Spaces:
Sleeping
Sleeping
File size: 8,795 Bytes
f927995 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 | """
Flash-only Gemini wrapper with retries, caching, schema ordering, and fast failover.
"""
from __future__ import annotations
import atexit
import copy
import hashlib
import json
import logging
import re
from concurrent.futures import ThreadPoolExecutor, TimeoutError as FutureTimeout
from datetime import datetime, timedelta
from typing import Any, Dict, Optional
from tenacity import retry, stop_after_attempt, wait_exponential
logger = logging.getLogger("shockmap.gemini_flash")
class GeminiFlashClient:
"""
Centralized Gemini client for Flash-only generation paths.
Notes:
- Uses a short execution timeout so the app degrades quickly instead of hanging.
- Adds property ordering to JSON schema objects for more stable structured output.
"""
def __init__(
self,
genai: Any = None,
model_name: str = "gemini-2.0-flash",
cache_ttl_seconds: int = 1800,
request_timeout_seconds: int = 4,
):
self.genai = genai
self.model_name = model_name
self.cache_ttl_seconds = cache_ttl_seconds
self.request_timeout_seconds = request_timeout_seconds
self._cache: Dict[str, Dict[str, Any]] = {}
self._model_cache: Dict[str, Any] = {}
self._executor = ThreadPoolExecutor(max_workers=4, thread_name_prefix="gemini_flash")
atexit.register(self.close)
def is_available(self) -> bool:
return self.genai is not None
def _make_cache_key(self, namespace: str, payload: Dict[str, Any]) -> str:
blob = json.dumps(payload, sort_keys=True, default=str)
return f"{namespace}:{hashlib.sha256(blob.encode('utf-8')).hexdigest()}"
def _get_cached(self, key: str) -> Optional[Any]:
entry = self._cache.get(key)
if not entry:
return None
if entry["expiry"] <= datetime.now():
self._cache.pop(key, None)
return None
return copy.deepcopy(entry["value"])
def _set_cached(self, key: str, value: Any) -> None:
self._cache[key] = {
"value": copy.deepcopy(value),
"expiry": datetime.now() + timedelta(seconds=self.cache_ttl_seconds),
}
def _coerce_json(self, raw_text: str) -> Dict[str, Any]:
cleaned = raw_text.strip()
if cleaned.startswith("```"):
cleaned = re.sub(r"^```(?:json)?\s*", "", cleaned, flags=re.IGNORECASE)
cleaned = re.sub(r"\s*```$", "", cleaned)
try:
data = json.loads(cleaned)
if isinstance(data, dict):
return data
except Exception:
pass
match = re.search(r"\{.*\}", cleaned, flags=re.DOTALL)
if match:
data = json.loads(match.group(0))
if isinstance(data, dict):
return data
raise ValueError("Gemini response is not valid JSON object")
def _normalize_schema(self, schema: Dict[str, Any]) -> Dict[str, Any]:
"""Recursively normalize nested schema objects for the current SDK."""
normalized = copy.deepcopy(schema)
schema_type = normalized.get("type")
if schema_type == "object":
properties = normalized.get("properties", {}) or {}
normalized["properties"] = {
key: self._normalize_schema(value) if isinstance(value, dict) else value
for key, value in properties.items()
}
elif schema_type == "array" and isinstance(normalized.get("items"), dict):
normalized["items"] = self._normalize_schema(normalized["items"])
return normalized
def close(self) -> None:
try:
self._executor.shutdown(wait=False, cancel_futures=True)
except Exception:
pass
def _get_model(self, system_instruction: Optional[str]) -> Any:
if not self.is_available():
raise RuntimeError("Gemini unavailable")
key = system_instruction or "__default__"
if key in self._model_cache:
return self._model_cache[key]
if system_instruction:
model = self.genai.GenerativeModel(
self.model_name,
system_instruction=system_instruction,
)
else:
model = self.genai.GenerativeModel(self.model_name)
self._model_cache[key] = model
return model
def _invoke_model(
self,
prompt: str,
generation_config: Dict[str, Any],
system_instruction: Optional[str] = None,
) -> Any:
model = self._get_model(system_instruction=system_instruction)
return model.generate_content(prompt, generation_config=generation_config)
@retry(stop=stop_after_attempt(1), wait=wait_exponential(multiplier=1, min=1, max=4), reraise=True)
def _generate(
self,
prompt: str,
generation_config: Dict[str, Any],
system_instruction: Optional[str] = None,
) -> Any:
future = self._executor.submit(
self._invoke_model,
prompt,
generation_config,
system_instruction,
)
try:
return future.result(timeout=self.request_timeout_seconds)
except FutureTimeout as exc:
future.cancel()
raise TimeoutError(f"Gemini request timed out after {self.request_timeout_seconds}s") from exc
def generate_json(
self,
prompt: str,
response_schema: Dict[str, Any],
system_instruction: Optional[str] = None,
temperature: float = 0.1,
max_output_tokens: int = 1024,
min_confidence: Optional[float] = None,
fallback: Optional[Dict[str, Any]] = None,
cache_namespace: str = "json",
) -> Dict[str, Any]:
safe_fallback: Dict[str, Any] = copy.deepcopy(fallback) if fallback else {}
normalized_schema = self._normalize_schema(response_schema)
key = self._make_cache_key(
cache_namespace,
{
"prompt": prompt,
"schema": normalized_schema,
"system_instruction": system_instruction,
"temperature": temperature,
"max_output_tokens": max_output_tokens,
"min_confidence": min_confidence,
"model": self.model_name,
},
)
cached = self._get_cached(key)
if cached is not None:
return cached
if not self.is_available():
return safe_fallback
gen_config = {
"response_mime_type": "application/json",
"response_schema": normalized_schema,
"temperature": temperature,
"max_output_tokens": max_output_tokens,
}
try:
response = self._generate(
prompt=prompt,
generation_config=gen_config,
system_instruction=system_instruction,
)
parsed = self._coerce_json(getattr(response, "text", "") or "")
if min_confidence is not None:
confidence = float(parsed.get("confidence", 0.0))
if confidence < min_confidence:
return safe_fallback or parsed
self._set_cached(key, parsed)
return parsed
except Exception as exc:
logger.warning("Gemini JSON generation failed: %s", exc)
return safe_fallback
def generate_text(
self,
prompt: str,
system_instruction: Optional[str] = None,
temperature: float = 0.2,
max_output_tokens: int = 1500,
fallback: str = "",
cache_namespace: str = "text",
) -> str:
key = self._make_cache_key(
cache_namespace,
{
"prompt": prompt,
"system_instruction": system_instruction,
"temperature": temperature,
"max_output_tokens": max_output_tokens,
"model": self.model_name,
},
)
cached = self._get_cached(key)
if cached is not None:
return str(cached)
if not self.is_available():
return fallback
gen_config = {
"temperature": temperature,
"max_output_tokens": max_output_tokens,
}
try:
response = self._generate(
prompt=prompt,
generation_config=gen_config,
system_instruction=system_instruction,
)
text = (getattr(response, "text", "") or "").strip()
self._set_cached(key, text)
return text
except Exception as exc:
logger.warning("Gemini text generation failed: %s", exc)
return fallback
|