"""
email-deliverability-scorer — LeadsBlue Hugging Face Space.

Estimates an email list's expected deliverability (0-100) from honest, real
factors: verification claim, data age, country compliance friction, and
industry receptiveness (derived from benchmark open rates).

This is an estimate, not a guarantee (stated in the UI).

Data source: country-insights.json, industry-insights.json
Author: Luther Johnson (ORCID 0009-0008-9836-1280) · LeadsBlue Research
License: CC BY 4.0 · Dataset DOI: 10.5281/zenodo.20136256
"""

import gradio as gr

from leadsblue_common import (
    CITATION_FOOTER,
    softwareapplication_jsonld,
    load_country_insights,
    load_industry_insights,
    parse_pct_range,
)

try:
    COUNTRIES = load_country_insights()
    INDUSTRIES = load_industry_insights()
    LOAD_ERROR = None
except Exception as exc:  # pragma: no cover
    COUNTRIES, INDUSTRIES, LOAD_ERROR = {}, {}, str(exc)

COUNTRY_CHOICES = sorted(c.title() for c in COUNTRIES.keys())
INDUSTRY_CHOICES = sorted(INDUSTRIES.keys())
VERIFICATION_CHOICES = ["Yes — recently verified", "Partial", "No / unknown"]
AGE_CHOICES = ["Under 6 months", "6–12 months", "12–24 months", "24+ months"]

# GDPR / strict-consent jurisdictions (compliance friction for cold B2B email)
GDPR_COUNTRIES = {
    "austria", "belgium", "bulgaria", "croatia", "cyprus", "czech republic",
    "denmark", "estonia", "finland", "france", "germany", "greece", "hungary",
    "ireland", "italy", "latvia", "lithuania", "luxembourg", "malta",
    "netherlands", "poland", "portugal", "romania", "slovakia", "slovenia",
    "spain", "sweden", "united kingdom",
}
CANSPAM_COUNTRIES = {"united states", "canada"}

VERIFICATION_POINTS = {"Yes — recently verified": 30, "Partial": 10, "No / unknown": 0}
AGE_POINTS = {"Under 6 months": 0, "6–12 months": -10, "12–24 months": -20, "24+ months": -35}


def _compliance_friction(country):
    c = country.lower()
    if c in GDPR_COUNTRIES:
        return -10, "GDPR jurisdiction (consent expectations): -10"
    if c in CANSPAM_COUNTRIES:
        return -3, "CAN-SPAM/CASL jurisdiction: -3"
    return 0, "Lighter regulatory friction: 0"


def _industry_receptiveness(industry):
    perf = INDUSTRIES.get(industry, {}).get("cold_email_performance", {})
    if "open_rate" not in perf:
        return 0, "Industry receptiveness: no data (0)"
    lo, hi = parse_pct_range(perf["open_rate"])
    mid = (lo + hi) / 2
    if mid >= 25:
        return 10, "High-receptiveness industry (open ~%.0f%%): +10" % mid
    if mid >= 20:
        return 6, "Above-average receptiveness (open ~%.0f%%): +6" % mid
    if mid >= 15:
        return 3, "Moderate receptiveness (open ~%.0f%%): +3" % mid
    return 0, "Lower-receptiveness industry (open ~%.0f%%): 0" % mid


def score(country, industry, verification, data_age, list_source):
    if LOAD_ERROR:
        return "**Data failed to load.** %s" % LOAD_ERROR
    if not country or not industry or not verification or not data_age:
        return "Please complete country, industry, verification, and data age."

    base = 50
    parts = ["Baseline: 50"]

    v = VERIFICATION_POINTS[verification]
    parts.append("Verification (%s): %+d" % (verification, v))

    a = AGE_POINTS[data_age]
    parts.append("Data age (%s): %+d" % (data_age, a))

    cf, cf_note = _compliance_friction(country)
    parts.append(cf_note)

    ir, ir_note = _industry_receptiveness(industry)
    parts.append(ir_note)

    total = max(0, min(100, base + v + a + cf + ir))

    if total >= 75:
        verdict = "Strong — list profile supports good deliverability."
    elif total >= 55:
        verdict = "Moderate — workable with verification and warm-up."
    elif total >= 35:
        verdict = "Caution — refresh and re-verify before sending at volume."
    else:
        verdict = "High risk — re-source or heavily clean before use."

    recs = []
    if v < 30:
        recs.append("Verify the list with a real-time validation pass before sending.")
    if a < 0:
        recs.append("Refresh records older than 6 months; B2B data decays ~2-3%/month.")
    if cf < 0:
        recs.append("For this jurisdiction, ensure a legitimate-interest basis and easy opt-out.")
    if not recs:
        recs.append("Maintain list hygiene and monitor bounce/complaint rates per send.")

    body = (
        "## Deliverability estimate: %d / 100\n\n"
        "**%s**\n\n"
        "### Score breakdown\n%s\n\n"
        "### Recommended actions\n%s\n\n"
        "> _This is an estimate based on benchmark factors, not a guarantee. "
        "Actual deliverability depends on sender reputation, content, and ESP "
        "configuration._\n\n---\n%s"
    ) % (
        total,
        verdict,
        "\n".join("- " + p for p in parts),
        "\n".join("- " + r for r in recs),
        CITATION_FOOTER,
    )
    return body


HEAD = softwareapplication_jsonld(
    "LeadsBlue Email Deliverability Scorer",
    "Estimates B2B email list deliverability from verification, data age, jurisdiction, and industry.",
)

_blocks_kwargs = {"title": "LeadsBlue Email Deliverability Scorer"}
try:
    demo = gr.Blocks(head=HEAD, **_blocks_kwargs)
except TypeError:
    demo = gr.Blocks(**_blocks_kwargs)  # older/newer gradio: head set at launch()
with demo:
    gr.Markdown(
        "# LeadsBlue Email Deliverability Scorer\n"
        "An honest, factor-based estimate of how a B2B list is likely to perform. "
        "No guarantees — just transparent scoring."
    )
    with gr.Row():
        country = gr.Dropdown(COUNTRY_CHOICES, label="Country", value=("United States" if "United States" in COUNTRY_CHOICES else None))
        industry = gr.Dropdown(INDUSTRY_CHOICES, label="Industry", value=(INDUSTRY_CHOICES[0] if INDUSTRY_CHOICES else None))
    with gr.Row():
        verification = gr.Dropdown(VERIFICATION_CHOICES, label="Email verification", value="Partial")
        data_age = gr.Dropdown(AGE_CHOICES, label="Data age", value="6–12 months")
    list_source = gr.Textbox(label="List source (optional)", placeholder="e.g. opt-in webinar, scraped directory, purchased...")
    btn = gr.Button("Score deliverability", variant="primary")
    out = gr.Markdown()
    btn.click(score, [country, industry, verification, data_age, list_source], out)
    gr.Markdown("---\n_%s_" % CITATION_FOOTER)

if __name__ == "__main__":
    try:
        demo.launch(head=HEAD)
    except TypeError:
        demo.launch()