Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -49,17 +49,17 @@ def parse_text_file(filepath):
|
|
| 49 |
def parse_html_file(filepath):
|
| 50 |
"""
|
| 51 |
Parseo para archivos HTML (por ejemplo, de Wikipedia).
|
| 52 |
-
Extrae el t铆tulo
|
| 53 |
-
|
| 54 |
"""
|
| 55 |
with open(filepath, 'r', encoding='utf-8') as f:
|
| 56 |
content = f.read()
|
| 57 |
|
| 58 |
-
#
|
| 59 |
soup = BeautifulSoup(content, "html.parser")
|
| 60 |
title = soup.title.string.strip() if soup.title else os.path.splitext(os.path.basename(filepath))[0]
|
| 61 |
|
| 62 |
-
#
|
| 63 |
coord_match = re.search(r'"wgCoordinates":\s*{\s*"lat":\s*([0-9\.-]+),\s*"lon":\s*([0-9\.-]+)', content)
|
| 64 |
if coord_match:
|
| 65 |
lat_str = coord_match.group(1)
|
|
@@ -75,11 +75,24 @@ def parse_html_file(filepath):
|
|
| 75 |
else:
|
| 76 |
lat, lon = None, None
|
| 77 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 78 |
return {
|
| 79 |
"name": title,
|
| 80 |
"lat": lat,
|
| 81 |
"lon": lon,
|
| 82 |
-
"content": content
|
|
|
|
| 83 |
}
|
| 84 |
|
| 85 |
def load_data_from_directory(directory, parser_func):
|
|
@@ -106,7 +119,7 @@ news_dir = os.path.join("data", "news")
|
|
| 106 |
landmarks = load_data_from_directory(landmarks_dir, parse_text_file)
|
| 107 |
news_data = load_data_from_directory(news_dir, parse_text_file)
|
| 108 |
|
| 109 |
-
# Para municipalities, se usa el parseo de HTML
|
| 110 |
municipalities = load_data_from_directory(municipalities_dir, parse_html_file)
|
| 111 |
|
| 112 |
# =============================================================================
|
|
@@ -126,12 +139,15 @@ for landmark in landmarks:
|
|
| 126 |
icon=folium.Icon(color='blue', icon='info-sign')
|
| 127 |
).add_to(m)
|
| 128 |
|
| 129 |
-
# Agregar marcadores para Municipios
|
| 130 |
for municipality in municipalities:
|
| 131 |
if municipality["lat"] is not None and municipality["lon"] is not None:
|
|
|
|
|
|
|
|
|
|
| 132 |
folium.Marker(
|
| 133 |
location=[municipality["lat"], municipality["lon"]],
|
| 134 |
-
popup=folium.Popup(
|
| 135 |
icon=folium.Icon(color='green', icon='home')
|
| 136 |
).add_to(m)
|
| 137 |
|
|
@@ -162,18 +178,19 @@ if st.button("Enviar consulta"):
|
|
| 162 |
if user_input:
|
| 163 |
# Configura tu clave API de OpenAI (def铆nela en st.secrets)
|
| 164 |
openai.api_key = st.secrets["OPENAI_API_KEY"]
|
| 165 |
-
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
|
|
|
| 178 |
else:
|
| 179 |
st.warning("Por favor, ingresa una consulta.")
|
|
|
|
| 49 |
def parse_html_file(filepath):
|
| 50 |
"""
|
| 51 |
Parseo para archivos HTML (por ejemplo, de Wikipedia).
|
| 52 |
+
Extrae el t铆tulo, las coordenadas desde el bloque "wgCoordinates" y, si es posible,
|
| 53 |
+
la secci贸n hist贸rica (buscando un encabezado que contenga "History" o "Historia").
