snapshot fundamentals for one ticker from /stock?t=<T>. None if Finviz does not know the ticker."""
    t = ticker.strip().upper()
    if not re.fullmatch(r"[A-Z0-9.\-]{1,10}", t):
        raise ValueError(f"bad ticker {ticker!r}")
    url = "https://finviz.com/stock?" + urllib.parse.urlencode({"t": t})
    landed = urllib.parse.urlparse(fv_open(url))
    if landed.path != "/stock":
        return None  # unknown tickers redirect to /search?p=..., which robots.txt disallows: not read
    if urllib.parse.parse_qs(landed.query).get("t", [""])[0].upper() != t:
        raise RuntimeError(f"asked for {t}, landed on {landed.geturl()}")
    d = js_retry(SNAPSHOT_JS)
    if len(d["rows"]) < 80:
        raise RuntimeError(f"{t}: only {len(d['rows'])} snapshot cells; layout changed?")
    snap = {}
    for label, value, tip in d["rows"]:
        key = label if label not in snap else (tip or label + " #2")  # 'EPS next Y' appears twice
        snap[key] = value
    return {"ticker": t, "company": d["company"], "url": landed.geturl(), "sector": d["sector"], "industry": d["industry"],
            "country": d["country"], "exchange": d["exchange"], "website": d["website"],
            "last_price": d["price"], "price_time": d["price_date"],
            "snapshot": snap, "income_statement": d["income"]}  # last 2 fiscal years, millions of income["currency"]
 
tab = new_tab("about:blank")
out = {}
try:
    for i, t in enumerate(["AAPL", "NVDA", "ZZZZQ"]):
        if i:
            time.sleep(3)
        s = finviz_stock(t)
        if s is None:
            print(t, "-> not on Finviz")
            continue
        out[t] = s
        k = s["snapshot"]
        print(t, s["company"], s["exchange"], s["sector"], "/", s["industry"], "| price", s["last_price"], s["price_time"],
              "| cells", len(k), "| P/E", k["P/E"], "| Market Cap", k["Market Cap"], "| EPS next Y", k["EPS next Y"],
              "| EPS growth next year", k["EPS growth next year"])
finally:
    cdp("Target.closeTarget", targetId=tab)  # an open Finviz tab keeps running ads and can stall a small browser
print(json.dumps({x: out["AAPL"]["snapshot"][x] for x in list(out["AAPL"]["snapshot"])[:12]}, indent=1))
inc = out["AAPL"]["income_statement"]
print(inc["currency"], inc["data"]["Period"], "Total Revenue", inc["data"]["Total Revenue"], "Net Income", inc["data"]["Net Income"])
path = os.path.join(tempfile.gettempdir(), "finviz_stocks.json")
with open(path, "w") as f:
    json.dump(out, f, indent=1)
print("saved", path)

import html, json, os, re, tempfile, time
 
def js_retry(expr, tries=6):
    # Finviz pages are ad-heavy; CDP calls sometimes miss the 5 s daemon timeout while the page is busy.
    for _ in range(tries):
        try:
            return js(expr)
        except TimeoutError:
            pass
        except RuntimeError as e:
            if "timed out" not in str(e):
                raise
        time.sleep(2)
    raise RuntimeError(f"page stayed busy for {tries} tries: {expr[:60]}")
 
def fv_open(url, timeout=45, settle=2.5):
    """Open url; return the landed URL. Raises on a Cloudflare challenge or if the page never loads."""
    prev = js_retry("location.href")
    try:
        goto_url(url)
    except TimeoutError:
        pass  # Page.navigate can exceed 5 s while the previous page unloads; the navigation still runs
    end = time.time() + timeout
    while True:
        if time.time() > end:
            raise RuntimeError(f"{url} did not finish loading in {timeout}s")
        time.sleep(1)
        href, state = js_retry("[location.href, document.readyState]")
        if state == "complete" and (href != prev or prev == url):
            break
    time.sleep(settle)
    title = js_retry("document.title") or ""
    challenge = js_retry("!!document.querySelector('#challenge-form, iframe[src*=\"challenges.cloudflare.com\"]')")
    if "Just a moment" in title or challenge:
        raise RuntimeError(f"Cloudflare challenge on {url}: stop and tell the user; do not retry around it")
    return js_retry("location.href")
 
# The presets robots.txt allows (checked 2026-10-31). Every other /screener?... or /screener/stocks?... URL is disallowed,
# so the view (v) and signal (s) pair must be exactly one of these; nothing here changes v, filters or sort order.
ALLOWED_PRESETS = {
    "ta_topgainers": 340, "ta_toplosers": 340, "ta_newhigh": 340, "ta_newlow": 340, "n_upgrades": 340,
    "n_downgrades": 340, "n_earningsbefore": 340, "n_earningsafter": 340, "it_latestbuys": 340,
    "it_latestsales": 340, "f_ssr": 340, "f_halted": 340,
    "ta_mostvolatile": 320, "ta_mostactive": 320, "ta_unusualvolume": 320, "n_majornews": 320,
    "ta_overbought": 210, "ta_oversold": 210, "ta_p_tlsupport": 210, "ta_p_tlresistance": 210,
    "ta_p_horizontal": 210, "ta_p_wedgeup": 210, "ta_p_wedge": 210, "ta_p_wedgedown": 210,
    "ta_p_wedgeresistance": 210, "ta_p_wedgesupport": 210, "ta_p_channelup": 210, "ta_p_channel": 210,
    "ta_p_channeldown": 210, "ta_p_doubletop": 210, "ta_p_multipletop": 210, "ta_p_doublebottom": 210,
    "ta_p_multiplebottom": 210, "ta_p_headandshoulders": 210, "ta_p_headandshouldersinv": 210,
}
 
