marcopanandClaude Opus 4.6
1e4bc11418
Embeddings timeout was 120s, causing multi-minute hangs when Ollama was down during F4/F6/F7 solved-search. Now: connect_timeout=5s across all HTTP clients (REST + Ollama), read_timeout reduced to 30s for embeddings, and OllamaEmbeddings auto-restarts the server on ConnectionError before degrading gracefully. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
105 lines
3.7 KiB
Python
105 lines
3.7 KiB
Python
import subprocess
|
|
import sys
|
|
import time
|
|
|
|
import requests
|
|
|
|
from tht.config import EmbeddingsConfig
|
|
|
|
DOC_PREFIX = "search_document: "
|
|
QUERY_PREFIX = "search_query: "
|
|
|
|
_RESTART_WAIT = 8 # secondi di attesa dopo aver avviato Ollama
|
|
_RESTART_POLL = 1.0
|
|
|
|
|
|
class EmbeddingsError(Exception):
|
|
pass
|
|
|
|
|
|
class OllamaEmbeddings:
|
|
"""Client embeddings via Ollama. Applica i prefissi di task richiesti da nomic v2:
|
|
ometterli degrada il retrieval in modo silenzioso."""
|
|
|
|
def __init__(self, cfg: EmbeddingsConfig):
|
|
self.cfg = cfg
|
|
|
|
def _is_up(self) -> bool:
|
|
try:
|
|
r = requests.get(
|
|
f"{self.cfg.base_url.rstrip('/')}/api/tags",
|
|
timeout=(self.cfg.connect_timeout, 5),
|
|
)
|
|
return r.status_code == 200
|
|
except requests.RequestException:
|
|
return False
|
|
|
|
def _try_restart(self) -> bool:
|
|
"""Tenta di avviare Ollama e attende che sia raggiungibile."""
|
|
start_cmd = self.cfg.start_cmd if self.cfg.start_cmd is not None else [self.cfg.bin, "serve"]
|
|
if not start_cmd:
|
|
return False
|
|
try:
|
|
subprocess.Popen( # noqa: S603
|
|
start_cmd, start_new_session=True,
|
|
stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL,
|
|
)
|
|
except Exception: # noqa: BLE001
|
|
return False
|
|
print("[embeddings] Ollama non raggiungibile, avvio in corso…", file=sys.stderr)
|
|
deadline = time.monotonic() + _RESTART_WAIT
|
|
while time.monotonic() < deadline:
|
|
time.sleep(_RESTART_POLL)
|
|
if self._is_up():
|
|
return True
|
|
return False
|
|
|
|
def _post(self, url: str, batch: list[str]) -> requests.Response:
|
|
resp = requests.post(
|
|
url, json={"model": self.cfg.model, "input": batch},
|
|
timeout=(self.cfg.connect_timeout, self.cfg.timeout),
|
|
)
|
|
resp.raise_for_status()
|
|
return resp
|
|
|
|
def _embed(self, texts: list[str]) -> list[list[float]]:
|
|
url = f"{self.cfg.base_url.rstrip('/')}/api/embed"
|
|
out: list[list[float]] = []
|
|
for i in range(0, len(texts), self.cfg.batch_size):
|
|
batch = texts[i : i + self.cfg.batch_size]
|
|
try:
|
|
resp = self._post(url, batch)
|
|
except requests.ConnectionError:
|
|
if not self._try_restart():
|
|
raise EmbeddingsError(
|
|
f"Ollama non raggiungibile su {self.cfg.base_url} "
|
|
f"(modello {self.cfg.model}), avvio automatico fallito"
|
|
)
|
|
try:
|
|
resp = self._post(url, batch)
|
|
except requests.RequestException as e:
|
|
raise EmbeddingsError(
|
|
f"Ollama non raggiungibile su {self.cfg.base_url} "
|
|
f"(modello {self.cfg.model}): {e}"
|
|
) from e
|
|
except requests.RequestException as e:
|
|
raise EmbeddingsError(
|
|
f"Ollama non raggiungibile su {self.cfg.base_url} "
|
|
f"(modello {self.cfg.model}): {e}"
|
|
) from e
|
|
embeddings = resp.json().get("embeddings", [])
|
|
for v in embeddings:
|
|
if len(v) != self.cfg.dim:
|
|
raise EmbeddingsError(
|
|
f"dimensione embedding inattesa: {len(v)} != {self.cfg.dim} "
|
|
f"(modello {self.cfg.model})"
|
|
)
|
|
out.extend(embeddings)
|
|
return out
|
|
|
|
def embed_documents(self, texts: list[str]) -> list[list[float]]:
|
|
return self._embed([DOC_PREFIX + t for t in texts])
|
|
|
|
def embed_query(self, text: str) -> list[float]:
|
|
return self._embed([QUERY_PREFIX + text])[0]
|