""" title: Gemini Manifold Companion id: gemini_manifold_companion description: A companion filter for "Gemini Manifold google_genai" pipe providing enhanced functionality. author: suurt8ll author_url: https://github.com/suurt8ll funding_url: https://github.com/suurt8ll/open_webui_functions license: MIT version: 1.7.0 """ VERSION = "1.7.0" # This filter can detect that a feature like web search or code execution is enabled in the front-end, # set the feature back to False so Open WebUI does not run it's own logic and then # pass custom values to "Gemini Manifold google_genai" that signal which feature was enabled and intercepted. import copy import json from google.genai import types import sys import time import asyncio import aiohttp from fastapi import Request from fastapi.datastructures import State from loguru import logger from pydantic import BaseModel, Field import pydantic_core from collections.abc import Awaitable, Callable from typing import Any, Literal, TYPE_CHECKING, cast from open_webui.models.functions import Functions if TYPE_CHECKING: from loguru import Record from loguru._handler import Handler # type: ignore from utils.manifold_types import * # My personal types in a separate file for more robustness. # According to https://ai.google.dev/gemini-api/docs/models ALLOWED_GROUNDING_MODELS = { "gemini-2.5-pro", "gemini-flash-latest", "gemini-2.5-flash-preview-09-2025", "gemini-2.5-flash", "gemini-flash-lite-latest", "gemini-2.5-flash-lite-preview-09-2025", "gemini-2.5-flash-lite", "gemini-2.5-flash-lite-preview-06-17", "gemini-2.5-pro-preview-06-05", "gemini-2.5-flash-preview-05-20", "gemini-2.5-pro-preview-05-06", "gemini-2.5-flash-preview-04-17", "gemini-2.5-pro-preview-03-25", "gemini-2.5-pro-exp-03-25", "gemini-2.0-pro-exp", "gemini-2.0-pro-exp-02-05", "gemini-exp-1206", "gemini-2.0-flash", "gemini-2.0-flash-exp", "gemini-2.0-flash-001", "gemini-1.5-pro", "gemini-1.5-flash", "gemini-1.0-pro", } ALLOWED_CODE_EXECUTION_MODELS = { "gemini-2.5-pro", "gemini-flash-latest", "gemini-2.5-flash-preview-09-2025", "gemini-2.5-flash", "gemini-flash-lite-latest", "gemini-2.5-flash-lite-preview-09-2025", "gemini-2.5-flash-lite", "gemini-2.5-flash-lite-preview-06-17", "gemini-2.5-pro-preview-06-05", "gemini-2.5-flash-preview-05-20", "gemini-2.5-pro-preview-05-06", "gemini-2.5-flash-preview-04-17", "gemini-2.5-pro-preview-03-25", "gemini-2.5-pro-exp-03-25", "gemini-2.0-pro-exp", "gemini-2.0-pro-exp-02-05", "gemini-exp-1206", "gemini-2.0-flash-thinking-exp-01-21", "gemini-2.0-flash", "gemini-2.0-flash-exp", "gemini-2.0-flash-001", } # Default timeout for URL resolution # TODO: Move to Pipe.Valves. DEFAULT_URL_TIMEOUT = aiohttp.ClientTimeout(total=10) # 10 seconds total timeout # Setting auditable=False avoids duplicate output for log levels that would be printed out by the main log. log = logger.bind(auditable=False) class Filter: class Valves(BaseModel): FORCE_NON_STREAM_FOR_IMAGE_MODELS: bool = Field( default=True, description="""Automatically disable streaming for image generation models (e.g., gemini-2.5-flash-image-preview) to prevent 'Chunk too big' errors. Set to False to attempt streaming with these models.""", ) SET_TEMP_TO_ZERO: bool = Field( default=False, description="""Decide if you want to set the temperature to 0 for grounded answers, Google reccomends it in their docs.""", ) GROUNDING_DYNAMIC_RETRIEVAL_THRESHOLD: float | None = Field( default=None, description="""See https://ai.google.dev/gemini-api/docs/grounding?lang=python#dynamic-threshold for more information. Only supported for 1.0 and 1.5 models""", ) USE_PERMISSIVE_SAFETY: bool = Field( default=False, description="""Whether to request relaxed safety filtering. Default value is False.""", ) BYPASS_BACKEND_RAG: bool = Field( default=True, description="""Decide if you want ot bypass Open WebUI's RAG and send your documents directly to Google API. Default value is True.""", ) LOG_LEVEL: Literal[ "TRACE", "DEBUG", "INFO", "SUCCESS", "WARNING", "ERROR", "CRITICAL" ] = Field( default="INFO", description="Select logging level. Use `docker logs -f open-webui` to view logs.", ) # TODO: Support user settting through UserValves. def __init__(self): # This hack makes the valves values available to the `__init__` method. # TODO: Get the id from the frontmatter instead of hardcoding it. valves = Functions.get_function_valves_by_id("gemini_manifold_companion") self.valves = self.Valves(**(valves if valves else {})) self.log_level = self.valves.LOG_LEVEL self._add_log_handler() log.success("Function has been initialized.") log.trace("Full self object:", payload=self.