Update app.py

app.py

@@ -1,33 +1,39 @@
 import streamlit as st
-import groq
-from jobspy import scrape_jobs
 import pandas as pd
 import json
-import time
 import smtplib
 from email.message import EmailMessage
 from typing import Dict
 
+from jobspy import scrape_jobs
+import groq
+
 
-# --------------------------------------------------
+# ======================================================
 # Utilities
-# --------------------------------------------------
+# ======================================================
 
 def remove_duplicates(df: pd.DataFrame) -> pd.DataFrame:
-    df["dedup_key"] = (
-        df["title"].fillna("") + "|" +
-        df["company"].fillna("") + "|" +
-        df["location"].fillna("") + "|" +
-        df["job_url"].fillna("")
+    """
+    Remove duplicate jobs using title + company + location + URL
+    """
+    df["__dedup_key__"] = (
+        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_key").drop(columns="dedup_key")
+    return df.drop_duplicates("__dedup_key__").drop(columns="__dedup_key__")
 
 
-# --------------------------------------------------
+# ======================================================
 # Email helper
-# --------------------------------------------------
+# ======================================================
 
 def send_email_with_csv(recipient_email: str, df: pd.DataFrame):
+    """
+    Send CSV file via SMTP email
+    """
     smtp_server = st.secrets["SMTP_SERVER"]
     smtp_port = int(st.secrets["SMTP_PORT"])
     smtp_user = st.secrets["SMTP_USER"]
@@ -47,7 +53,11 @@ def send_email_with_csv(recipient_email: str, df: pd.DataFrame):
     )
 
     csv_data = df.to_csv(index=False)
-    msg.add_attachment(csv_data, subtype="csv", filename="job_results.csv")
+    msg.add_attachment(
+        csv_data,
+        subtype="csv",
+        filename="job_results.csv"
+    )
 
     with smtplib.SMTP(smtp_server, smtp_port) as server:
         server.starttls()
@@ -55,14 +65,18 @@ def send_email_with_csv(recipient_email: str, df: pd.DataFrame):
         server.send_message(msg)
 
 
-# --------------------------------------------------
-# AI helper
-# --------------------------------------------------
+# ======================================================
+# AI helper (intent extraction only)
+# ======================================================
 
-def convert_prompt_to_parameters(client, prompt: str) -> Dict[str, str]:
+def extract_search_parameters(client, prompt: str) -> Dict[str, str]:
+    """
+    Extract search_term and location from free text
+    """
     system_prompt = """
-    Extract structured job search parameters.
+    Extract job search parameters.
     Return JSON ONLY:
+
     {
       "search_term": "<job title or keywords>",
       "location": "<city, province/state, or country>"
@@ -82,15 +96,22 @@ def convert_prompt_to_parameters(client, prompt: str) -> Dict[str, str]:
     try:
         return json.loads(response.choices[0].message.content)
     except Exception:
-        return {"search_term": prompt, "location": "Canada"}
+        # Safe fallback
+        return {
+            "search_term": prompt,
+            "location": "Canada"
+        }
 
 
-# --------------------------------------------------
+# ======================================================
 # Job scraping
-# --------------------------------------------------
+# ======================================================
 
 @st.cache_data(ttl=3600)
 def get_indeed_jobs(search_term: str, location: str) -> pd.DataFrame:
+    """
+    Scrape Indeed jobs using JobSpy
+    """
     try:
         jobs = scrape_jobs(
             site_name=["indeed"],
@@ -105,23 +126,29 @@ def get_indeed_jobs(search_term: str, location: str) -> pd.DataFrame:
         return pd.DataFrame()
 
 
-# --------------------------------------------------
-# Streamlit UI
-# --------------------------------------------------
+# ======================================================
+# Streamlit App
+# ======================================================
 
 def main():
-    st.set_page_config(page_title="Private Job Alert", layout="centered")
+    st.set_page_config(
+        page_title="Private Job Alert",
+        layout="centered"
+    )
 
     st.title("📧 Private Job Search & Email Alert")
 
     job_prompt = st.text_area(
         "Describe the job you are looking for",
-        placeholder="e.g., Civil Engineer, Transportation Planner, Water Resources Engineer in Alberta",
+        placeholder=(
+            "e.g. Civil Engineer, Transportation Planner, "
+            "Water Resources Engineer in Alberta"
+        ),
         height=120
     )
 
     email_address = st.text_input(
-        "Email address (CSV will be sent here)"
+        "Email address (CSV results will be sent here)"
     )
 
     api_key = st.text_input(
@@ -129,12 +156,15 @@ def main():
         type="password"
     )
 
-    if st.button("🔍 Search & Email Jobs", disabled=not (job_prompt and email_address and api_key)):
+    if st.button(
+        "🔍 Search & Email Jobs",
+        disabled=not (job_prompt and email_address and api_key)
+    ):
 
         client = groq.Client(api_key=api_key)
 
         with st.spinner("Understanding your request..."):
-            params = convert_prompt_to_parameters(client, job_prompt)
+            params = extract_search_parameters(client, job_prompt)
 
         with st.spinner("Searching jobs..."):
             jobs_df = get_indeed_jobs(
@@ -153,7 +183,7 @@ def main():
             try:
                 send_email_with_csv(email_address, jobs_df)
                 st.success(
-                    f"✅ {len(jobs_df)} jobs found and emailed successfully to {email_address}"
+                    f"✅ {len(jobs_df)} jobs found and emailed to {email_address}"
                 )
             except Exception as e:
                 st.error(f"Failed to send email: {e}")