Update seo_keywords.py

seo_keywords.py

@@ -1,142 +1,139 @@
-from langchain_community.tools import TavilySearchResults
-from langchain_google_genai import ChatGoogleGenerativeAI
-from langchain_core.prompts import PromptTemplate
-from tavily import TavilyClient
-import asyncio
-import sys
-
-import os
-from dotenv import load_dotenv
-
-# files
-from crawl import seo_crawling
-
-# Secret Key
-load_dotenv(override=True)
-tavily_api_key = os.getenv("TAVILY_API_KEY")
-gemini_api_key = os.getenv("GEMINI_API_KEY")
-
-
-llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash-exp", api_key = gemini_api_key) # type: ignore
-
-
-# def tavily_search(query: str):
-    
-#     tavily = TavilySearchResults(
-#         max_results=10,
-#         search_depth="advanced",
-#         include_answer=True,
-#         include_images=True,
-#         include_links=True, # type: ignore
-#         api_key=tavily_api_key, # type: ignore
-#     )
-    
-#     results = tavily.invoke({"query": f"{query}"})
-#     return results
-
-# results = tavily_search("UK Air Source Heat Pump Market Trends 2025")
-# print(results)
-
-
-def tavily_search(query):
-    tavily_client = TavilyClient(api_key = tavily_api_key)
-    response = tavily_client.search(query, max_results=10)
-    # print(response["results"])
-    return response["results"]
-
-
-def match_title(title, titles):
-    title_prompt_template = PromptTemplate.from_template("""
-        Your task is to find the title in the List that semantically matches the User_title.
-        - Don't change the title name
-        - Don't give extra content. Only give the title name.
-        - Only give **One title**
-
-        List = {list}
-        User_title = {title}
-
-    """)
-
-    prompt = title_prompt_template.invoke({"list": titles, "title": title})
-    response = llm.invoke(prompt)
-    return response.content
-
-
-def keywords(content):
-    keywords_prompt_template = PromptTemplate.from_template("""
-    # **Instruction:**
-    Analyze the given text and extract keywords based on their relevance to SEO. Categorize them into the following three groups:
-
-    ## **1. Primary Keywords (High-Impact, Industry-Specific):**
-       - Broad, high-volume search terms that are directly related to the main topic.
-       - Common industry terms that people search for when looking for services or information.
-       - Maximum **10-12 keywords**.
-
-    ## **2. Secondary Keywords (Supporting SEO & Long-Tail Queries):**
-       - More specific, longer phrases related to the main topic.
-       - Keywords that provide contextual depth and support for primary keywords.
-       - Maximum **10-12 keywords**.
-
-    ## **3. Local SEO Keywords (Boosting Regional Visibility):**
-       - Keywords that include location-specific terms.
-       - Phrases that help rank in local search results.
-       - Maximum **5-8 keywords**.
