Implement Google Search for Answers to Questions

examples/google_search_example.md

@@ -0,0 +1,111 @@
+# Google Search Bot Feature Examples
+
+This document demonstrates how the Google search functionality works in the VK bot.
+
+## Usage Examples
+
+### Basic Search Commands
+
+```
+search python list comprehension
+google javascript async await  
+найди react hooks tutorial
+поиск how to use git
+```
+
+### Command Structure
+
+- **Trigger words**: `search`, `google`, `найди`, `поиск`
+- **Minimum words**: 3 words required in the query
+- **Maximum results**: Up to 3 links returned
+- **Whitelisted sites only**: Results filtered to trusted programming sites
+
+### Response Format
+
+When you send: `search python list comprehension`
+
+The bot responds with:
+```
+Результаты поиска для 'python list comprehension':
+
+1. https://docs.python.org/3/tutorial/datastructures.html#list-comprehensions
+2. https://stackoverflow.com/questions/34835951/what-does-list-comprehension-mean-how-does-it-work-and-how-can-i-use-it
+3. https://medium.com/@python_guide/python-list-comprehensions-explained-765fb6ca5c8a
+```
+
+### Error Cases
+
+**Too few words:**
+```
+User: search python
+Bot: Пожалуйста, используйте не менее 3 слов для поиска.
+```
+
+**No whitelisted results found:**
+```
+User: search obscure topic nobody talks about
+Bot: К сожалению, не найдено ссылок с проверенных сайтов для запроса 'obscure topic nobody talks about'.
+```
+
+**Network error:**
+```
+User: search python programming
+Bot: Произошла ошибка при поиске. Попробуйте позже.
+```
+
+## Whitelisted Sites
+
+The bot only returns results from these trusted programming sites:
+
+- stackoverflow.com
+- github.com
+- docs.python.org
+- developer.mozilla.org
+- w3schools.com
+- medium.com
+- dev.to
+- geeksforgeeks.org
+- tutorialspoint.com
+- programiz.com
+
+## Technical Implementation
+
+### Configuration
+
+The feature is configured in `config.py`:
+
+```python
+GOOGLE_SEARCH_WHITELISTED_SITES = [
+    'stackoverflow.com',
+    'github.com',
+    # ... other trusted sites
+]
+GOOGLE_SEARCH_MIN_WORDS = 3
+GOOGLE_SEARCH_MAX_RESULTS = 3
+GOOGLE_SEARCH_TIMEOUT = 10  # seconds
+```
+
+### Pattern Matching
+
+The command is recognized using a regex pattern in `patterns.py`:
+
+```python
+GOOGLE_SEARCH = recompile(
+    r'\A\s*(search|найди|поиск|google)\s+(?P<query>[\S][\S\s]*?)\??\s*\Z', IGNORECASE)
+```
+
+### Search Process
+
+1. **Query Validation**: Check minimum word count
+2. **Google Search**: Make HTTP request to Google with proper headers
+3. **Link Extraction**: Parse HTML to find result URLs
+4. **Whitelist Filtering**: Keep only links from trusted domains
+5. **Relevance Scoring**: Prioritize results with more matching query words
+6. **Response Formatting**: Send formatted results to user
+
+### Security Features
+
+- **Request Limiting**: Built-in timeout protection
+- **Domain Filtering**: Only whitelisted domains returned
+- **Query Sanitization**: URLs properly encoded
+- **Error Handling**: Graceful fallback for network issues

python/__main__.py

@@ -63,7 +63,8 @@ def __init__(
             (patterns.WHAT_IS, self.commands.what_is),
             (patterns.WHAT_MEAN, self.commands.what_is),
             (patterns.APPLY_KARMA, self.commands.apply_karma),
-            (patterns.GITHUB_COPILOT, self.commands.github_copilot)
+            (patterns.GITHUB_COPILOT, self.commands.github_copilot),
+            (patterns.GOOGLE_SEARCH, self.commands.google_search)
         )
 
     def message_new(

python/config.py

@@ -153,5 +153,22 @@
 GITHUB_COPILOT_RUN_COMMAND = 'bash -c "./copilot.sh {input_file} {output_file}"'
 GITHUB_COPILOT_TIMEOUT = 120  # seconds
 