__dict__) def inlet(self, body: "Body", __metadata__: dict[str, Any]) -> "Body": """Modifies the incoming request payload before it's sent to the LLM. Operates on the `form_data` dictionary.""" # Detect log level change inside self.valves if self.log_level != self.valves.LOG_LEVEL: log.info( f"Detected log level change: {self.log_level=} and {self.valves.LOG_LEVEL=}. " "Running the logging setup again." ) self._add_log_handler() log.debug( f"inlet method has been called. Gemini Manifold Companion version is {VERSION}" ) canonical_model_name, is_manifold = self._get_model_name(body) # Exit early if we are filtering an unsupported model. if not is_manifold: log.debug( "Returning the original body object because conditions for proceeding are not fulfilled." ) return body # Check if it's a relevant model (supports either feature) is_grounding_model = canonical_model_name in ALLOWED_GROUNDING_MODELS is_code_exec_model = canonical_model_name in ALLOWED_CODE_EXECUTION_MODELS log.debug(f"{is_grounding_model=}, {is_code_exec_model=}") features = body.get("features", {}) log.debug(f"body.features:", payload=features) # Ensure features field exists metadata = body.get("metadata") metadata_features = metadata.get("features") if metadata_features is None: metadata_features = cast(Features, {}) metadata["features"] = metadata_features # Add the companion version to the payload for the pipe to consume. metadata_features["gemini_manifold_companion_version"] = VERSION if is_grounding_model: web_search_enabled = ( features.get("web_search", False) if isinstance(features, dict) else False ) if web_search_enabled: log.info( "Search feature is enabled, disabling it and adding custom feature called grounding_w_google_search." ) # Disable web_search features["web_search"] = False # Use "Google Search Retrieval" for 1.0 and 1.5 models and "Google Search as a Tool for >=2.0 models". if "1.0" in canonical_model_name or "1.5" in canonical_model_name: metadata_features["google_search_retrieval"] = True metadata_features["google_search_retrieval_threshold"] = ( self.valves.GROUNDING_DYNAMIC_RETRIEVAL_THRESHOLD ) else: metadata_features["google_search_tool"] = True # Google suggest setting temperature to 0 if using grounding: # https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/ground-with-google-search#:~:text=For%20ideal%20results%2C%20use%20a%20temperature%20of%200.0. if self.valves.SET_TEMP_TO_ZERO: log.info("Setting temperature to 0.") body["temperature"] = 0 # type: ignore if is_code_exec_model: code_execution_enabled = ( features.get("code_interpreter", False) if isinstance(features, dict) else False ) if code_execution_enabled: log.info( "Code interpreter feature is enabled, disabling it and adding custom feature called google_code_execution." ) # Disable code_interpreter features["code_interpreter"] = False metadata_features["google_code_execution"] = True if self.valves.USE_PERMISSIVE_SAFETY: log.info("Adding permissive safety settings to body.metadata") metadata["safety_settings"] = self._get_permissive_safety_settings( canonical_model_name ) if self.valves.BYPASS_BACKEND_RAG: if __metadata__["chat_id"] == "local": # TODO toast notification log.warning( "Bypassing Open WebUI's RAG is not possible for temporary chats. " "The Manifold pipe requires a database entry to access uploaded files, " "which temporary chats do not have. Falling back to Open WebUI's RAG." ) metadata_features["upload_documents"] = False else: log.info( "BYPASS_BACKEND_RAG is enabled, bypassing Open WebUI RAG to let the Manifold pipe handle documents." ) if files := body.get("files"): log.info( f"Removing {len(files)} files from the Open WebUI RAG pipeline." ) body["files"] = [] metadata_features["upload_documents"] = True else: log.info( "BYPASS_BACKEND_RAG is disabled. Open WebUI's RAG will be used if applicable." ) metadata_features["upload_documents"] = False # The manifold pipe requires the backend to be in streaming mode to correctly # process the AsyncGenerator it returns. We save the user's original # streaming intent and then force the backend into streaming