Update app2.py

app2.py

@@ -1253,22 +1253,14 @@ def get_trend_description(df):
         return "The number of cases has remained relatively stable over the five-year period."
 
 PRACTICE_AREAS = [
-    "Personal Injury", "Medical Malpractice", "Criminal Law", "DUI & DWI", "Family Law",
-    "Divorce", "Bankruptcy", "Business Law", "Consumer Law", "Employment Law",
-    "Estate Planning", "Foreclosure Defense", "Immigration Law", "Intellectual Property",
-    "Nursing Home Abuse", "Probate", "Products Liability", "Real Estate Law", "Tax Law",
-    "Traffic Tickets", "Workers' Compensation", "Agricultural Law", "Animal & Dog Law",
-    "Antitrust Law", "Appeals & Appellate", "Arbitration & Mediation", "Asbestos & Mesothelioma",
-    "Cannabis & Marijuana Law", "Civil Rights", "Collections", "Communications & Internet Law",
-    "Construction Law", "Domestic Violence", "Education Law", "Elder Law",
-    "Energy, Oil & Gas Law", "Entertainment & Sports Law", "Environmental Law",
-    "Gov & Administrative Law", "Health Care Law", "Insurance Claims", "Insurance Defense",
-    "International Law", "Juvenile Law", "Landlord Tenant", "Legal Malpractice",
-    "Maritime Law", "Military Law", "Municipal Law", "Native American Law", "Patents",
-    "Securities Law", "Social Security Disability", "Stockbroker & Investment Fraud",
-    "Trademarks", "White Collar Crime"
+    "Bankruptcy", "Business", "Civil Rights", "Consumer", "Criminal Defense", "DUI",
+    "Employment", "Estate Planning", "Family", "Government", "Immigration", "Injury",
+    "Insurance", "Intellectual Property", "International", "Juvenile", "Landlord Tenant",
+    "Military", "Products Liability", "Real Estate", "Social Security Disability", "Tax",
+    "Traffic Tickets", "Workers' Compensation"
 ]
 
+# All US states and territories
 STATES = [
     "Alabama", "Alaska", "Arizona", "Arkansas", "California", "Colorado", "Connecticut",
     "Delaware", "Florida", "Georgia", "Hawaii", "Idaho", "Illinois", "Indiana", "Iowa",
@@ -1277,9 +1269,8 @@ STATES = [
     "New Jersey", "New Mexico", "New York", "North Carolina", "North Dakota", "Ohio",
     "Oklahoma", "Oregon", "Pennsylvania", "Rhode Island", "South Carolina", "South Dakota",
     "Tennessee", "Texas", "Utah", "Vermont", "Virginia", "Washington", "West Virginia",
-    "Wisconsin", "Wyoming", "Washington, DC"
+    "Wisconsin", "Wyoming", "District of Columbia", "Puerto Rico"
 ]
-
 CITIES_BY_STATE = {
     "Alabama": ["Birmingham", "Montgomery", "Mobile", "Huntsville", "Tuscaloosa", "Hoover", "Dothan", "Auburn", "Decatur", "Madison", "Florence", "Gadsden", "Vestavia Hills", "Prattville", "Phenix City", "Alabaster", "Bessemer", "Prichard", "Opelika", "Enterprise"],
     "Alaska": ["Anchorage", "Fairbanks", "Juneau", "Sitka", "Ketchikan", "Wasilla", "Kenai", "Kodiak", "Bethel", "Palmer", "Homer", "Unalaska", "Barrow", "Soldotna", "Valdez", "Nome", "Kotzebue", "Seward", "Wrangell", "Dillingham"],
@@ -1334,13 +1325,16 @@ CITIES_BY_STATE = {
     "Washington, DC": ["Washington"]
 }
 