+# Google search configuration
+GOOGLE_SEARCH_WHITELISTED_SITES = [
+    'stackoverflow.com',
+    'github.com',
+    'docs.python.org',
+    'developer.mozilla.org',
+    'w3schools.com',
+    'medium.com',
+    'dev.to',
+    'geeksforgeeks.org',
+    'tutorialspoint.com',
+    'programiz.com'
+]
+GOOGLE_SEARCH_MIN_WORDS = 3
+GOOGLE_SEARCH_MAX_RESULTS = 3
+GOOGLE_SEARCH_TIMEOUT = 10  # seconds
+
 DEFAULT_PROGRAMMING_LANGUAGES_PATTERN_STRING = "|".join(DEFAULT_PROGRAMMING_LANGUAGES)
 GITHUB_COPILOT_LANGUAGES_PATTERN_STRING = "|".join([i for i in GITHUB_COPILOT_LANGUAGES.keys()])

python/modules/commands.py

@@ -5,10 +5,11 @@
 import os
 
 from regex import Pattern, Match, split, match, search, IGNORECASE, sub
-from requests import post
+from requests import post, get
 from social_ethosa import BetterUser
 from saya import Vk
 import wikipedia
+import urllib.parse
 
 from .commands_builder import CommandsBuilder
 from .data_service import BetterBotBaseDataService
@@ -379,6 +380,100 @@ def github_copilot(self) -> NoReturn:
             f'Пожалуйста, подождите {round(config.GITHUB_COPILOT_TIMEOUT - (now - self.now))} секунд', self.peer_id
         )
 
+    def google_search(self) -> NoReturn:
+        """Search Google for answers and return whitelisted links"""
+        query = self.matched.group('query').strip()
+
+        # Check minimum word count
+        words = query.split()
+        if len(words) < config.GOOGLE_SEARCH_MIN_WORDS:
+            self.vk_instance.send_msg(
+                f'Пожалуйста, используйте не менее {config.GOOGLE_SEARCH_MIN_WORDS} слов для поиска.',
+                self.peer_id
+            )
+            return
+
+        try:
+            # Build Google search URL
+            search_url = f"https://www.google.com/search?q={urllib.parse.quote(query)}&num=20"
+
+            # Set headers to mimic a browser
+            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'
+            }
+
+            # Make the search request
+            response = get(search_url, headers=headers, timeout=config.GOOGLE_SEARCH_TIMEOUT)
+            response.raise_for_status()
+
+            # Extract links from search results using multiple patterns
+            import re
+
+            # Try multiple patterns to extract URLs from Google results
+            url_patterns = [
+                r'href="(/url\?q=[^"]+)"',  # Standard Google redirect
+                r'href="(https?://[^"]*)"',  # Direct URLs
+                r'data-href="([^"]+)"',     # Alternative href attribute
+            ]
+
+            all_links = []
+            for pattern in url_patterns:
+                matches = re.findall(pattern, response.text)
+                all_links.extend(matches)
+
+            filtered_links = []
+            seen_domains = set()
+
+            for link in all_links:
+                if len(filtered_links) >= config.GOOGLE_SEARCH_MAX_RESULTS:
+                    break
+
+                # Clean up URL
+                actual_url = link
+                if '/url?q=' in link:
+                    # Extract from Google redirect
+                    try:
+                        actual_url = link.split('/url?q=')[1].split('&')[0]
+                        actual_url = urllib.parse.unquote(actual_url)
+                    except:
+                        continue
+                elif link.startswith('/'):
+                    # Skip relative URLs
+                    continue
+
+                # Validate URL format
+                if not actual_url.startswith(('http://', 'https://')):
+                    continue
+
+                # Check if URL is from whitelisted domain
+                for whitelisted_site in config.GOOGLE_SEARCH_WHITELISTED_SITES:
+                    if whitelisted_site in actual_url and whitelisted_site not in seen_domains:
+                        # Count matching words in URL for relevance
+                        word_matches = sum(1 for word in words if word.lower() in actual_url.lower())
+                        if word_matches >= 0:  # Allow URLs even without exact word matches
+                            filtered_links.append(actual_url)
+                            seen_domains.add(whitelisted_site)
+                            break
+
+            # Send results
+            if filtered_links:
+                result_message = f"Результаты поиска для '{query}':\n\n"
+                for i, link in enumerate(filtered_links, 1):
+                    result_message += f"{i}. {link}\n"
+                self.vk_instance.send_msg(result_message, self.peer_id)
+            else:
+                self.vk_instance.send_msg(
+                    f"К сожалению, не найдено ссылок с проверенных сайтов для запроса '{query}'.",
+                    self.peer_id
+                )
+
+        except Exception as e:
+            print(f"Google search error: {e}")
+            self.vk_instance.send_msg(
+                "Произошла ошибка при поиске. Попробуйте позже.",
+                self.peer_id
+            )
+
     def match_command(
             self,
             pattern: Pattern

python/patterns.py

@@ -65,3 +65,7 @@
 GITHUB_COPILOT = recompile(
     r'\A\s*(code|код)\s+(?P<lang>(' + COPILOT_LANGUAGES +
     r'))(?P<text>[\S\s]+)\Z', IGNORECASE)
+
+# Google search pattern
+GOOGLE_SEARCH = recompile(
+    r'\A\s*(search|найди|поиск|google)\s+(?P<query>[\S][\S\s]*?)\??\s*\Z', IGNORECASE)