mode. user_stream_intent = body.get("stream", True) image_generation_models = { "gemini-2.0-flash-preview-image-generation", "gemini-2.5-flash-image-preview", "gemini-2.5-flash-image", } # Check if the current model is an image generation model and if the # user has enabled the non-streaming override for them. if ( self.valves.FORCE_NON_STREAM_FOR_IMAGE_MODELS and canonical_model_name in image_generation_models ): log.info( f"Image generation model '{canonical_model_name}' detected. " "Forcing non-streaming mode to prevent potential 'Chunk too big' errors." ) # Override the user's intent to ensure stability. user_stream_intent = False log.info( f"Storing user's stream intent ({user_stream_intent}) into __metadata__. " "Backend will be forced down the streaming path." ) metadata_features["stream"] = user_stream_intent body["stream"] = True # TODO: Filter out the citation markers here. log.debug("inlet method has finished.") return body def stream(self, event: dict) -> dict: """Modifies the streaming response from the LLM in real-time. Operates on individual chunks of data.""" return event async def outlet( self, body: "Body", __request__: Request, __metadata__: dict[str, Any], __event_emitter__: Callable[["Event"], Awaitable[None]], ) -> "Body": """Modifies the complete response payload after it's received from the LLM. Operates on the final `body` dictionary.""" log.debug("outlet method has been called.") chat_id: str = __metadata__.get("chat_id", "") message_id: str = __metadata__.get("message_id", "") grounding_key = f"grounding_{chat_id}_{message_id}" time_key = f"pipe_start_time_{chat_id}_{message_id}" app_state: State = __request__.app.state log.debug(f"Checking for attributes for message {message_id} in request state.") stored_metadata: types.GroundingMetadata | None = getattr( app_state, grounding_key, None ) pipe_start_time: float | None = getattr(app_state, time_key, None) if stored_metadata: log.info("Found grounding metadata, processing citations.") log.trace("Stored grounding metadata:", payload=stored_metadata) current_content = body["messages"][-1]["content"] if isinstance(current_content, list): text_to_use = "" for item in current_content: if item.get("type") == "text": item = cast("TextContent", item) text_to_use = item["text"] break else: text_to_use = current_content # Insert citation markers into the response text cited_text = self._get_text_w_citation_markers( stored_metadata, text_to_use, ) if cited_text: content = body["messages"][-1]["content"] if isinstance(content, list): for item in content: if item.get("type") == "text": item = cast("TextContent", item) item["text"] = cited_text break else: body["messages"][-1]["content"] = cited_text # Emit sources to the front-end. gs_supports = stored_metadata.grounding_supports gs_chunks = stored_metadata.grounding_chunks if gs_supports and gs_chunks: await self._resolve_and_emit_sources( grounding_chunks=gs_chunks, supports=gs_supports, event_emitter=__event_emitter__, pipe_start_time=pipe_start_time, ) else: log.info( "Grounding metadata missing supports or chunks (checked in outlet); " "skipping source resolution and emission." ) # Emit status event with search queries await self._emit_status_event_w_queries(stored_metadata, __event_emitter__) # Clean up state delattr(app_state, grounding_key) if hasattr(app_state, time_key): delattr(app_state, time_key) else: log.info("No grounding metadata found in request state.") log.debug("outlet method has finished.") return body # region 1. Helper methods inside the Filter class # region 1.1 Add citations def _get_text_w_citation_markers( self, grounding_metadata: types.GroundingMetadata, raw_str: str, ) -> str | None: """ Returns the model response with citation markers. Thoughts, if present as THOUGHT_START_TAG...THOUGHT_END_TAG at the beginning of raw_str, are preserved but excluded from the citation indexing process. Everything up to the *last* THOUGHT_END_TAG tag is considered part of the thought. """ supports = grounding_metadata.grounding_supports grounding_chunks = grounding_metadata.grounding_chunks if not supports or not grounding_chunks: log.info( "Grounding metadata missing supports or chunks, can't insert citation markers. " "Response was probably just not grounded." ) return None log.trace("raw_str:", payload=raw_str, _log_truncation_enabled=False) thought_prefix = "" content_for_citation_processing = raw_str THOUGHT_START_TAG = "<details" THOUGHT_END_TAG = "</details>\n" if raw_str.startswith(THOUGHT_START_TAG): last_end_thought_tag_idx = raw_str.rfind(THOUGHT_END_TAG) if ( last_end_thought_tag_idx != -1 and last_end_thought_tag_idx >= len(THOUGHT_START_TAG) - 1 ): thought_block_end_offset = last_end_thought_tag_idx + len( THOUGHT_END_TAG ) thought_prefix = raw_str[:thought_block_end_offset] content_for_citation_processing = raw_str[thought_block_end_offset:] log.info( "Model thoughts detected at the beginning of the response. " "Citations will be processed on the content following the last thought block." ) else: log.warning( "Detected THOUGHT_START_TAG at the start of raw_str without a subsequent closing THOUGHT_END_TAG " "or a malformed thought block. The entire raw_str will be processed for citations. " "This might lead to incorrect marker placement if thoughts were intended and indices " "are relative to content after thoughts." ) processed_content_part_with_markers = content_for_citation_processing if content_for_citation_processing: try: modified_content_bytes = bytearray( content_for_citation_processing.encode("utf-8") ) for support in reversed(supports): segment = support.segment indices = support.grounding_chunk_indices if not ( indices is not None and segment and segment.end_index is not None ): log.debug(f"Skipping support due to missing data: {support}") continue end_pos = segment.end_index if not (0 <= end_pos <= len(modified_content_bytes)): log.warning( f"Support segment end_index ({end_pos}) is out of bounds for the processable content " f"(length {len(modified_content_bytes)} bytes after potential thought stripping). " f"Content (first 50 chars): '{content_for_citation_processing[:50]}...'. Skipping this support. Support: {support}" ) continue citation_markers = "".join(f"[{index + 1}]" for index in indices) encoded_citation_markers = citation_markers.encode("utf-8") modified_content_bytes[end_pos:end_pos] = encoded_citation_markers processed_content_part_with_markers = modified_content_bytes.decode( "utf-8" ) except Exception as e: log.error( f"Error injecting citation markers into content: {e}. " f"Using content part (after potential thought stripping) without new markers." ) else: if raw_str and not content_for_citation_processing: log.info( "Content for citation processing is empty (e.g., raw_str contained only thoughts). " "No citation markers will be injected." ) elif not raw_str: log.warning("Raw string is empty, cannot inject citation markers.") final_result_str = thought_prefix + processed_content_part_with_markers return final_result_str async def _resolve_url( self, session: aiohttp.ClientSession, url: str, timeout: aiohttp.ClientTimeout = DEFAULT_URL_TIMEOUT, max_retries: int = 3, base_delay: float = 0.5, ) -> tuple[str, bool]: """ Resolves a given URL using the provided aiohttp session, with multiple retries on failure. Returns the final URL and a boolean indicating success. """ if not url: return "", False for attempt in range(max_retries + 1): try: async with session.get( url, allow_redirects=True, timeout=timeout, ) as response: final_url = str(response.url) log.debug( f"Resolved URL '{url}' to '{final_url}' after {attempt} retries" ) return final_url, True except (asyncio.TimeoutError, aiohttp.ClientError) as e: if attempt == max_retries: log.error( f"Failed to resolve URL '{url}' after {max_retries + 1} attempts: {e}" ) return url, False else: delay = min(base_delay * (2**attempt), 10.0) log.warning( f"Retry {attempt + 1}/{max_retries + 1} for URL '{url}': {e}. Waiting {delay:.1f}s..." ) await asyncio.sleep(delay) except Exception as e: log.error(f"Unexpected error resolving URL '{url}': {e}") return url, False return url, False async def _resolve_and_emit_sources( self, grounding_chunks: list[types.GroundingChunk], supports: list[types.GroundingSupport], event_emitter: Callable[["Event"], Awaitable[None]], pipe_start_time: float | None, ): """ Resolves URLs in the background and emits a chat completion event containing only the source information, along with status updates. """ initial_metadatas: list[tuple[int, str]] = [] for i, g_c in enumerate(grounding_chunks): uri = None if (web_info := g_c.web) and web_info.uri: uri = web_info.uri elif (maps_info := g_c.maps) and maps_info.uri: uri = maps_info.uri