-       
-    # Don't give the extra content only give the SEO keywords
-
-    # **Input:**
-    {text}
-
-    """)
-    
-    
-    prompt = keywords_prompt_template.invoke({"text": content})
-    response = llm.invoke(prompt)
-    return response.content
-
-
-
-
-def seo_keywords(state):
-    
-    topic = state["final_topic"]
-    # topic = state
-    
-    results = tavily_search(topic)
-
-    titles = []
-    titles_url = []
-    final_url = ""
-    
-    for t in results:
-        titles.append(t['title'])
-    
-    for t in results:
-        titles_url.append({
-            "title": t['title'],
-            "url": t['url']
-        })  
-    
-    print(titles)
-    print(titles_url)
-   
-    text = match_title(topic, titles)
-    print(text)
-
-
-    for title in titles_url:
-        if title['title'] == text:
-            final_url = title['url']
-
-    print(final_url)
-    
-    if sys.platform == "win32":
-        asyncio.set_event_loop_policy(asyncio.WindowsProactorEventLoopPolicy())
-        
-    crawled_content = asyncio.run(seo_crawling(final_url))
-    response = keywords(crawled_content)
-    print(response)
-    return{"seo_keywords":response}
-
-# a = seo_keywords("UK Air Source Heat Pump Market Trends 2025")
+from langchain_community.tools import TavilySearchResults
+from langchain_google_genai import ChatGoogleGenerativeAI
+from langchain_core.prompts import PromptTemplate
+from tavily import TavilyClient
+import asyncio
+import sys
+
+import os
+from dotenv import load_dotenv
+
+# files
+from crawl import seo_crawling
+
+# Secret Key
+load_dotenv(override=True)
+tavily_api_key = os.getenv("TAVILY_API_KEY")
+gemini_api_key = os.getenv("GEMINI_API_KEY")
+
+
+llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash-exp", api_key = gemini_api_key) # type: ignore
+
+
+# def tavily_search(query: str):
+    
+#     tavily = TavilySearchResults(
+#         max_results=10,
+#         search_depth="advanced",
+#         include_answer=True,
+#         include_images=True,
+#         include_links=True, # type: ignore
+#         api_key=tavily_api_key, # type: ignore
+#     )
+    
+#     results = tavily.invoke({"query": f"{query}"})
+#     return results
+
+# results = tavily_search("UK Air Source Heat Pump Market Trends 2025")
+# print(results)
+
+
+def tavily_search(query):
+    tavily_client = TavilyClient(api_key = tavily_api_key)
+    response = tavily_client.search(query, max_results=10)
+    # print(response["results"])
+    return response["results"]
+
+
+def match_title(title, titles):
+    title_prompt_template = PromptTemplate.from_template("""
+        Your task is to find the title in the List that semantically matches the User_title.
+        - Don't change the title name
+        - Don't give extra content. Only give the title name.
+        - Only give **One title**
+
+        List = {list}
+        User_title = {title}
+
+    """)
+
+    prompt = title_prompt_template.invoke({"list": titles, "title": title})
+    response = llm.invoke(prompt)
+    return response.content
+
+
+def keywords(content):
+    keywords_prompt_template = PromptTemplate.from_template("""
+    # **Instruction:**
+    Analyze the given text and extract keywords based on their relevance to SEO. Categorize them into the following three groups:
+
+    ## **1. Primary Keywords (High-Impact, Industry-Specific):**
+       - Broad, high-volume search terms that are directly related to the main topic.
+       - Common industry terms that people search for when looking for services or information.
+       - Maximum **10-12 keywords**.
+
+    ## **2. Secondary Keywords (Supporting SEO & Long-Tail Queries):**
+       - More specific, longer phrases related to the main topic.
+       - Keywords that provide contextual depth and support for primary keywords.
+       - Maximum **10-12 keywords**.
+
+    ## **3. Local SEO Keywords (Boosting Regional Visibility):**
+       - Keywords that include location-specific terms.
+       - Phrases that help rank in local search results.
+       - Maximum **5-8 keywords**.
+       
+    # Don't give the extra content only give the SEO keywords
+
+    # **Input:**
+    {text}
+
+    """)
+    
+    
+    prompt = keywords_prompt_template.invoke({"text": content})
+    response = llm.invoke(prompt)
+    return response.content
+
+
+
+
+def seo_keywords(state):
+    
+    topic = state["final_topic"]
+    # topic = state
+    
+    results = tavily_search(topic)
+
+    titles = []
+    titles_url = []
+    final_url = ""
+    
+    for t in results:
+        titles.append(t['title'])
+    
+    for t in results:
+        titles_url.append({
+            "title": t['title'],
+            "url": t['url']
+        })  
+    
+    print(titles)
+    print(titles_url)
+   
+    text = match_title(topic, titles)
+    print(text)
+
+
+    for title in titles_url:
+        if title['title'] == text:
+            final_url = title['url']
+
+    print(final_url)
+        
+    crawled_content = asyncio.run(seo_crawling(final_url))
+    response = keywords(crawled_content)
+    print(response)
+    return{"seo_keywords":response}
+
+# a = seo_keywords("UK Air Source Heat Pump Market Trends 2025")
 # print(a)