"""
Copyright 2026 OÜ KAVAL AI (registry code 17393877)
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
"""
import os
import time
from typing import Any, Dict, List, Optional, Tuple, Type
from google import genai
from google.genai import types
from loguru import logger
from pydantic import BaseModel
from kavalai.llm_clients.base_client import (
ensure_user_turn,
BaseLlmClient,
LlmClientException,
ChatHistory,
LlmClientParameters,
ModelStatsReceiver,
)
from kavalai.llm_clients.common import walk_json_schema
from kavalai.llm_clients.streamer import Streamer
_SERVICE_TIERS = {
"priority": types.ServiceTier.PRIORITY,
"standard": types.ServiceTier.STANDARD,
"flex": types.ServiceTier.FLEX,
}
[docs]
class GeminiClient(BaseLlmClient):
"""
Gemini LLM client implementation using the Streamer.
"""
provider = "gemini"
def __init__(
self,
model: str,
llm_client_parameters: Optional[LlmClientParameters] = None,
model_stats_receiver: Optional[ModelStatsReceiver] = None,
api_key: Optional[str] = None,
):
"""
Initialize the Gemini client.
Args:
model: The Gemini model name (e.g., 'gemini-1.5-flash').
llm_client_parameters: Optional parameters like temperature, top_p, etc.
model_stats_receiver: Optional receiver for model call statistics.
api_key: Optional API key (falls back to GEMINI_API_KEY env var).
"""
super().__init__(llm_client_parameters, model_stats_receiver)
self.model = model
self.api_key = api_key or os.getenv("GEMINI_API_KEY")
try:
self.client = genai.Client(api_key=self.api_key)
except ValueError as e:
raise LlmClientException(str(e)) from e
async def _run_chat_completions(
self,
chat_history: ChatHistory,
response_model: Optional[Type[BaseModel]],
streamer: Streamer,
):
"""
Background task to handle the actual Gemini API call and stream results.
"""
start_time = time.perf_counter()
value_streamer = streamer.get_value_streamer(
"response", response_model=response_model
)
system_instruction, contents = convert_messages(
[msg.model_dump() for msg in ensure_user_turn(chat_history.messages)]
)
config_kwargs = {}
params = self.parameters
if params.temperature is not None:
config_kwargs["temperature"] = params.temperature
if params.top_p is not None:
config_kwargs["top_p"] = params.top_p
if params.service_tier is not None:
service_tier = params.service_tier.lower()
if service_tier in _SERVICE_TIERS:
config_kwargs["service_tier"] = _SERVICE_TIERS[service_tier]
logger.info(f"Gemini Service Tier: {service_tier.upper()}")
if params.max_output_tokens is not None:
config_kwargs["max_output_tokens"] = params.max_output_tokens
if params.reasoning_effort is not None:
# Gemini 3 models take the effort as `thinking_level` (minimal,
# low, medium, high); any other value goes out as given for the
# API to reject. The thoughts themselves are streamed on the
# "thought" value streamer below.
config_kwargs["thinking_config"] = types.ThinkingConfig(
include_thoughts=True,
thinking_level=params.reasoning_effort,
)
if system_instruction:
config_kwargs["system_instruction"] = system_instruction
if response_model:
config_kwargs["response_mime_type"] = "application/json"
schema = response_model.model_json_schema()
remove_additional_properties(schema)
config_kwargs["response_schema"] = schema
config = types.GenerateContentConfig(**config_kwargs)
thought_streamer = None
if config_kwargs.get("thinking_config"):
thought_streamer = streamer.get_value_streamer("thought")
prompt_tokens = 0
completion_tokens = 0
cached_prompt_tokens = None
reasoning_tokens = None
finish_reason = None
full_response = ""
try:
async for chunk in await self.client.aio.models.generate_content_stream(
model=self.model,
contents=contents,
config=config,
):
if chunk.candidates:
candidate = chunk.candidates[0]
if candidate.finish_reason is not None:
finish_reason = candidate.finish_reason
if candidate.content and candidate.content.parts:
for part in candidate.content.parts:
if part.thought:
if thought_streamer:
await thought_streamer.stream_partial(part.text)
elif part.text:
full_response += part.text
await value_streamer.stream_partial(part.text)
if chunk.usage_metadata:
usage = chunk.usage_metadata
prompt_tokens = usage.prompt_token_count or 0
# Absent when the cap is spent on thoughts alone.
completion_tokens = usage.candidates_token_count or 0
# Context cache hits and "thinking" tokens, when reported.
cached_prompt_tokens = getattr(
usage, "cached_content_token_count", None
)
reasoning_tokens = getattr(usage, "thoughts_token_count", None)
except Exception as e:
logger.error(f"Gemini Stream Error: {e}")
raise
request_data = {
"model": self.model,
"contents": [str(c) for c in contents],
"config": config_kwargs,
}
usage_counts = {
"prompt_tokens": prompt_tokens,
"completion_tokens": completion_tokens,
"cached_prompt_tokens": cached_prompt_tokens,
"reasoning_tokens": reasoning_tokens,
}
if finish_reason == types.FinishReason.MAX_TOKENS:
raise self._output_truncated(
finish_reason.value,
full_response,
request_data=request_data,
**usage_counts,
)
await value_streamer.stream_complete()
if thought_streamer:
await thought_streamer.stream_complete()
await self._record_completed_call(
request_data=request_data,
response_data=full_response,
started=start_time,
**usage_counts,
)
[docs]
def convert_messages(
messages: List[Dict[str, Any]],
) -> Tuple[Optional[str], List[types.Content]]:
system_instruction = None
contents = []
for msg in messages:
role = msg.get("role")
content = msg.get("content")
if role == "system":
if system_instruction:
system_instruction += "\n" + content
else:
system_instruction = content
continue
# Convert role to Gemini format (user or model)
gemini_role = "user" if role == "user" else "model"
parts = []
if isinstance(content, str):
parts.append(types.Part.from_text(text=content))
elif isinstance(content, list):
for item in content:
if item.get("type") == "text":
parts.append(types.Part.from_text(text=item.get("text")))
contents.append(types.Content(role=gemini_role, parts=parts))
if not contents:
if system_instruction:
contents.append(
types.Content(
role="user", parts=[types.Part.from_text(text=system_instruction)]
)
)
system_instruction = None
else:
contents.append(
types.Content(role="user", parts=[types.Part.from_text(text="...")])
)
return system_instruction, contents
[docs]
def remove_additional_properties(schema: Dict[str, Any]) -> None:
"""Strip ``additionalProperties`` from every object schema, in place.
Gemini's response schema does not accept the field.
"""
walk_json_schema(schema, lambda node: node.pop("additionalProperties", None))