if uri: initial_metadatas.append((i, uri)) if not initial_metadatas: log.info("No source URIs found, skipping source emission.") return urls_to_resolve = [ uri for _, uri in initial_metadatas if uri.startswith( "https://vertexaisearch.cloud.google.com/grounding-api-redirect/" ) ] resolved_uris_map = {} if urls_to_resolve: num_urls = len(urls_to_resolve) self._emit_status_update( event_emitter, f"Resolving {num_urls} source URLs...", pipe_start_time, ) try: log.info(f"Resolving {num_urls} source URLs...") async with aiohttp.ClientSession() as session: tasks = [self._resolve_url(session, url) for url in urls_to_resolve] results = await asyncio.gather(*tasks) log.info("URL resolution completed.") resolved_uris = [res[0] for res in results] resolved_uris_map = dict(zip(urls_to_resolve, resolved_uris)) success_count = sum(1 for _, success in results if success) final_status_msg = ( "URL resolution complete" if success_count == num_urls else f"Resolved {success_count}/{num_urls} URLs" ) self._emit_status_update( event_emitter, final_status_msg, pipe_start_time, done=True ) except Exception as e: log.error(f"Error during URL resolution: {e}") resolved_uris_map = {url: url for url in urls_to_resolve} self._emit_status_update( event_emitter, "URL resolution failed", pipe_start_time, done=True ) source_metadatas_template: list["SourceMetadata"] = [ {"source": None, "original_url": None, "supports": []} for _ in grounding_chunks ] populated_metadatas = [m.copy() for m in source_metadatas_template] for chunk_index, original_uri in initial_metadatas: final_uri = resolved_uris_map.get(original_uri, original_uri) if 0 <= chunk_index < len(populated_metadatas): populated_metadatas[chunk_index]["original_url"] = original_uri populated_metadatas[chunk_index]["source"] = final_uri else: log.warning( f"Chunk index {chunk_index} out of bounds when populating resolved URLs." ) # Create a mapping from each chunk index to the text segments it supports. chunk_index_to_segments: dict[int, list[types.Segment]] = {} for support in supports: segment = support.segment indices = support.grounding_chunk_indices if not (segment and segment.text and indices is not None): continue for index in indices: if index not in chunk_index_to_segments: chunk_index_to_segments[index] = [] chunk_index_to_segments[index].append(segment) populated_metadatas[index]["supports"].append(support.model_dump()) # type: ignore valid_source_metadatas: list["SourceMetadata"] = [] doc_list: list[str] = [] for i, meta in enumerate(populated_metadatas): if meta.get("original_url") is not None: valid_source_metadatas.append(meta) content_parts: list[str] = [] chunk = grounding_chunks[i] if maps_info := chunk.maps: title = maps_info.title or "N/A" place_id = maps_info.place_id or "N/A" content_parts.append(f"Title: {title}\nPlace ID: {place_id}") supported_segments = chunk_index_to_segments.get(i) if supported_segments: if content_parts: content_parts.append("") # Add a blank line for separation # Use a set to show each unique snippet only once per source. unique_snippets = { (seg.text, seg.start_index, seg.end_index) for seg in supported_segments if seg.text is not None } # Sort snippets by their appearance in the text. sorted_snippets = sorted(unique_snippets, key=lambda s: s[1] or 0) snippet_strs = [ f'- "{text}" (Indices: {start}-{end})' for text, start, end in sorted_snippets ] content_parts.append("Supported text snippets:") content_parts.extend(snippet_strs) doc_list.append("\n".join(content_parts)) sources_list: list["Source"] = [] if valid_source_metadatas: sources_list.append( { "source": {"name": "web_search"}, "document": doc_list, "metadata": valid_source_metadatas, } ) event: "ChatCompletionEvent" = { "type": "chat:completion", "data": {"sources": sources_list}, } await event_emitter(event) log.info("Emitted sources event.") log.trace("ChatCompletionEvent:", payload=event) async def _emit_status_event_w_queries( self, grounding_metadata: types.GroundingMetadata, event_emitter: Callable[["Event"], Awaitable[None]], ) -> None: """ Creates a StatusEvent with search URLs based on the web_search_queries in the GroundingMetadata. This covers both Google Search and Google Maps grounding. """ if not grounding_metadata.web_search_queries: log.debug("Grounding metadata does not contain any search queries.") return search_queries = grounding_metadata.web_search_queries if not search_queries: log.debug("web_search_queries list is empty.") return # The queries are used for grounding, so we link them to a general Google search page. google_search_urls = [ f"https://www.google.com/search?q={query}" for query in search_queries ] status_event_data: StatusEventData = { "action": "web_search", "description": "This response was grounded with a Google tool", "urls": google_search_urls, } status_event: StatusEvent = { "type": "status", "data": status_event_data, } await event_emitter(status_event) log.info("Emitted grounding queries.") log.trace("StatusEvent:", payload=status_event) # endregion 1.1 Add citations # region 1.2 Remove citation markers # TODO: Remove citation markers from model input. # endregion 1.2 Remove citation markers # region 1.3 Get permissive safety settings def _get_permissive_safety_settings( self, model_name: str ) -> list[types.SafetySetting]: """Get safety settings based on model name and permissive setting.""" # Settings supported by most models category_threshold_map = { types.HarmCategory.HARM_CATEGORY_HARASSMENT: types.HarmBlockThreshold.OFF, types.HarmCategory.HARM_CATEGORY_HATE_SPEECH: types.HarmBlockThreshold.OFF, types.HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: types.HarmBlockThreshold.OFF, types.HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: types.HarmBlockThreshold.OFF, types.HarmCategory.HARM_CATEGORY_CIVIC_INTEGRITY: types.HarmBlockThreshold.BLOCK_NONE, } # Older models use BLOCK_NONE if model_name in [ "gemini-1.5-pro-001", "gemini-1.5-flash-001", "gemini-1.5-flash-8b-exp-0827", "gemini-1.5-flash-8b-exp-0924", "gemini-pro", "gemini-1.0-pro", "gemini-1.0-pro-001", ]: for category in category_threshold_map: category_threshold_map[category] = types.HarmBlockThreshold.BLOCK_NONE # Gemini 2.0 Flash supports CIVIC_INTEGRITY OFF if model_name in [ "gemini-2.0-flash", "gemini-2.0-flash-001", "gemini-2.0-flash-exp", ]: category_threshold_map[types.HarmCategory.HARM_CATEGORY_CIVIC_INTEGRITY] = ( types.HarmBlockThreshold.OFF ) log.debug( f"Safety settings: {str({k.value: v.value for k, v in category_threshold_map.items()})}" ) safety_settings = [ types.SafetySetting(category=category, threshold=threshold) for category, threshold in category_threshold_map.items() ] return safety_settings # endregion 1.3 Get permissive safety settings # region 1.4 Utility helpers def _emit_status_update( self, event_emitter: Callable[["Event"], Awaitable[None]], description: str, pipe_start_time: float | None, *, done: bool = False, ): """Constructs and emits a status event in a non-blocking task.""" async def emit_task(): time_str = ( f" (+{(time.monotonic() - pipe_start_time):.2f}s)" if pipe_start_time is not None else "" ) full_description = f"{description}{time_str}" status_event: "StatusEvent" = { "type": "status", "data": {"description": full_description, "done": done}, } try: await event_emitter(status_event) log.debug(f"Emitted status:", payload=status_event) except Exception: log.exception("Error emitting status.") # Fire-and-forget the emission task. asyncio.create_task(emit_task()) def _get_first_candidate( self, candidates: list[types.Candidate] | None ) -> types.Candidate | None: """Selects the first candidate, logging a warning if multiple exist.""" if not candidates: log.warning("Received chunk with no candidates, skipping processing.") return None if len(candidates) > 1: log.warning("Multiple candidates found, defaulting to first candidate.") return candidates[0] def _get_model_name(self, body: "Body") -> tuple[str, bool]: """ Extracts the effective and canonical model name from the request body. Handles standard model names and custom workspace models by prioritizing the base_model_id found in metadata. Args: body: The request body dictionary. Returns: A tuple containing: - The canonical model name (prefix removed). - A boolean indicating if the effective model name contained the 'gemini_manifold_google_genai.' prefix. """ # 1. Get the initially requested model name from the top level effective_model_name: str = body.get("model", "") initial_model_name = effective_model_name base_model_name = None # 2. Check for a base model ID in the metadata for custom models # If metadata exists, attempt to extract the base_model_id if metadata := body.get("metadata"): # Safely navigate the