+def format_url_component(s):
+    return s.lower().replace(' ', '-').replace(',', '').replace('&', 'and')
+
 def find_lawyers(practice_area, state, city=None):
     base_url = "https://lawyers.findlaw.com"
     formatted_practice_area = format_url_component(practice_area)
-    formatted_state = state.lower().replace(' ', '-')
+    formatted_state = format_url_component(state)
     
     if city:
-        formatted_city = city.lower().replace(' ', '-')
+        formatted_city = format_url_component(city)
         search_url = f"{base_url}/{formatted_state}/{formatted_city}/{formatted_practice_area}-lawyer.html"
     else:
         search_url = f"{base_url}/{formatted_state}/{formatted_practice_area}-lawyer.html"
@@ -1375,15 +1369,20 @@ def find_lawyers(practice_area, state, city=None):
             phone_elem = card.find('p', class_='phone')
             phone = phone_elem.text.strip() if phone_elem else "N/A"
             
-            # FindLaw doesn't provide ratings, so we'll generate a random one for demonstration
-            rating = f"{random.uniform(3.0, 5.0):.1f}/5.0"
+            # Extract years of experience if available
+            experience = "N/A"
+            exp_elem = card.find('p', class_='experience')
+            if exp_elem:
+                exp_match = re.search(r'(\d+)\s+years', exp_elem.text)
+                if exp_match:
+                    experience = f"{exp_match.group(1)} years"
             
             lawyers.append({
                 'name': name,
                 'location': location,
                 'practice_areas': practice_areas,
                 'phone': phone,
-                'rating': rating,
+                'experience': experience,
                 'profile_url': profile_url
             })
         except AttributeError as e:
@@ -1395,40 +1394,52 @@ def find_lawyers(practice_area, state, city=None):
     
     return lawyers
 
-def format_url_component(s):
-    return s.lower().replace(' ', '-').replace(',', '').replace('&', 'and')
+def get_cities(state):
+    formatted_state = format_url_component(state)
+    url = f"https://lawyers.findlaw.com/{formatted_state}/"
+    
+    headers = {
+        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'
+    }
+    
+    try:
+        response = requests.get(url, headers=headers, timeout=10)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.content, 'html.parser')
+        city_links = soup.find_all('a', class_='city-link')
+        return [city.text.strip() for city in city_links]
+    except requests.RequestException as e:
+        print(f"Error fetching cities: {e}")
+        return []
 
 def lawyer_finder_ui():
-    st.subheader("Find the Best Lawyer for Your Case")
+    st.title("FindLaw Lawyer Search")
     
     practice_area = st.selectbox("Select a Practice Area:", PRACTICE_AREAS)
     state = st.selectbox("Select a State:", STATES)
     
-    cities = CITIES_BY_STATE.get(state, [])
-    if cities:
-        city = st.selectbox("Select a City (optional):", [""] + cities)
-    else:
-        city = None
-        st.info("No major cities available for the selected state. The search will be performed statewide.")
+    cities = get_cities(state)
+    city = st.selectbox("Select a City (optional):", [""] + cities)
     
     if st.button("Find Lawyers", type="primary"):
-        with st.spinner("Searching for the best lawyers..."):
-            lawyers = find_lawyers(practice_area, state, city)
+        with st.spinner("Searching for lawyers..."):
+            lawyers = find_lawyers(practice_area, state, city if city else None)
             
             if lawyers:
-                st.success(f"Found {len(lawyers)} potential lawyers for your case.")
+                st.success(f"Found {len(lawyers)} lawyers matching your criteria.")
                 for lawyer in lawyers:
                     with st.expander(f"{lawyer['name']} - {lawyer['location']}"):
                         col1, col2 = st.columns([2, 1])
                         with col1:
                             st.markdown(f"**Practice Areas:** {lawyer['practice_areas']}")
                             st.markdown(f"**Phone:** {lawyer['phone']}")
-                            st.markdown(f"**Rating:** {lawyer['rating']} (Estimated)")
+                            st.markdown(f"**Experience:** {lawyer['experience']}")
                         with col2:
                             st.markdown(f"[![View Profile](https://img.shields.io/badge/View_Profile-FindLaw-blue?style=for-the-badge)]({lawyer['profile_url']})")
             else:
                 st.warning("No lawyers found matching your criteria. Try broadening your search.")
 
+
 class LegalDataRetriever:
     def __init__(self):
         self.session = requests.Session()