""" title: GPT-4o Search Preview Companion id: gpt4o_search_preview_companion description: A companion filter that intercepts search-enabled messages and reroutes them to OpenAI's GPT-4o Search Preview model. author: Justin Kropp (inspired by suurt8ll's Google Companion filter) funding_url: https://github.com/suurt8ll/open_webui_functions license: MIT version: 1.1.2 This Open WebUI filter reroutes requests to OpenAI's gpt-4o-search-preview model when the built-in web search feature is active. It disables Open WebUI's native search handler, injects search context and timezone, and emits cleaned citation markers using URLs from the assistant's final message. Change Log: v1.1.2 - Improved code structure and readability v1.1.1 - Fixed final status not always displaying - Refactored outlet handler for better flow and readability v1.1.0 - Switched citation capture from streaming to post-response outlet. - Removed `stream()` method to improve performance - Removed OpenAI tracking parameters from URLs for clean URLs. - Skipped duplicate citations - Added fallback message when no links are found TODO: - Add fallback logic for detecting links without utm_source - Allow dynamic country/region detection - Improve log granularity using structured debug levels """ import re from urllib.parse import urlparse, urlunparse from pydantic import BaseModel, Field class Filter: class Valves(BaseModel): ENABLE_LOGGING: bool = Field( default=False, description="Enable debug logging output." ) SEARCH_CONTEXT_SIZE: str = Field( default="medium", description="Granularity of search context (low, medium, high).", ) def __init__(self): self.valves = self.Valves() self._search_rerouted = False if self.valves.ENABLE_LOGGING: print("[gpt-4o-search-preview] ✅ Filter initialized") async def inlet(self, body: dict, __event_emitter__=None) -> dict: metadata = body.get("metadata", {}) features = metadata.get("features", {}) # Exit fast if web search isn't active if not features.get("web_search", False): return body self._search_rerouted = True if __event_emitter__: await __event_emitter__( { "type": "status", "data": { "description": "🔍 Web search detected — rerouting to GPT-4o Search Preview...", "done": False, "hidden": False, }, } ) # 🧠 Override native search and explicitly set GPT-4o route features["web_search"] = False body["model"] = "gpt-4o-search-preview" # Use chat metadata or a default timezone timezone = metadata.get("variables", {}).get( "{{CURRENT_TIMEZONE}}", "America/Vancouver" ) body["web_search_options"] = { "user_location": { "type": "approximate", "approximate": {"country": "CA", "timezone": timezone}, }, "search_context_size": self.valves.SEARCH_CONTEXT_SIZE.lower(), } return body async def outlet(self, body: dict, __event_emitter__=None) -> dict: if not self._search_rerouted: return body # This message didn't trigger search reroute # Safely get the last assistant message, if it's the final message in the list messages = body.get("messages", []) if not messages or messages[-1].get("role") != "assistant": return body content = messages[-1].get("content", "") if not content: return body # Extract and clean tracked OpenAI URLs matches = re.findall(r"https?://[^\s)\"'>]+?\?utm_source=openai", content) citation_urls = { urlunparse(urlparse(url)._replace(query="")) for url in matches } if self.valves.ENABLE_LOGGING: print( f"[gpt-4o-search-preview][outlet] 🔗 {len(citation_urls)} unique URLs found" ) if __event_emitter__: # Emit each citation for url in sorted(citation_urls): await __event_emitter__( { "type": "citation", "data": { "document": [url], "metadata": [{"source": url, "html": False}], "source": {"name": url}, }, } ) # Emit final status message citation_count = len(citation_urls) status_message = ( f"✅ Web search complete — {citation_count} source{'s' if citation_count != 1 else ''} cited." if citation_urls else "Search not used — answer based on model's internal knowledge." ) await __event_emitter__( { "type": "status", "data": { "description": status_message, "done": True, "hidden": False, }, } ) self._search_rerouted = False # Clear internal reroute flag for next cycle return body