Update app.py

app.py

@@ -2,6 +2,7 @@ import gradio as gr
 import requests
 from bs4 import BeautifulSoup
 from urllib.parse import quote
+from itertools import combinations
 
 # Paleta de colores
 color_palette = ['#1f77b4', '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2', '#bcbd22', '#17becf', '#ff6666', '#4dff4d', '#6666ff', '#ffcc00', '#993399', '#00cc99', '#ff5050', '#33adff', '#ff9966']
@@ -29,43 +30,63 @@ def buscar_google(query, hl='es', num_results=10):
 
     return serp_data
 
-def generar_html_con_colores(keywords):
-    keywords_list = keywords.split(',')
-    serp_results = {keyword.strip(): buscar_google(keyword.strip()) for keyword in keywords_list}
-    
-    max_length = max(len(results) for results in serp_results.values()) if serp_results else 0
+def calcular_coincidencias(serps):
+    urls = [set(entry["URL"] for entry in serp) for serp in serps]
+    common_urls = set.intersection(*urls)
+    return common_urls
 
+def generar_html_con_colores(serp_results, common_urls, color_palette):
     html_table = "<style>td, th {border: 1px solid #ddd; text-align: left; padding: 8px;} tr:nth-child(even) {background-color: #f2f2f2;} th {padding-top: 11px; padding-bottom: 11px; background-color: #4CAF50; color: white;}</style>"
     html_table += "<table><tr><th>Posición</th>"
 
-    for keyword in keywords_list:
-        html_table += f"<th>{keyword.strip()}</th>"
+    keywords = list(serp_results.keys())
+    for keyword in keywords:
+        html_table += f"<th>{keyword}</th>"
     html_table += "</tr>"
 
-    url_colors = {}
+    max_length = max(len(serp_results[keyword]) for keyword in keywords)
+
     for i in range(max_length):
         html_table += f"<tr><td>{i + 1}</td>"
-        for keyword in keywords_list:
-            keyword = keyword.strip()
-            results = serp_results.get(keyword, [])
-            if i < len(results):
-                url = results[i]["URL"]
-                if url not in url_colors:
-                    url_colors[url] = color_palette[len(url_colors) % len(color_palette)]
-                color = url_colors[url]
-                html_table += f"<td><span style='color:{color}'>{results[i]['Título']}</span></td>"
+        for keyword in keywords:
+            if i < len(serp_results[keyword]):
+                entry = serp_results[keyword][i]
+                url = entry["URL"]
+                if url in common_urls:
+                    color = color_palette[0]  # Use a specific color for common URLs
+                    html_table += f"<td><span style='color:{color}'>{entry['Título']}</span></td>"
+                else:
+                    html_table += f"<td>{entry['Título']}</td>"
             else:
                 html_table += "<td></td>"
         html_table += "</tr>"
     html_table += "</table>"
     return html_table
 
+def analyze_keywords(keywords):
+    keywords_list = [keyword.strip() for keyword in keywords.split(',')]
+    serp_results = {keyword: buscar_google(keyword) for keyword in keywords_list}
+    serps = list(serp_results.values())
+    common_urls = calcular_coincidencias(serps)
+    
+    html_table = generar_html_con_colores(serp_results, common_urls, color_palette)
+
+    # Calcular y mostrar porcentajes de coincidencia
+    coincidencias_info = ""
+    for combo in combinations(keywords_list, 2):
+        serp1, serp2 = serp_results[combo[0]], serp_results[combo[1]]
+        coincidencias = calcular_coincidencias([serp1, serp2])
+        porcentaje = round(len(coincidencias) / min(len(serp1), len(serp2)) * 100, 2)
+        coincidencias_info += f"Coincidencia entre '{combo[0]}' y '{combo[1]}': {porcentaje}%<br>"
+
+    return html_table, coincidencias_info
+
 iface = gr.Interface(
-    fn=generar_html_con_colores,
+    fn=analyze_keywords,
     inputs="text",
-    outputs="html",
+    outputs=["html", "html"],
     title="Comparador de Keywords en Google",
-    description="Introduce las keywords separadas por comas para encontrar coincidencias en los resultados de búsqueda de Google."
+    description="Introduce las keywords separadas por comas para encontrar coincidencias en los resultados de búsqueda de Google y calcular el porcentaje de similaridad entre ellos."
 )
 
 iface.launch()