app.py

import streamlit as st
import pandas as pd
import json
import smtplib
from email.message import EmailMessage
from typing import Dict, List

from jobspy import scrape_jobs
import groq


# ======================================================
# Utilities
# ======================================================

def remove_duplicates(df: pd.DataFrame) -> pd.DataFrame:
    df["__dedup__"] = (
        df.get("title", "").astype(str) + "|" +
        df.get("company", "").astype(str) + "|" +
        df.get("location", "").astype(str) + "|" +
        df.get("job_url", "").astype(str)
    )
    return df.drop_duplicates("__dedup__").drop(columns="__dedup__")


def compute_keyword_score(text: str, keywords: List[str]) -> int:
    text_l = (text or "").lower()
    return sum(text_l.count(k.lower()) for k in keywords if k)


# ======================================================
# Optional Email Helper
# ======================================================

def email_secrets_available() -> bool:
    required = [
        "SMTP_SERVER",
        "SMTP_PORT",
        "SMTP_USER",
        "SMTP_PASSWORD",
        "EMAIL_FROM",
    ]
    return all(key in st.secrets for key in required)


def send_email_with_csv(recipient_email: str, df: pd.DataFrame):
    smtp_server = st.secrets["SMTP_SERVER"]
    smtp_port = int(st.secrets["SMTP_PORT"])
    smtp_user = st.secrets["SMTP_USER"]
    smtp_password = st.secrets["SMTP_PASSWORD"]
    email_from = st.secrets["EMAIL_FROM"]

    msg = EmailMessage()
    msg["Subject"] = "Your Job Search Results"
    msg["From"] = email_from
    msg["To"] = recipient_email
    msg.set_content(
        "Hello,\n\nAttached is the CSV file containing your job search results.\n\nRegards,\nPrivate Job Search Tool"
    )

    csv_data = df.to_csv(index=False)
    msg.add_attachment(csv_data, subtype="csv", filename="job_results.csv")

    with smtplib.SMTP(smtp_server, smtp_port) as server:
        server.starttls()
        server.login(smtp_user, smtp_password)
        server.send_message(msg)


# ======================================================
# AI helper (intent extraction)
# ======================================================

def extract_search_parameters(client, prompt: str) -> Dict[str, str]:
    system_prompt = """
    Extract job search parameters.
    Return JSON ONLY:

    {
      "search_term": "<job title or keywords>",
      "location": "<city, province/state, or country>"
    }
    """

    response = client.chat.completions.create(
        model="meta-llama/llama-4-scout-17b-16e-instruct",
        temperature=0.2,
        max_tokens=200,
        messages=[
            {"role": "system", "content": system_prompt},
            {"role": "user", "content": prompt}
        ]
    )

    try:
        return json.loads(response.choices[0].message.content)
    except Exception:
        return {"search_term": prompt, "location": "Canada"}


# ======================================================
# Job scraping
# ======================================================

@st.cache_data(ttl=3600)
def get_indeed_jobs(
    search_term: str,
    location: str,
    radius_km: int,
    posted_within_days: int
) -> pd.DataFrame:
    try:
        jobs = scrape_jobs(
            site_name=["indeed"],
            search_term=search_term,
            location=location,
            results_wanted=100,
            hours_old=posted_within_days * 24,
            country_indeed="Canada",
            radius=radius_km
        )
        return pd.DataFrame(jobs)
    except Exception:
        return pd.DataFrame()


# ======================================================
# Streamlit App
# ======================================================

def main():
    st.set_page_config(page_title="Private Job Search", layout="centered")
    st.title("📄 Private Job Search, Rank & Download")

    # --------------------------------------------------
    # Job description
    # --------------------------------------------------
    job_prompt = st.text_area(
        "Describe the job you are looking for",
        placeholder="e.g. Civil Engineer, Water Resources, Transportation",
        height=120
    )

    api_key = st.text_input("Groq API Key", type="password")