STATE_JS = """(() => { const s = [...document.scripts].find(s => s.type === 'application/json' && s.textContent.includes('"screenerType"'));
  return s ? s.textContent : null })()"""
TOTAL_JS = """(() => { const e = [...document.querySelectorAll('span')].find(e => /^#\\d+ \\/ \\d+ Total$/.test(e.innerText.trim()));
  return e ? e.innerText.trim() : null })()"""
 
def finviz_preset_page(signal, row=1):
    """One page of a robots-allowed preset. row is the site's own pager offset: 1, 11, 21 ... (v=340/320) or 1, 13, 25 ... (v=210)."""
    if signal not in ALLOWED_PRESETS:
        raise ValueError(f"{signal}: not a robots.txt-allowed preset")
    v = ALLOWED_PRESETS[signal]
    url = f"https://finviz.com/screener/stocks?v={v}&s={signal}" + (f"&r={row}" if row > 1 else "")
    landed = fv_open(url)
    if landed != url:
        raise RuntimeError(f"asked for {url}, landed on {landed}; not reading it")
    raw = js_retry(STATE_JS)
    if not raw:
        raise RuntimeError(f"{url}: no screener state JSON; layout changed?")
    st = json.loads(raw)
    if "data" in st:  # v=340 snapshot and v=320 news views: full rows with price, change, volume
        d = st["data"]
        num = lambda r, k, nd=2: None if r["numbers"].get(k) is None else round(r["numbers"][k], nd)
        rows = [{"ticker": r["text"]["ticker"], "company": r["text"]["company"], "industry": r["text"].get("industry"),
                 "country": r["text"].get("country"), "exchange": r["text"].get("exchange"),
                 "price": num(r, "price", 4), "change_pct": num(r, "change"),
                 "volume": r["numbers"].get("volume"), "avg_volume": num(r, "averageVolume", 0),
                 "market_cap_musd": num(r, "marketCap"), "pe": num(r, "PE")} for r in d["rows"]]
        total, first, size = d["totalCount"], d["firstRow"] + 1, d["pageSize"]
    else:  # v=210 chart views: tickers only (price data is in chart images)
        rows = []
        for c in st["charts"]["charts"]:
            # boxoverHtml: "<b>Century Casinos Inc</b><br>Resorts &amp; Casinos • USA • 14.63M"
            name, _, rest = html.unescape(c["boxoverHtml"]).partition("<br>")
            rows.append({"ticker": c["ticker"], "company": re.sub(r"<[^>]+>", "", name).strip()}
                        | dict(zip(("industry", "country", "market_cap"), rest.split(" • "))))
        m = re.fullmatch(r"#(\d+) / (\d+) Total", js_retry(TOTAL_JS) or "")
        if not m:
            raise RuntimeError(f"{url}: no '#n / N Total' counter")
        first, total, size = int(m.group(1)), int(m.group(2)), 12
    if first != row:
        raise RuntimeError(f"{url}: page starts at row {first}, expected {row}")
    return {"url": url, "view": v, "total": total, "first_row": first, "page_size": size, "rows": rows}
 
def finviz_preset(signal, max_pages=2, pause=3):
    pages = [finviz_preset_page(signal)]
    size, total = pages[0]["page_size"], pages[0]["total"]
    while len(pages) < max_pages and len(pages) * size < total:
        time.sleep(pause)
        pages.append(finviz_preset_page(signal, row=len(pages) * size + 1))
    rows = [r for p in pages for r in p["rows"]]
    return {"signal": signal, "view": pages[0]["view"], "total": total, "rows": rows, "truncated": len(rows) < total}
 
tab = new_tab("about:blank")
out = {}
try:
    for i, (sig, n) in enumerate([("ta_topgainers", 2), ("ta_unusualvolume", 1), ("ta_oversold", 2)]):
        if i:
            time.sleep(3)
        p = finviz_preset(sig, max_pages=n)
        out[sig] = p
        print(f'{sig:17} v={p["view"]} total={p["total"]:<4} rows={len(p["rows"]):<3} truncated={p["truncated"]}')
        for r in p["rows"][:3]:
            print("   ", r)
finally:
    cdp("Target.closeTarget", targetId=tab)  # an open Finviz tab keeps running ads and can stall a small browser
print("top gainers, change % in page order:", [r["change_pct"] for r in out["ta_topgainers"]["rows"]])
try:
    finviz_preset_page("cap_mega")  # a custom filter, not a preset: refused before any request
except ValueError as e:
    print("refused:", e)
path = os.path.join(tempfile.gettempdir(), "finviz_presets.json")
with open(path, "w") as f:
    json.dump(out, f, indent=1)
print("saved", path)