nested structure: metadata -> model -> info -> base_model_id base_model_name = ( metadata.get("model", {}).get("info", {}).get("base_model_id", None) ) # If a base model ID is found, it overrides the initially requested name if base_model_name: effective_model_name = base_model_name # 3. Determine if the effective model name contains the manifold prefix. # This flag indicates if the model (after considering base_model_id) # appears to be one defined or routed via the manifold pipe function. is_manifold_model = "gemini_manifold_google_genai." in effective_model_name # 4. Create the canonical model name by removing the manifold prefix # from the effective model name. canonical_model_name = effective_model_name.replace( "gemini_manifold_google_genai.", "" ) # 5. Log the relevant names for debugging purposes log.debug( f"Model Name Extraction: initial='{initial_model_name}', " f"base='{base_model_name}', effective='{effective_model_name}', " f"canonical='{canonical_model_name}', is_manifold={is_manifold_model}" ) # 6. Return the canonical name and the manifold flag return canonical_model_name, is_manifold_model def _is_flat_dict(self, data: Any) -> bool: """ Checks if a dictionary contains only non-dict/non-list values (is one level deep). """ if not isinstance(data, dict): return False return not any(isinstance(value, (dict, list)) for value in data.values()) def _truncate_long_strings( self, data: Any, max_len: int, truncation_marker: str, truncation_enabled: bool ) -> Any: """ Recursively traverses a data structure (dicts, lists) and truncates long string values. Creates copies to avoid modifying original data. Args: data: The data structure (dict, list, str, int, float, bool, None) to process. max_len: The maximum allowed length for string values. truncation_marker: The string to append to truncated values. truncation_enabled: Whether truncation is enabled. Returns: A potentially new data structure with long strings truncated. """ if not truncation_enabled or max_len <= len(truncation_marker): # If truncation is disabled or max_len is too small, return original # Make a copy only if it's a mutable type we might otherwise modify if isinstance(data, (dict, list)): return copy.deepcopy(data) # Ensure deep copy for nested structures return data # Primitives are immutable if isinstance(data, str): if len(data) > max_len: return data[: max_len - len(truncation_marker)] + truncation_marker return data # Return original string if not truncated elif isinstance(data, dict): # Process dictionary items, creating a new dict return { k: self._truncate_long_strings( v, max_len, truncation_marker, truncation_enabled ) for k, v in data.items() } elif isinstance(data, list): # Process list items, creating a new list return [ self._truncate_long_strings( item, max_len, truncation_marker, truncation_enabled ) for item in data ] else: # Return non-string, non-container types as is (they are immutable) return data def plugin_stdout_format(self, record: "Record") -> str: """ Custom format function for the plugin's logs. Serializes and truncates data passed under the 'payload' key in extra. """ # Configuration Keys LOG_OPTIONS_PREFIX = "_log_" TRUNCATION_ENABLED_KEY = f"{LOG_OPTIONS_PREFIX}truncation_enabled" MAX_LENGTH_KEY = f"{LOG_OPTIONS_PREFIX}max_length" TRUNCATION_MARKER_KEY = f"{LOG_OPTIONS_PREFIX}truncation_marker" DATA_KEY = "payload" original_extra = record["extra"] # Extract the data intended for serialization using the chosen key data_to_process = original_extra.get(DATA_KEY) serialized_data_json = "" if data_to_process is not None: try: serializable_data = pydantic_core.to_jsonable_python( data_to_process, serialize_unknown=True ) # Determine truncation settings truncation_enabled = original_extra.get(TRUNCATION_ENABLED_KEY, True) max_length = original_extra.get(MAX_LENGTH_KEY, 256) truncation_marker = original_extra.get(TRUNCATION_MARKER_KEY, "[...]") # If max_length was explicitly provided, force truncation enabled if MAX_LENGTH_KEY in original_extra: truncation_enabled = True # Truncate long strings truncated_data = self._truncate_long_strings( serializable_data, max_length, truncation_marker, truncation_enabled, ) # Serialize the (potentially truncated) data if self._is_flat_dict(truncated_data) and not isinstance( truncated_data, list ): json_string = json.dumps( truncated_data, separators=(",", ":"), default=str ) # Add a simple prefix if it's compact serialized_data_json = " - " + json_string else: json_string = json.dumps(truncated_data, indent=2, default=str) # Prepend with newline for readability serialized_data_json = "\n" + json_string except (TypeError, ValueError) as e: # Catch specific serialization errors serialized_data_json = f" - {{Serialization Error: {e}}}" except ( Exception ) as e: # Catch any other unexpected errors during processing serialized_data_json = f" - {{Processing Error: {e}}}" # Add the final JSON string (or error message) back into the record record["extra"]["_plugin_serialized_data"] = serialized_data_json # Base template base_template = ( "<green>{time:YYYY-MM-DD HH:mm:ss.SSS}</green> | " "<level>{level: <8}</level> | " "<cyan>{name}</cyan>:<cyan>{function}</cyan>:<cyan>{line}</cyan> - " "<level>{message}</level>" ) # Append the serialized data base_template += "{extra[_plugin_serialized_data]}" # Append the exception part base_template += "\n{exception}" # Return the format string template return base_template.rstrip() def _add_log_handler(self): """ Adds or updates the loguru handler specifically for this plugin. Includes logic for serializing and truncating extra data. """ def plugin_filter(record: "Record"): """Filter function to only allow logs from this plugin (based on module name).""" return record["name"] == __name__ # Get the desired level name and number desired_level_name = self.valves.LOG_LEVEL try: # Use the public API to get level details desired_level_info = log.level(desired_level_name) desired_level_no = desired_level_info.no except ValueError: log.error( f"Invalid LOG_LEVEL '{desired_level_name}' configured for plugin {__name__}. Cannot add/update handler." ) return # Stop processing if the level is invalid # Access the internal state of the log handlers: dict[int, "Handler"] = log._core.handlers # type: ignore handler_id_to_remove = None found_correct_handler = False for handler_id, handler in handlers.items(): existing_filter = handler._filter # Access internal attribute # Check if the filter matches our plugin_filter # Comparing function objects directly can be fragile if they are recreated. # Comparing by name and module is more robust for functions defined at module level. is_our_filter = ( existing_filter is not None # Make sure a filter is set and hasattr(existing_filter, "__name__") and existing_filter.__name__ == plugin_filter.__name__ and hasattr(existing_filter, "__module__") and existing_filter.__module__ == plugin_filter.__module__ ) if is_our_filter: existing_level_no = handler.levelno log.trace( f"Found existing handler {handler_id} for {__name__} with level number {existing_level_no}." ) # Check if the level matches the desired level if existing_level_no == desired_level_no: log.debug( f"Handler {handler_id} for {__name__} already exists with the correct level '{desired_level_name}'." ) found_correct_handler = True break # Found the correct handler, no action needed else: # Found our handler, but the level is wrong. Mark for removal. log.info( f"Handler {handler_id} for {__name__} found, but log level differs " f"(existing: {existing_level_no}, desired: {desired_level_no}). " f"Removing it to update." ) handler_id_to_remove = handler_id break # Found the handler to replace, stop searching # Remove the old handler if marked for removal if handler_id_to_remove is not None: try: log.remove(handler_id_to_remove) log.debug(f"Removed handler {handler_id_to_remove} for {__name__}.") except ValueError: # This might happen if the handler was somehow removed between the check and now log.warning( f"Could not remove handler {handler_id_to_remove} for {__name__}. It might have already been removed." ) # If removal failed but we intended to remove, we should still proceed to add # unless found_correct_handler is somehow True (which it shouldn't be if handler_id_to_remove was set). # Add a new handler if no correct one was found OR if we just removed an incorrect one if not found_correct_handler: self.log_level = desired_level_name log.add( sys.stdout, level=desired_level_name, format=self.plugin_stdout_format, filter=plugin_filter, ) log.debug( f"Added new handler to loguru for {__name__} with level {desired_level_name}." ) # endregion 1.4 Utility helpers # endregion 1. Helper methods inside the Filter class