    # --------------------------------------------------
    # City selection
    # --------------------------------------------------
    st.subheader("Location")

    predefined_cities = [
        "Use AI / Prompt Location",
        "Calgary, AB",
        "Edmonton, AB",
        "Toronto, ON",
        "Vancouver, BC",
        "Mississauga, ON",
        "Brampton, ON",
        "Ottawa, ON",
        "Hamilton, ON",
        "Custom city..."
    ]

    selected_city = st.selectbox("Select city", predefined_cities)

    custom_city = ""
    if selected_city == "Custom city...":
        custom_city = st.text_input(
            "Enter city (e.g., Red Deer, AB or Surrey, BC)"
        )

    # --------------------------------------------------
    # Job boards
    # --------------------------------------------------
    st.subheader("Job Boards")
    use_indeed = st.checkbox("Indeed", value=True)

    # --------------------------------------------------
    # Filters
    # --------------------------------------------------
    st.subheader("Filters")

    posted_within_days = st.slider(
        "Posted within last (days)",
        min_value=1,
        max_value=30,
        value=7
    )

    radius_km = st.slider(
        "Search radius (km)",
        min_value=5,
        max_value=100,
        value=25,
        step=5
    )

    # --------------------------------------------------
    # Keyword ranking
    # --------------------------------------------------
    keywords_raw = st.text_input(
        "Keyword ranking (comma-separated)",
        placeholder="water, wastewater, stormwater, EPANET"
    )
    keywords = [k.strip() for k in keywords_raw.split(",") if k.strip()]

    # --------------------------------------------------
    # Optional email
    # --------------------------------------------------
    send_email = st.checkbox("📧 Send results by email (optional)")
    email_address = st.text_input("Email address") if send_email else None

    # --------------------------------------------------
    # Action
    # --------------------------------------------------
    if st.button(
        "🔍 Search Jobs",
        disabled=not job_prompt or not api_key
    ):
        client = groq.Client(api_key=api_key)

        with st.spinner("Understanding your request..."):
            params = extract_search_parameters(client, job_prompt)

        # Resolve final location
        if selected_city == "Use AI / Prompt Location":
            location = params.get("location", "Canada")
        elif selected_city == "Custom city...":
            location = custom_city if custom_city else params.get("location", "Canada")
        else:
            location = selected_city

        if not use_indeed:
            st.warning("No job boards selected.")
            return

        with st.spinner("Searching jobs..."):
            jobs_df = get_indeed_jobs(
                params["search_term"],
                location,
                radius_km,
                posted_within_days
            )

            if jobs_df.empty:
                st.warning("No jobs found.")
                return

            jobs_df.fillna("", inplace=True)
            jobs_df = remove_duplicates(jobs_df)

        # Keyword ranking
        jobs_df["keyword_score"] = jobs_df.apply(
            lambda r: compute_keyword_score(
                f"{r.get('title','')} {r.get('description','')}",
                keywords
            ),
            axis=1
        )

        jobs_df = jobs_df.sort_values(
            by="keyword_score",
            ascending=False
        )

        st.success(f"✅ Found {len(jobs_df)} jobs for **{location}**")

        # --------------------------------------------------
        # Download
        # --------------------------------------------------
        csv_data = jobs_df.to_csv(index=False).encode("utf-8")
        st.download_button(
            label="⬇️ Download Jobs (CSV)",
            data=csv_data,
            file_name="job_results.csv",
            mime="text/csv"
        )

        # --------------------------------------------------
        # Optional email
        # --------------------------------------------------
        if send_email:
            if not email_address:
                st.warning("Please enter an email address.")
            elif not email_secrets_available():
                st.warning("Email not configured. Download is still available.")
            else:
                with st.spinner("Sending email..."):
                    try:
                        send_email_with_csv(email_address, jobs_df)
                        st.success(f"📧 Results emailed to {email_address}")
                    except Exception as e:
                        st.error(f"Failed to send email: {e}")

        # --------------------------------------------------
        # Preview
        # --------------------------------------------------
        st.subheader("Preview (Top 20 Results)")
        preview_cols = [
            c for c in [
                "title", "company", "location",
                "keyword_score", "date_posted", "job_url"
            ] if c in jobs_df.columns
        ]
        st.dataframe(
            jobs_df[preview_cols].head(20),
            use_container_width=True
        )


if __name__ == "__main__":
    main()