From b4ac40d1acfd9228297735f3e951faa54e667055 Mon Sep 17 00:00:00 2001
From: Isaac Wasserman <isaacrwasserman@gmail.com>
Date: Thu, 16 May 2024 16:16:54 -0400
Subject: [PATCH 1/2] Added fallbacks for finding description and title

---
 googlesearch/__init__.py | 65 ++++++++++++++++++++++++++++------------
 1 file changed, 46 insertions(+), 19 deletions(-)

diff --git a/googlesearch/__init__.py b/googlesearch/__init__.py
index 7efa821..6760ab7 100644
--- a/googlesearch/__init__.py
+++ b/googlesearch/__init__.py
@@ -1,4 +1,5 @@
 """googlesearch is a Python library for searching Google, easily."""
+
 from time import sleep
 from bs4 import BeautifulSoup
 from requests import get
@@ -9,9 +10,7 @@
 def _req(term, results, lang, start, proxies, timeout, safe, ssl_verify):
     resp = get(
         url="https://www.google.com/search",
-        headers={
-            "User-Agent": get_useragent()
-        },
+        headers={"User-Agent": get_useragent()},
         params={
             "q": term,
             "num": results + 2,  # Prevents multiple requests
@@ -37,10 +36,12 @@ def __repr__(self):
         return f"SearchResult(url={self.url}, title={self.title}, description={self.description})"
 
 
-def search(term, num_results=10, lang="en", proxy=None, advanced=False, sleep_interval=0, timeout=5, safe="active", ssl_verify=None):
+def search(
+    term, num_results=10, lang="en", proxy=None, advanced=False, sleep_interval=0, timeout=5, safe="active", ssl_verify=None
+):
     """Search the Google search engine"""
 
-    escaped_term = urllib.parse.quote_plus(term) # make 'site:xxx.xxx.xxx ' works.
+    escaped_term = urllib.parse.quote_plus(term)  # make 'site:xxx.xxx.xxx ' works.
 
     # Proxy
     proxies = None
@@ -54,28 +55,54 @@ def search(term, num_results=10, lang="en", proxy=None, advanced=False, sleep_in
     start = 0
     while start < num_results:
         # Send request
-        resp = _req(escaped_term, num_results - start,
-                    lang, start, proxies, timeout, safe, ssl_verify)
+        resp = _req(escaped_term, num_results - start, lang, start, proxies, timeout, safe, ssl_verify)
 
         # Parse
         soup = BeautifulSoup(resp.text, "html.parser")
+        with open("output.html", "w") as f:
+            f.write(resp.text)
+
         result_block = soup.find_all("div", attrs={"class": "g"})
-        if len(result_block) ==0:
+        if len(result_block) == 0:
+            print("Result block empty")
             start += 1
         for result in result_block:
-            # Find link, title, description
+
+            # Find link. If link is not found, we skip the result because this is the only necessary output
             link = result.find("a", href=True)
+            if not link:
+                continue
+            link = link["href"]
+
+            # Find title. If title is not found, we use the link as the title
             title = result.find("h3")
-            description_box = result.find(
-                "div", {"style": "-webkit-line-clamp:2"})
-            if description_box:
-                description = description_box.text
-                if link and title and description:
-                    start += 1
-                    if advanced:
-                        yield SearchResult(link["href"], title.text, description)
-                    else:
-                        yield link["href"]
+            if not title:
+                title = link
+            else:
+                title = title.text
+
+            # Find description. If description is not found, we attempt to reconstruct it by searching and combining spans.
+            # These messy descriptions are not perfect, but they're adequate when scraping results for AI agents.
+            description_box = result.find("div", {"style": "-webkit-line-clamp:2"})
+            description = description_box.text if description_box else None
+            if not description_box:
+                description_box_candidates = result.find_all("span")
+                # The description almost always has an <em> tag in it, so we use that as a heuristic to find the description
+                spans_with_em_child = [
+                    candidate for candidate in description_box_candidates if candidate.find("em", recursive=False)
+                ]
+                if len(spans_with_em_child) > 0:
+                    description_box = spans_with_em_child[0]
+                    description = description_box.text
+                # If we can't find an <em> tag, we just concatenate all the spans
+                else:
+                    description = "".join([span.text for span in result.find_all("span")[5:]])
+                    description = description.replace(title.text, "", 1)
+
+            if advanced:
+                yield SearchResult(link, title, description)
+            else:
+                yield link
         sleep(sleep_interval)
 
         if start == 0:

From ca27535b2c35fac9f1dedc374ac2d697b343c7a9 Mon Sep 17 00:00:00 2001
From: Isaac Wasserman <isaacrwasserman@gmail.com>
Date: Thu, 16 May 2024 16:26:54 -0400
Subject: [PATCH 2/2] removed logging

---
 googlesearch/__init__.py | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/googlesearch/__init__.py b/googlesearch/__init__.py
index 6760ab7..0849656 100644
--- a/googlesearch/__init__.py
+++ b/googlesearch/__init__.py
@@ -59,9 +59,6 @@ def search(
 
         # Parse
         soup = BeautifulSoup(resp.text, "html.parser")
-        with open("output.html", "w") as f:
-            f.write(resp.text)
-
         result_block = soup.find_all("div", attrs={"class": "g"})
         if len(result_block) == 0:
             print("Result block empty")