|
| 54 |
"""
|
| 55 |
with open(filepath, 'r', encoding='utf-8') as f:
|
| 56 |
content = f.read()
|
| 57 |
|
| 58 |
+
# Parsear con BeautifulSoup
|
| 59 |
soup = BeautifulSoup(content, "html.parser")
|
| 60 |
title = soup.title.string.strip() if soup.title else os.path.splitext(os.path.basename(filepath))[0]
|
| 61 |
|
| 62 |
+
# Extraer coordenadas (pueden venir en el bloque "wgCoordinates")
|
| 63 |
coord_match = re.search(r'"wgCoordinates":\s*{\s*"lat":\s*([0-9\.-]+),\s*"lon":\s*([0-9\.-]+)', content)
|
| 64 |
if coord_match:
|
| 65 |
lat_str = coord_match.group(1)
|
|
|
|
| 75 |
else:
|
| 76 |
lat, lon = None, None
|
| 77 |
|
| 78 |
+
# Intentar extraer datos hist贸ricos: se busca un encabezado que contenga "History" o "Historia"
|
| 79 |
+
history = ""
|
| 80 |
+
history_header = soup.find(lambda tag: tag.name in ["h2", "h3"] and ("History" in tag.get_text() or "Historia" in tag.get_text()))
|
| 81 |
+
if history_header:
|
| 82 |
+
history_parts = []
|
| 83 |
+
# Recorrer los hermanos hasta llegar a otro encabezado del mismo nivel (o superior)
|
| 84 |
+
for sibling in history_header.find_next_siblings():
|
| 85 |
+
if sibling.name and sibling.name.startswith("h"):
|
| 86 |
+
break
|
| 87 |
+
history_parts.append(sibling.get_text(strip=True))
|
| 88 |
+
history = "\n".join(history_parts)
|
| 89 |
+
|
| 90 |
return {
|
| 91 |
"name": title,
|
| 92 |
"lat": lat,
|
| 93 |
"lon": lon,
|
| 94 |
+
"content": content,
|
| 95 |
+
"history": history
|
| 96 |
}
|
| 97 |
|
| 98 |
def load_data_from_directory(directory, parser_func):
|
|
|
|
| 119 |
landmarks = load_data_from_directory(landmarks_dir, parse_text_file)
|
| 120 |
news_data = load_data_from_directory(news_dir, parse_text_file)
|
| 121 |
|
| 122 |
+
# Para municipalities, se usa el parseo de HTML (que ahora incluye datos hist贸ricos)
|
| 123 |
municipalities = load_data_from_directory(municipalities_dir, parse_html_file)
|
| 124 |
|
| 125 |
# =============================================================================
|
|
|
|
| 139 |
icon=folium.Icon(color='blue', icon='info-sign')
|
| 140 |
).add_to(m)
|
| 141 |
|
| 142 |
+
# Agregar marcadores para Municipios, incluyendo datos hist贸ricos si est谩n disponibles
|
| 143 |
for municipality in municipalities:
|
| 144 |
if municipality["lat"] is not None and municipality["lon"] is not None:
|
| 145 |
+
history_text = municipality.get("history", "")
|
| 146 |
+
history_excerpt = history_text[:300] + "..." if len(history_text) > 300 else history_text
|
| 147 |
+
popup_content = f"<b>{municipality['name']}</b><br><b>Historia:</b><br>{history_excerpt}"
|
| 148 |
folium.Marker(
|
| 149 |
location=[municipality["lat"], municipality["lon"]],
|
| 150 |
+
popup=folium.Popup(popup_content, max_width=300),
|
| 151 |
icon=folium.Icon(color='green', icon='home')
|
| 152 |
).add_to(m)
|
| 153 |
|
|
|
|
| 178 |
if user_input:
|
| 179 |
# Configura tu clave API de OpenAI (def铆nela en st.secrets)
|
| 180 |
openai.api_key = st.secrets["OPENAI_API_KEY"]
|
| 181 |
+
with st.spinner("Consultando a OpenAI..."):
|
| 182 |
+
try:
|
| 183 |
+
response = openai.ChatCompletion.create(
|
| 184 |
+
model="gpt-3.5-turbo",
|
| 185 |
+
messages=[
|
| 186 |
+
{"role": "system", "content": "Eres un asistente experto en recomendaciones tur铆sticas y culturales de Puerto Rico. Ayudas a los usuarios a descubrir lugares de inter茅s y actividades en la isla."},
|
| 187 |
+
{"role": "user", "content": user_input}
|
| 188 |
+
]
|
| 189 |
+
)
|
| 190 |
+
answer = response.choices[0].message.content.strip()
|
| 191 |
+
st.markdown("**Respuesta:**")
|
| 192 |
+
st.write(answer)
|
| 193 |
+
except Exception as e:
|
| 194 |
+
st.error(f"Error al conectarse con la API de OpenAI: {e}")
|
| 195 |
else:
|
| 196 |
st.warning("Por favor, ingresa una consulta.")
|